PL3238 Summarised Study Guide

Tuesday, 10 August 2021
Week 1 Readings Summary

Fiske & Taylor (2021)
I. Social Cognition is the scienti c study of how people make sense of other people and
themselves. It focuses on how ordinary people think about people and how they think
they think about other people. (Perceive, interpret, remember, and represent information
about themselves and other individuals/groups.
A. People’s understanding of the social world can be studied by asking them how they
make sense of others. (Heider, 1958). This is the route of phenomenology. To
describe how ordinary people say they experience the world. If you are right, you
can build formal theories by making their insights scienti c, by pulling together
patterns across many people’s intuitions. Even if people are wrong, they can study
other’s common-sense theories in and of themselves, how people think and not
their accuracy, is the phenomenon of interest.
B. Social Cognition research is also concerned, secondly, with common sense theory
or “naive psychology”, for its own sake. That is, people’s everyday theories about
each other are themselves interesting to study. Thus, if the person at the party has
some ideas about how people form impressions of each other, the person’s informal
ideas are interesting in their own right. Hence, common-sense psychology is useful
for 2 reasons.
C. Social cognition research also goes beyond naive psychology. It also entails a ne
grained analysis of how people think about themselves and others, and it leans
heavily on the theories and methods of cognitive psychology. The in ux of ne-
grained or detailed models from cognitive psychology is one of the hallmarks of the
current approaches to social cognition. These models are important because they
describe mechanisms of learning and thinking that apply in a wide variety of
areas, perhaps including social perception. These models are general and because
cognitive processes presumably in uence social behaviour heavily, it makes sense
to adapt cognitive theory to social settings.
1. Both naive psychology and cognitive viewpoints are important themes in social
cognition research. These 2 viewpoints characterise the double appeal of social
cognition, the entertaining part of studying how people think about tohers is its
appeal to your intuitions, and the ne grained part forces you to be accurate and
precise.
II. 2 broad intellectual approaches to studying social cognition exists. Elemental and
Holistic. Elemental approach is characterised by breaking scienti c problems down to
pieces and analysing the pieces in separate detail before combining them. Holistic
approach is characterised by analysing the pieces in the context of other pieces and
focusing on the entire con guration of relationships among them. Distinction becomes
clearer as the 2 approaches are described.
A. Elemental Origins of Social Cognition Research.

1. Psychology used to be a branch of philosophy. Philosophers provided some
basic principles of mind that still carry weight today. In elemental tradition of the
British philosophers, the mind is likened to chemistry, in which ideas are the
elements. Any concept, whether concrete like “salt” or abstract like “shame” is a
1
fi
fi
fi
fl
fi
fi
fl
fi
fi
basic element. Any element associated with any other element. The bonds
between concepts create mental chemistry.
2. Ideas here rst come from our sensations and perceptions. Then they are
associated by contiguity in space and time. If salt is next to pepper on a table,
the 2 can become a unity through contiguity. Repetition is the key the moving
from simple contiguity to a mental compound. When you subsequently think of
salt, you will think of pepper. Inevitably. Something like pairing.
3. Proper methods of testing psychological science emerged. Empirical methods.

Laboratory psychologists such as Wilhelm Wundt and Hermann Ebbinghaus,
trained themselves and graduate students to observe their own thought
processes, to introspect on how they committed ideas to memory and on how
they retrieved ideas from memory. They analysed experiences into its elements,
determined how they connected, and determined the laws that governed the
connections.
B. Holistic Origins of Social Cognition Research.
1. Tackles the whole mind at once, by Immanuel Kant. In his POV, mental
phenomena are inherently subjective. The mind actively constructs a reality that
goes beyond the original thing in and of itself. E.g. a bunch of grapes is
perceived as a unit, but that perception is a construction of the mind. Perceiving
a bowl of grapes di ers from perceiving each individual grape separately.
Similarly, if someone cuts o some grapes and the remaining ones topple out of
the bowl, the 2 movements are perceived as linked in a cause-e ect relationship.
Again, that perception is furnished by the mind, it is not inherent in the stimulus.
(Something else may have caused the remaining grapes to topple).
2. Gestalt Psychology drew on these initial holistic insights. In contrast to analysis

into elements, psychologists who used Gestalt methods rst describe the
phenomenon of interest, the immediate experience of perception, without
analysis. This method, called phenomenology, focuses on systematically
describing people’s experience of perceiving and thinking. It later became
one of the major foundations of social cognition research. The reliance on asking
people how they made sense of the world.
3. Although both the elemental and holistic groups drew on introspections, Gestalt
focused on people’s experience of dynamic wholes & elementalists focused
on the expert’s ability to break the whole into pieces. Imagine a song in your
mind. A song can be perceived as a series of individual notes (elemental) or as a
melody that emerges from the relationships among the notes (Gestalt). The
emergent structure is lost by analysing it into sensory elements, in the Gestalt
view. Gestalt psychologists saw the mental chemistry metaphor of the
elementalists as misguided because a chemical compound has properties not
predictable from its isolated elements. Similarly, the perceptual whole has
properties not discernible from the isolated parts. For example, the note middle
C can seem high in the context of many lower notes or low in the context of
many higher notes, but it would not stand out at all in the context of other notes
close to it. Psychological meaning goes beyond raw sensory parts to include the
organisation people impose on the whole.
4. Focused on a person’s own subjective perceptions, not on objective analysis.

Emphasised the in uence of social environment as perceived by the individual,
2
fi
fl
ff
ff
fi
ff
which is called the “psychological eld”. (Kurt Lewin, 1951). A full
understanding of a person’s psychological eld thus cannot result from an
objective description by others of what surrounds the person. The crucial factor
is the person’s own interpretation. However, it isn’t the case that the person can
necessarily verbalise his or her own perceived environment, but that the person’s
own reports typically provide better clues than do the researcher’s intuitions. For
instance, a researcher may objectively report that Barb complimented Ann on
her appearance. The researcher may even have strong hunches why Barb did
so. But Ann’s reaction will depend on her own perception of Barb ’s intent,
ingratiation, envy, reassurance, or friendliness. A prime way to nd that out is to
ask Ann to describe what happened in her own terms. Just as in Gestalt
psychology, Lewin emphasised the individual’s phenomenology, the individual’s
construction of the situation.
5. Kurt Lewin also insisted on describing the total situation, not its isolated
elements. A person exists within a psychological eld that is a con guration of
forces. One must understand that all the psychological forces operating on the
person in any given situation in order to predict anything. E.g. some forces might
motivate one to study (an upcoming exam, the sight of one’s roommate
studying), but other forces might motivate one to spend the evening another way
(a group of friends suggesting a movie). No one force predicts a certain action,
but the dynamic equilibrium among them, the everchanging balance of forces,
does predict action.
6. The Total Psychological Field, (and hence behaviour), is determined by 2 pairs

of factors. The rst pair consists of the person in the situation. Neither alone is
su cient to predict behaviour. The person contributes needs, beliefs, perceptual
abilities, and more. These act on the environment to constitute the psychological
eld. Thus, to know what a particular person is motivated to study does not
predict whether or how much he/she will study. But, a motivated person in a
library is extremely likely to study a lot. Both the person and the situation are
key to predicting behaviour. The study of social cognition focuses on
perceiving, thinking, and remembering the function of who and where one
is.
7. Second pair of factors consist of cognition and motivation. Both are joint
functions of person and situation. Both are essential to predicting behaviour.
Cognition provides the perceiver’s own interpretation of the world, without clear
cognitions, behaviour is not predictable. If a person has incomplete cognitions or
confused cognition about a new setting, behaviour will be unstable. For
example, if you do not have the foggiest cognition about what an upcoming
exam in music composition will be like, you may behave erratically and hence
unpredictably, you may try several study strategies. None of them very
systematically. Cognitions help determine what a person will do, which
direction behaviour will take. If a musician friend explains what composition
exams typically contain, your cognitions and hence your studying will settle
views over time.
III. The Ebb and Flow of Cognition in Psychology.
A. Psychologists have not always agreed that it is important to get inside the mind. The
study of cognition has received both good and bad reviews over time. To prevent an
3
fi
ffi
fi
fi
fi
fi
fi
fi
overly myopic view of the importance of cognition, we need to take a good look at
experimental and social psychology. Early psychologists tend to rely heavily on
introspection as a central tool for understanding human thought. However, it soon
developed a bad reputation, and with it, cognition fell into disrepute.
B. Cognition in Experimental Psychology.
1. Wundt’s work at the dawn of empirical psychology as already described, relied

heavily on trained introspection. The use of introspection was linked to the fact
that Wundt’s goal was emphatically cognitive. People’s experience was the
subject matter of interest. Wundt and others gathered data about mental events
and constructed theories to account for those data. However, introspection was
ultimately abandoned as a methodology in experimental psychology because it
did not conform to the principles appropriate to scienti c investigation. By usual
scienti c standards and methods, one’s data should be publicly reproducible.
In early experimental psychology, theories were required to account for
introspections (i.e. self observations), and therein lay the problem. If the criteria
for a theory’s success depended on private experience, the evidence could not
be produced in public. The research could not be checked by others. The most
absurd version of the problem would be this, if my theory accounts for my
introspections and your theory accounts for yours, how do we decide who is
right?
2. Behaviourists believe that specifying an observable stimulus (S) and response ®,

for every part of one’s theory is the strict scienti c discipline necessary to the
advancement of psychology, including social psychology (Berger & Lambert,
1968). For example, a behaviourist might approach the topic of racial and ethnic
discrimination by noting that some children are punished for playing with
children of certain other ethnic groups and rewarded for playing with children of
the family’s own ethnic group. A simpli ed model of this would include “the other
ethnic group” as the stimulus and “not playing together” as the response. A
behaviourist would not consider the possible role of stereotyping (cognition).
3. Linguists criticised the failure of the stimulus response framework’s attempts to

account for language. It became clear that the complex, symbolic, and uniquely
human phenomenon of language would not easily yield to behaviourist
approaches.
4. Information Processing arose out of work on how people acquire knowledge

and skills (Broadbent, 1958). Information processing refers to the idea that
mental operations can be broken down into sequential stages. Information
processing theory might represent my cognitive operations as: understand the
question’s meaning > search for information on that topic > verify answer > state
answer. The point of this is that one tries to specify the steps intervening
between stimulus (question) and response (answer). From this POV, the
important feature is the sequential processing of information. Information
processing approaches entail the e ort to specify cognitive processes, which
behaviourists would not do.
5. New scienti c tools have developed that allow cognitive psychologists to trace
the non-observable processes presumed to intervene between stimulus and
response. The most important of these tools is the computer, which has become
a methodological tool as well as a theoretical metaphor. It serves as a tool in that
4
fi
fi
ff
fi
fi
fi
cognitive scientists actually use computers to stimulate human cognitive
processes, they write complex programs that play chess, learn geometry, and
summarise the news. Social cognition researchers have also developed
computer simulations of how people form impressions and memories of each
other (Hastie, 1988; Linville, Fischer, & Salovey, 1989; E.R. Smith, 1988). The
computer is also a metaphor in that it provides a framework and a jargon for
characterising mental processes, psychologists talk about input-output
operations or memory storage and retrieval, with respect to human cognition.
More importantly, most current theory builds on the idea that human cognition
resembles computer information processing in important ways.
C. Cognition in Social Psychology 

Social psychology has consistently leaned on cognitive concepts, even when most
psychology was behaviourist. Social psychology has always been cognitive in at
least 3 ways.  
-1st, since Lewin, social psychologists have decided that social behaviour is more
usefully understood as a function of people’s perceptions of their world, rather than
as a function of objective descriptions of their stimulus environment. E.g. an
objective reward like money or praise that people perceive as a bribe or as attery
will in uence them di erently than a reward they perceive as without manipulative
intent. What predicts their reaction is their perception and not just the giver’s
actions. 
-2nd, other people can in uence a person’s actions without even being present,
which is the ultimate reliance on perceptions to the exclusion of objective stimuli.
Thus, someone may react to a pro ered bribe or to attery by imagining the
reactions of others (“What would my mother say?”, “what would my friends think?”).
The causes of social behaviour are doubly cognitive, our perceptions of others
actually present and our imagination of their presence both predict behaviour. 
Social psychologists view not only causes but also the end result of social
perception and interaction in heavily cognitive terms, and this is a second way in
which social psychology has always been cognitive. Thought often comes before
feeling and behaving as the main reaction that social researchers measure. A person
may worry about a bribe (thought), hate the idea (feeling), and reject it (behaviour),
but social psychologists often mainly ask. “What do you think about it?”. Even when
they focus on behaviour and feelings, their questions are often, “What do you intend
to do?” and “How would you label your feeling?”. These arguably are not behaviour
and feelings but cognitions about them. Results are largely cognitive as a result. 
-3rd. The person in between the presumed cause and the result is viewed as a
thinking organism, this view contrasts with viewing the person as an emotional
organism or a mindless automaton (Manis, 1977). Many social psychological
theories paint a portrait of the typical person as reasoning (perhaps badly) before
acting. In attempting to deal with complex human problems, as social psychology
always has, complex mental processes seem essential. How else can one account
fo stereotyping, prejudice, propaganda and persuasion, altruism and aggression,
and more? A strict stimulus-response (S-R) theory does not include the thinking
organism that seems essential to account for such problems. In several senses
then, social psychology contrasts with strict S-R theories in its reliance on S-O-R
theories that include stimulus, organism, and response. Consequently, the thinker,
5
fl
ff
fl
ff
fl
fl
who comes between stimulus and response, has always been paramount in social
psychology.
1. Social Thinkers has taken many guises. These describe the various roles of
cognition in social psychology. Besides the varied roles of cognition, motivation
has played di erent roles in the view of the social thinker. Keeping in mind these
2 components of cognition and motivation. We can identify 4 general views of
the thinker in social psychology; consistency seeker, naive scientist, cognitive
miser, and motivated tactician.
a) Consistency Seekers are motivated by perceived discrepancies among

their cognitions. E.g. if David knows he is on a diet and knows that he has
just eaten a hot fudge sundae, he must do some thinking to bring those 2
cognitions into line. 
Places cognitive activity in a central role. Objective inconsistency is therefore
not important. E.g. if would-be dieters can convince themselves that one
splurge will not matter, eating a sundae is not inconsistent for them. 
Subjective inconsistency will be central to these theories. 
However, once inconsistency is perceived, the person is presumed to feel
uncomfortable, and subsequently motivated to reduce the inconsistency.
b) Naive Scientist consist of attribution theories that concern how people

explain their own and others’ behaviour. Attribution theories describe
people’s causal analyses of or attributions about the social world. 
Long story short, the naive scientist model is an outcome of fairly rational
analysis, and involve weighing many things, including situation setting,
emotional in uence of di erent parties, evidence for situational/dispositional
cause, e.g. why is my girlfriend so irritable today?.  
Even if you are wrong, early theories would have viewed your error as an
emotion-based departure from the normal process or as a simple error in
available information. Errors arise mainly as interference from non-rational
motivations. 
In general, basically scientists who have to deal with subjectivity and
incomplete information, whilst carefully pursuing the truth.
c) Cognitive Miser is the idea that people are limited in their capacity to
process information, so they take shortcuts whenever they can. People
adopt strategies that simplify complex problems; the strategies may not be
normatively correct or produce normatively correct answers, but this is
because they emphasise e ciency. 
Ideally, we want to be right, but in the real world, we need to search for rapid
adequate solutions, rather than slow accurate ones. 
Consequently, errors and biases stem from inherent features of the cognitive
system, not necessarily from motivations.  
This model does not address the issue of motivations or feelings of any sort.
d) Motivated Tactician is a fully engaged thinker who has multiple cognitive

strategies available and chooses among them based on goals, motives and
needs. Sometimes the motivated tactician chooses wisely, in the interests of
adaptability and accuracy, and sometimes the motivated tactician chooses
defensively, in the interests of speed or self-esteem. 
6
fl
ff
ff
ffi
IV. What is Social Cognition? 
Object of stiudy concerns how people make sense of other people and themselves. As
a topic, it is relevant to the study of attitudes, small groups, and so much more.  
Contains unabashed mentalism, orientation toward process, cross fertilisation between
cognitive and social psychologies, and at least some concern on real world social
issues.
A. Mentalism: A Commitment to Cognitive Elements
1. First assumption is that people have an unabashed commitment to mentalism

(cognition). Cognitive elements people naturally use to make sense of other
people constitute the rst third of this book, the “what” of social cognition. 
Attributions are people’s causal explanations for events in the social world.
2. Social Schema. Schema is a cognitive structure that represents one’s general

knowledge about a given concept or stimulus domain. 
E.g. your friend John, you may organise him into a schema that includes your
view of him as independent but not a loner, friendly but not saccharine, athletic
but not a star. A schema for a concept for John includes relevant attributes (e.g.
independent, friendly, athletic), and the relationships among these attributes (e.g.
what his independence has to do with his friendliness, he is friendly without
being overly reliant or dependent on relationships) help us provide expectations
that enable us to function and depict the real world in our minds.
B. Cognitive Processes in Social Settings
1. Formation of cognitive elements. 

In behaviourism, scientists avoid discussion of internal processes because they
were concerned with predicted a publicly observable response from a publicly
observable stimulus. In that sense, they were response or outcome oriented,
rather than process oriented that may not be readily observable. 
In Social Cognition research, one must attend to the stimulus and encode it as
an internal representation of external reality. This eld attempts to measure the
stages of social information processing. When one is faced with a social
stimulus, there are several steps posited to occur before one responds. 
One must attend to the stimulus and encode it as an internal representation of
external reality, but also sometimes we can operate without much thought. 
After encoding the stimulus, one can store it away in memory. This memory is
hence organised for information about other people, contents of memory for
others, and the relationships among memory, judgement, and goals. What we
remember and how we judge are surprisingly independent processes. 
People make inferences about social events. Social perceivers make inferences
by specifying relevant information, sampling that information, and combine it into
some judgement. Heuristics, provide e ciency at some cost to accuracy, but
many errors are inconsequential, self-correcting, or subject to improvement.
C. Cross Fertilisation, Studying Social Cognitive Processes
1. Social Cognition borrows relatively ne-grained cognitive theory and methods

that have proved fruitful for social psychological research. They specify the steps
in a presumed process model and also attempt to measure the steps in some
detail. Various traditional and newer experimental methods allow researchers to
support di ering aspects of process models. 
We now attempt to trace the processes of attention, memory, inference.
D. Beyond Cognition, Real World Social Issues
7
ff
fi
fi
ffi
fi
1. Application to the real world. Research in social cognition informs us about
important social trends, issues, and phenomena. It applies heavily cognitive
theory and method to real world social problems. 
Social cognition guides work in areas such as psychotherapy, healthcare, legal
system, stereotyping, advertising, political campaigns, altruism, and romantic
involvements. These wide array of areas covered illustrate the exibility of social
cognition research. Also demonstrates how some otherwise highly technical or
abstract ideas generalise outside the laboratory. 
Social cognition applications to real world issues not only contain a purely
cognitive analysis, but also considers other factors such as interpersonal
settings, environment, emotional factors, etc. E.g. what happens when an
cognitive miser encounters feelings? What relationship does social information
processing have to situations of intense personal involvement? How do social
cognitions or how much do these theories translate into actual behaviour? 
V. People Are Not Things
A. People intentionally in uence the environment, they attempt to control it for their
own purpose. Objects are not intentional causal agents.
B. People perceive back, as you are busy forming impressions of them, they are doing
the same to you. Mutual cognition.
C. Social cognition implicates the self, because the target is judging you, because the
target may provide you with information about yourself, and because the target is
more similar to you than any object could be.
D. Social stimuli may change upon being the target of cognition. People worry about
how they come across and may adjust their behaviour and appearances
accordingly. Co ee cups do not.
E. People’s traits are non-observable attributes that are vital to thinking about them.
Object’s non-observable attributes are somewhat less crucial. Both a person and a
cup can be fragile, but that inferred characteristic is both less important and more
directly seen in the cup.
F. People change over time and circumstance more than objects typically do. This can
make cognitions rapidly obsolete or unreliable.
G. The accuracy of one’s cognitions about people is harder to check than the accuracy
of one’s cognitions about objects. Even psychologists nd it hard agreeing on
whether a given person is extraverted, sensitive, or honest, but most ordinary
people could easily test whether a cup is heat-resistant, fragile, or leaky.
H. People are unavoidable complex. Cognitions cannot be studied without making

numerous choices to simplify.
I. Because people are so complex, they have innate traits and intents hidden from
physical view, and because these a ect us in ways objects do not, social cognition
automatically involves social explanation.
Baumeister - Thinking About People
I. Social Psychologists and Thinking About People 

Social psychology is open to almost anything in the realm of normal human
behaviour. Many sub elds of psychology are de ned by a speci c focus, on
mental illness, on children, or on brain processes. Social psychology has no
8
ff
fl
fi
ff
fi
fi
fl
fi
speci c focus. Anything contributing to an increased understanding of how
people in general think, feel and act is welcome. 
 
Understanding People 
Psychology involves studies of motivation and cognition, and it is possible to
trace the history of psychology as a series of pendulum swings to emphasise
one or another. Wundtian introspectionist school focused on cognition. Freudian
Theory emphasised motivation. Learning in animals is considered cognitive. If
we put them all together, we are likely to get a balanced and probably fairly
accurate view.
A. The Consistency Seeker 

People are motivated to seek consistency. Dominant view in late 1950s and
1960s and has remained in uential ever since. View that emphasises
motivated cognition, or perhaps motivations about cognition. 
E.g. cognitive dissonance was one, but it was not really cognitive by modern
standards. 
The purest form of this idea was that people are interested in consistency
much of the time and are perhaps constantly alert for possible
inconsistencies in their knowledge and interpretation of events. So if
someone does something strikingly inconsistent with what they have said,
done, or thought about before, they get motivated to reduce this
inconsistency. 
No longer the major or central aspect of human social life.
B. The Self-Esteem Maximiser 

Motivation to maintain self-esteem was seen as driving task performance
and responses to failure, interpersonal strategies, defensive cognitive styles,
stress, emotion, risk taking, and much more. 
The Maximiser seeks above all to avoid losing self-esteem. They avoid
situations that may cause a downward revision in self-appraisal, and they
have the urge to enhance their favourable view of self, in some versions but
not in others. It is somewhat more controversial to say this than the urge to
avoid losing self-esteem. 
New research suggests that people not only want to think that they are great
individually, but they also want to believe that their close relationships are
exceptionally good.
C. The Terror Manager 

Highly distinctive, well-integrated theory of human nature has been
advanced under the rubric of Terror Management Theory. This was originally
inspired by writings of anthropologist Ernest Becker (1973), who proposed
9
fi
fl
that humans are unique among living things in knowing that they will
eventually die. Becker proposed that much of human behaviour can be
understood as a motivated response to the fear of death. 
Basically, the notion of death avoidance as the master motive. This provides
a basis for explaining a great many, and potentially all, human actions and
strivings.
D. The Information Seeker 

This is the relatively colder (i.e. cognitive images), rather than the relatively
warmer (i.e. motivational) image of humankind. 
This emphasises thinking and processing information as the paramount
human activity. Motivation is quietly downplayed in some variations on these
approaches, recognised but simply not considered in others, and actively
denied (for the most part) in still others. 
Theory is that it is important and helpful for people to understand their
worlds, and so they constantly go about trying to collect information. The
drive to understand the environment is probably present even in simple
animals, who bene t from being able to predict events in their physical
surroundings. 
Includes making inferences about other people, as well as social situations
and structures. 
Central assumption of the Information Seeker approach was that whenever
something happens, you pass a test, get rejected by a romantic partner,
meet someone new, have an argument, you respond by trying to determine
what it means what its implications are. Attribution theory was one of the
most dominant theories in social psychology from the late 1960s into the
1980s, took this approach. 
Basically a view that combines the basic cognitive, curious, avid learner with
the understanding that most individuals have a fairly strong set of
preferences for what to learn. The Information Seeker may want to learn the
truth about themself, regardless of what it is, but the Motivated Information
Seeker (like the Self-Esteem Maximiser) much prefers to hear favourable
rather than unfavourable things about the self.
E. The Information Processor 

Simple view of humans as Information Seekers gave way in the 1970s to the
realisation that information was not simply taken in but rather was subjected
to fairly extensive processing. These are a more sophisticated version of the
Information Seekers. 
Depicts humans as scouring the world for information, quickly guring it out
with a couple of attributions, and storing those conclusions for future use.
On top of this, it also recognises that considerable inner mental work
10
fi
fi
occurred when the information was rst encountered. Instead of an
attribution or two, the processing involves selective attention, extensive and
fallible interpretation processes, partial encoding into memory and at best
modestly reliable retrieval from memory, assimilation of new information to
existing knowledge, mental shortcuts, and numerous other processes. 
E.g. why might people take more responsibility for success than failure, in
the standard self-serving attributional bias e ect (Jones et. ak., 1972;
Zuckerman; 1979). Motivational explanation was that people want to believe
good things about themselves, so they more readily accept success than
failure as a true sign of their worth. But it was also possible to pose a purely
cognitive explanation. Maybe people expect success more than failure, and
so failure violates their expectancies in a way that success does not. Violated
expectancies cause them to engage in more cognitive processing after
failure than success, and the intensi ed scrutiny will sometimes reveal
reasons not to take the failure to heart. In that view, it has nothing to do with
wanting to think well of oneself. 
Basically, although these new views of the Motivated Information Processor
do allow some scope and in uence to motivation, it is treated as secondary,
because motivation is seen as something that mainly interferes with
cognitive processing or, at best, can occasionally focus cognitive processing
on things that are important.
F. The Foolish Mistake Maker 

Variation of the Information Processor, rede ning it as someone who
processes information badly. 
Social Psychologists searched for counterintuitive ndings that went against
what most people assumed and expected. Research on social cognition also
showed that even though people might reach the correct conclusion at
times, it was not very informative about the inner processes involved, and a
premium was placed on showing instances in which people came to false
conclusions or made other errors. 
The study of why and how people make mistakes in systematic, predictable
ways. This provides valuable insights and it would be unfair to stigmatise the
entire line of work based on some excesses and unfortunate tendencies. 
Basically, a general sense that that is not all that human beings are. There are
even reasoned, thoughtful critiques suggesting that much of what is called
error and bias should not be thus disparaged, partly because the same inner
processes that produce the occasional well-documented errors in studied of
social psychology also produce correct answers most of the time.
G. The Non-di erent Individual / Situational Responder 

Underlying theory is that behaviour is primarily a response to situations, how
11
ff
fl
fi
fi
fi
ff
fi
people think, feel, and act are direct results of situational pressures and
in uences. This view of humanity states that there is not a great deal inside
them, other than mechanisms to help them respond to their immediate
situation. 
Today, most psychologists recognise that both personality traits and
situational factors contribute important insights to predicting and
understanding human behaviour.
H. The Impression Manager 

People simply try to present themselves to others in ways that make a good
impression. People here again, do not have much personal depth in contrast
to Freudian or personality theories, but simply have the inner processes that
enable them to adapt to the situation. 
Impression Managers care greatly about what others think, and so in that
sense the theory has a strong motivational component. The Manager can be
a chameleon, changing colours to suit the situation. They do come equipped
with a possibly extensive set of inner mechanisms for discerning what others
prefer and for altering his or her own behaviour accordingly. Self-
presentational strategies and tactics are chosen according to what will work
best. 
Basically, what the Impression Manager does it simply expedient, and
dependent on situational forces almost entirely. 
Has gone out of fashion, as people realised that there were powerful inner
forces and processes at stake. 
Few really think that such processes provide anything close to a thorough
account of the human individual and human social behaviour. Albeit,
impression management does consist of helpful set of strategies and
behaviours that accompany the extensive inner cognitive processes and
serve its’ motivations.
I. The Naturally Selected Animal 

Basically similar to many other animals, although perhaps a bit more
complicated in view of its’ high intelligence, invention of language, and
mastery of technology. Still, the same basic principles apply. Naturally
Selected Animals still want to survive and reproduce. Many behaviour
patterns have become divorced from their overt connection to survival and
reproduction but remain in place because they contributed to survival and
reproduction in the past. 
E.g. many people desire sex without reproduction, and in fact quite a few of
them take extensive precautions to achieve this, although their patterns of
desire are still shaped by what produced the best reproductive results in the
past. 
12
fl
More than just the above, Naturally Selected Animal theory shows how
humans favour relatives over strangers, form groups easily, and is interested
in dominance (i.e. rising to the top of a group hierarchy).
J. Cultural Animal 
Synthesis and compromise among many other views, so it is less
provocative than most. Human psyche was shaped by evolution but also
recognises the importance of culture. 
Instead of the leash metaphor of the Naturally Selected Animal Theory, which
assumed that nature came rst, laying the foundation for human behaviour,
and culture following after the evolutionary process was done, the Cultural
Animal argues that culture in uenced our evolution. This does not require
that speci c cultural practices were produced by evolution, but rather that
culture became part of the selection environment, so that traits favourable to
culture evolved. 
E.g. following the advent of the human language in social environments,
people who were better able to talk and understand speech became more
successful at surviving and reproducing than people who lacked the
biological capabilities to use language well. 
Basically, main function of culture is to prevent actual death (not just the idea
of it). Culture is the way humans solve the basic natural problems of survival
and reproduction.
K. The Group Member 

Involves some loss of individuality within the group. The Group Member can
become deindividuated, engage in groupthink, and might even participate in
mob violence. These negative e ects reveal the group aspect of The Foolish
Decision Maker. 
E.g. social loa ng, crowding, social facilitation, and di usion of responsibility
in bystander intervention. 
The Group Member need not be a bad person, and the motivations of Group
Members di er somewhat depending on which of the 2 approaches are
taken. One approach considers processes within the group ( nd ways to be
accepted and liked by the other members, often requires determining how
the member is similar to them and can t with them). Other approach is to
look at processes between groups. Intergroup processes have become
dominant focus of social psychology in Europe and Australia and have also
been studied elsewhere. Emphasis is on how the individual identi es with the
group and relates to members of other groups.
L. The Benighted Layperson 

The everyday person who thinks or does socially undesirable things. Social
13
fi
ff
fi
fi
fl
ff
fi
ff
fi
fi
psychology will hence need to teach this person how to be a better person,
for the good of all. Not a dominant view in the eld, requires consensus that
the job of science is to instil social values into the general public. 
Someone who is prone to holding various prejudices, especially toward
women and minorities. Not environmentally friendly, aggressive, unhelpful,
does not treat others properly, does things that are harmful to the self.

• Heuristics
• We use heuristics because it is a mental shortcut that allows us to solve problem and
make judgements/decisions quickly and e ciently. These rule of thumb strategies
shorten decision making time and allow people to function without constantly stopping
to think about their next course of action.
• The world in which we make judgements is full of uncertainties, and we do not

spontaneously use statistical rules (too time consuming).
• Representativeness Heuristic
• Involves classifying things according to how similar they are to the typical case.
• Occurs because people ignore prior probabilities (base-rates) in the population.
• Representativeness and Medicine

• The cure should resemble the disease. Ancient Chinese medicine, people with vision
problems were fed ground bat. (Deutsch, 1977).
• Ratings of the Chandorans showed… rated as more irritable and aggressive when
described as “boar-eaters” than as “turtle eaters”
• Representativeness and the conjunction fallacy

• The conjunction fallacy causes failure to recognise that the co-occurrence of 2
outcomes cannot be greater than the probability of each outcome alone.  
E.g. Event A = Tornado, Event B = Hail. The probability of tornado AND hail is less
probable or equally than just a tornado or just hail.
Schwarz, N. et al. (1991). Ease of Retrieval as information. Availability

Heuristics. 
I. Experiment Background
A. Widely shared assumption in decision making as well as social judgement holds that
people estimate the frequency of an event, or the likelihood of its occurrence, by the
ease with which instances or associations come to mind.
1. Tversky & Kahneman, 1973. Most frequently studied study of Experiment 8,

subjects were read 2 lists of names, one presenting 19 famous men and 20 less
famous women, and the other presenting 19 famous women and 20 less famous
men. When asked, subjects reported that there were more men than women in
the rst list but more women than men in the second list, even though the
14
fi
ffi
fi
opposite was true by a di erence of 1. This meant that presumably, the famous
names were easier to recall than non-famous ones, resulting in an
overestimation. In fact, subjects were able to recall about 50% more of the
famous than of the non-famous names. It remains unclear what drove the
overestimate.
2. Tversky & Kahneman, 1973, Experiment 3. Subjects were found to overestimate

the number of words that began with the letter r but to underestimate the
number of words that had r as the third letter. 
Similarly, Gabrielcik and Fazio 1984 observed that exposing subjects to
subliminally presented words containing the letter t increased subjects’
estimates of the frequency of t words. Again, these ndings may re ect either
subjects could generate more words beginning with an r, or including a t if
primed or that they relied on the ease with which relevant exemplars could be
called to mind. 
Sherman & Corty, 1984; Strack, 1985; Taylor, 1982. These experiments show that
typically, the manipulations that are introduced to increase the subjectively
experienced ease of recall are also likely to a ect the amount of subjects recall.
As a result, it makes it di cult to evaluate if the obtained estimates of frequency,
likelihood, or typicality are based on subjects’ subjective experiences or on a
biased sample of recalled information. 
According to Taylor, 1982, the latter possibility would render the availability
heuristic rather trivial, as one’s judgements are always based on what comes to
mind.
II. Experiment 1: If It Is So Di cult to Recall, It Cannot Be Typical.

A. Suppose that people are asked to report a certain number of examples of
particularly assertive, or of particularly unassertive behaviours that they have
recently engaged in. Presumably, reporting these behaviours would increase their
cognitive accessibility in memory, making it more likely that these behaviours come
to mind when the people are later asked to evaluate their own assertiveness.  
As a result, one should nd that people who had to report assertive behaviours
report higher assertiveness than people who had to report unassertive behaviours,
re ecting that the previously reported examples resulted in a biased sample of
relevant behaviours.  
As long as people consider only the content of what they recall, the more examples
they have to report, the more pronounced this e ect should be. Such content based
predictions may be derived from numerous models of self-related judgement.
B. Suppose that however, people not only rely on what comes to mind but also pay
attention to the subjective experiences that accompany the recall process. If so, the
subjective experience that it is very di cult to recall examples of one’s own
assertive behaviours may imply that one cannot be that assertive after all, or
thinking of examples would not be that di cult. This also means that to the extent
that the experienced di culty of recall increases with the number of examples that
are to be reported, ease of recall and content of recall would lead to di erent
conclusions. 
Whereas the content of recalled examples would suggest that one is very assertive
(or very unassertive), the di culty experienced in recalling these examples would
suggest that they cannot be frequent and typical. Hence, one may conclude that
15
fl
ffi
fi
ffi
ff
ffi
ffi
ffi
ffi
ff
ff
fi
fl
ff
one is probably not as assertive (or unassertive) as the recalled behaviours would
seem to imply. Accordingly, the experienced di culty of recall may qualify the
implications of recalled content.
C. Process 
Asked subjects to describe either 6 or 12 examples of very assertive or very
unassertive behaviours in which they had engaged. Pre-tests indicated that most
subjects could easily generate 8 or 9 behaviours but found it very di cult to
generate more than 10. Subsequently, subjects were asked to rate their own
assertiveness along several items. 
Hypothesis 
-If subjects base assessments of their own assertiveness solely on the relevant
behaviour that comes to mind, subjects who have to report examples of assertive
behaviour should rate themselves as more assertive than subjects who have to
report examples of unassertive behaviour. Additionally, the more examples subjects
have to report, the more pronounced the impact of recalled content should be. 
-Thus, a content-based judgement process will predict additive e ects of type of
example and number of examples requested. If subjects consider the content of
their recall in the light of the ease or di culty with which they can generate the
requested examples, however, these additive e ects should not be obtained. Rather,
the impact of recalled content should be less pronounced, the more examples
subjects have to report because the di culty of doing so should imply that the
recalled examples are not very frequent and typical. Hence, subjects should rate
themselves as less assertive (or unassertive) after recalling 12 rather than 6
examples, indicating that the implications of recalled content are quali ed by the
ease or di culty with which this content could be brought to mind.
D. Method 
40 female students at German University participated in a 2 (examples of assertive
vs. unassertive behaviour) x 2(6 vs. 12 examples) factorial between-subjects
experiment. Subjects were randomly assigned to conditions and tested in groups of
4.
E. Results 
-Manipulation Check: Subjects found it easier to report 6 rather than 12 examples.
No other e ect emerged. 
-Mean di erences: Subjects who had to describe examples of assertive behaviours
rated themselves as more assertive after describing 6 rather than 12 examples.
Conversely, subjects who had to describe examples of unassertive behaviours rated
themselves as less assertive after describing 6 rather than 12 examples. 
Crossover pattern was re ected in a marginally signi cant Valence x Number of
Examples Requested interaction, whereas neither of the main e ects reached
signi cance. 
What does this mean? Indicate that subjects did consider the experienced ease of
recall in evaluating their own assertiveness. In fact, subjects rated themselves as
more assertive after describing examples of assertive rather than unassertive
behaviours only if the recall task was easy. When the recall task was di cult, their
self-rating was opposite to the implications of recalled content, despite the fact that
more examples had been recalled. This pattern could not be accounted for on the
basis of recalled content per se, but instead re ected that the implications of
16
fi
ff
ffi
ff
fl
ffi
ffi
fl
ffi
ff
fi
ff
ff
ffi
fi
ffi
recalled content were quali ed by the ease with which the respective content could
be brought to mind. 
Correlational Analysis: The more di cult subjects who had to recall assertive
behaviours found the recall task, the lower assertiveness they reported. In contrast,
the more di cult subjects who had to report examples of unassertive behaviours
found the task, the higher the assertiveness they reported. 
Discussion: Findings suggest that the content of recall a ected self-judgements in
the direction of the valence of the recalled behaviours only if the recall process itself
was experienced as easy. If the recall process elicited experiences of di culty on
the other hand, the content of recall a ected self-judgements in a direction opposite
to the implications of the recalled behaviours. Hence we conclude that the
phenomenal experience of ease or di culty of recall may qualify the
implications of what comes to mind, even to the extent that the inferences
drawn are opposite in valence to the implications of recalled content. 
III. Experiment 2: An Extended Replication 

Designed to provide replication and extend previous study by manipulating the
perceived diagnosticity of experienced ease of retrieval. Subjects were again asked to
report 6 or 12 examples of assertive or unassertive behaviour. However, some subjects
were informed that most participants of a previous study found it easy to complete this
task, whereas other subjects were told that most previous participants found the task
di cult. We expected that subjects who were told that their subjective experience of
ease or di culty of recall was shared by most other subjects would be likely to attribute
their experience to characteristics of the task. If so, they might perceive their subjective
experience of ease or di culty as being less diagnostic and might therefore be less
likely to consider it when making a self-judgement. (If I tell you instances of you being
assertive is hard to recall and you nd it hard to recall too, you will most likely not report
lower rates of assertiveness because you attribute this di culty to the task rather than
yourself, hence low diagnosticity).  
In contrast, subjects who were told that their own experience of ease or di culty
contradicted the typical experience of similar others might be more likely to perceive the
frequency of the respective behaviours in their own repertoire. If so, they would be more
likely to rely on their subjective experiences of ease or di culty of retrieval in making
self-judgements. (High diagnosticity).
A. Method. 158 students (113 women and 45 men) of a West German teachers’
college were asked to report either 6 or 12 examples of assertive or unassertive
behaviour and were informed either that most previous participants had found it
easy or that most previous participants had found it di cult to complete this task.
The latter manipulation should increase the perceived diagnosticity of ease or
di culty of retrieval under conditions in which subjects’ own experience deviates
from the alleged experience of previous participants, that is, in which subjects nd
the task easy (6 examples) but are told others found it di cult or in which subjects
nd the task di cult (12 examples) but are told others found it easy. Conversely, this
info should decrease perceived diagnosticity under conditions in which subjects’
experience coincides with the alleged experience of most other participants.
B. Results. 
Manipulation Check. Indicated that subjects experienced recalling 6 examples as
17
fi
ffi
ffi
ffi
ffi
ffi
ffi
fi
fi
ffi
ff
ffi
ffi
ffi
ffi
ffi
ff
ffi
ffi
fi
easier than recalling 12 examples. No other signi cant e ect. 
Mean Di erences. Revealed a main e ect of sex, indicating that men reported higher
assertiveness than did women. However, sex of subjects did not interact with any of
the experimental variables and was therefore ignored in the following analyses.  
Present study replicated the previous ndings. Subjects who had to describe
examples of assertive behaviours rated themselves as more assertive after
describing 6 examples rather than 12 examples. Conversely, subjects who had to
describe examples of unassertive behaviour rated themselves as less assertive after
describing 6 examples rather than 12 examples.  
Contrary to predictions, the manipulation of the perceived diagnosticity of the
experienced ease of recall had no signi cant impact on subjects’ self assessment.  
Correlational Analyses. The interpretation that subjects’ self-assessments of
assertiveness were mediated by the subjective experience of ease of retrieval is
again supported by correlational analyses. Speci cally, subjects who had to report
examples of unassertive behaviour reported higher assertiveness the more di cult
they found the task.  
In contrast, subjects who had to report examples of assertive behaviour reported
lower assertiveness the more di cult they found the task, and both correlations
di er signi cantly from one another. 
Self Perception. Recalling many examples is di cult, the more examples subjects
are to report, the more the representativeness of the recalled examples may
decrease. Hence, subjects who have to recall 12 examples may eventually include
examples that are less extreme to complete their task. If so, a content-based
judgemental process may produce a similar pattern of ndings, re ecting that this
inclusion of less extreme examples may dilute the impact of more extreme ones. 
What the experiment found was although di erences in the quality of the
recalled behaviours did emerge in the unassertive examples conditions, these
di erences were opposite to those found in subjects’ judgements, lending no
support to the hypothesis that subjects’ judgements were mediated by
di erential content rather than by the subjective experience of ease of recall. 
Discussion. Experiment 2 replicated the previously obtained interaction, indicating
that implications of recalled content were quali ed by the ease with which this
content could be brought to mind. In contrast to our expectations, however,
informing subjects that most previous participants found the task easy or di cult,
respectively, did not result in a signi cant triple interaction, although separate
analyses under each diagnosticity condition provided some support for their
reasoning. 
This new diagnosticity condition might have been less relevant for subjects’ self
judgements than we had assumed a priori, whereas the diagnosticity manipulation
referred to others’ experience, the requested judgement was not a comparative one.
Individuals may use their own feelings and aspirations, rather than the behaviour of
others, in evaluating how assertive they are. If they can easily recall situations in
which they behaved unassertively, this may be bothersome no matter if others can
do so just as easily or not. Moreover, informing subjects about others’ experienced
ease or di culty of recall may have focused their attention even more on their own
phenomenal experience, thus increasing its’ impact. Accordingly, a di erent
manipulation of the diagnosticity of ease of retrieval is used in Experiment 3. 
18
ff
ff
ff
ff
ffi
fi
ffi
fi
ff
fi
fi
fi
ffi
ff
fi
fi
fi
ff
fl
ff
ffi
ffi
At the same time, failure to obtain a pronounced impact of the alleged typical recall
performance renders a possible variation of the ease of recall account less plausible.
One might argue that asking subjects to report 12 examples of a given class of
behaviours may convey that most people are probably able to do so. If so, the
experienced di culty in meeting this expectation might imply that one has less of
the respective trait than many other people. Accordingly, the obtained ndings
would re ect the experience of ease or di culty in meeting a certain standard,
rather than the experience of ease or di culty of recall per se. If so however, one
would expect a di culty with which others can perform the recall task. That
manipulation shows little e ect shows that an ease of meeting a standard account
less compelling.
IV. Experiment 3: Misattributing the Ease of Recall. 

Re ecting assumption that others’ performance may not be germane to the
informational value of subjective experiences as discussed above, a di erent
manipulation of perceived diagnosticity of ease of retrieval was used in Experiment 3.
As previous research on informational functions of another subjective experience,
namely mood, indicated, the diagnosticity of subjective states can be manipulated by
misattribution manipulations. Attributing one’s feelings correctly or incorrectly to a
transient source that is irrelevant to the judgement at hand was found to eliminate the
impact of a ective states on subsequent judgements. Similarly, misattributing one’s
arousal to an irrelevant source was found to eliminate the e ects of arousal states. 
Design rationale. Informed subjects that study was concerned with impact of di erent
types of music on the recall of autobiographical experiences. All subjects were exposed
by means of headphones to a piece of meditative music. Some subjects were informed
that this music facilitated the recall of situations in which one behaved assertively and
felt as ease, whereas others were informed that it facilitated the recall of situations in
which one behaved unassertively and felt insecure. After instructions, subjects had to
describe either 6 or 12 examples of assertive or unassertive behaviours, replicating the
previous experiments. 
In summary, discrediting the informative value of experienced ease of recall may result
in a data pattern that is opposite to the pattern observed in the previous experiments,
either within or between valence conditions, whereas the previously obtained pattern
should replicate when the informational value of subjects’ phenomenal experience is not
called into question.
A. Method. 78 female students of a West German University. 2 (examples of assertive

vs. unassertive behaviours) x 2(6 vs 12 examples) x 2(high vs. low diagnosticity of
experienced ease of recall) factorial between subjects experiment. Subjects
randomly assigned to conditions and received all instructions from cassette players
by means of headphones. 
Same as the previous ones, but to manipulate the perceived diagnosticity of
experienced ease of recall, subjects were told that the music was known to facilitate
the recall of autobiographical memories that pertain to experiences characterised by
assertiveness or by insecurity, respectively. If the music is said to facilitate the
respective recall task, this alleged side e ect should result in low diagnosticity of the
experienced ease of retrieval associated with recalling 6 examples but high
diagnosticity of the di culty associated with recalling 12 examples. Conversely, if
the music is said to facilitate recall of experiences that are opposite in valence to
those the person is asked to recall, this should imply high diagnosticity of easy recall
19
fl
fl
ff
ffi
ffi
ffi
ff
ffi
ff
ffi
ff
fi
ff
ff
but low diagnosticity of di cult recall. (E.g. participant asked to recall assertive
moments, but told that music would help to recall unassertive moments. If the
participant nds it easy to recall assertive moments despite this, it means that there
is high diagnosticity of recall, that their assertiveness is truly shining through despite
the blockade attempt by the music).
B. Results and Discussion. 

Manipulation check. Same. 
Mean di erences. Subjects self-ratings of assertiveness are the experienced ease of
retrieval, resulting in reversal between both valence conditions. Finally, such a
reversal was not obtained when subjects had to recall 12 examples, as suggested
by the reasoning outlined in the introduction to this study. 
When the alleged side e ects of the music did not discredit the implications of
subjects’ subjective experience, subjects did not rely on their subjective experience.
Rather, subjects who had to recall 6 examples of assertive behaviour reported lower
assertiveness than subjects who had to recall 6 examples of unassertive behaviour,
presumably re ecting that they discounted the experience of ease of retrieval. 
Correlational Analyses. Subjects who had to recall examples of unassertive
behaviours reported higher assertiveness the more di cult they found the recall
task, provided that the diagnosticity of ease of recall was not called into question.
When diagnosticity of ease of recall was discredited on the other hand, both
variables were uncorrelated as would be expected on the basis of the present
theorising, although the di erence between correlations did not reach signi cance.  
(E.g. subjects who were asked to report unassertive behaviour, who found it di cult
to recall unassertiveness reported themselves as being more assertive, provided
that the music that played was told to make them recall more unassertive
behaviour). 
Finally, no signi cant correlation of reported ease of recall and assertiveness
emerged under either high or low diagnosticity conditions for subjects who had to
report examples of assertive behaviour, in contrast to the previous experiments.  
 
V. Final Notes and General Discussion

A. Ease of Recall as Information. 
Reported ndings provide consistent support for assumption that implications of
recalled content may be quali ed by the ease or di culty with which content may be
quali ed by the ease or di culty with which that content can be brought to mind.
Although assumption has enjoyed great popularity since Tversky and Kahneman,
1973 introduction of the availability heuristic, it had not been adequately tested. 
In these series of experiments, conditions were constructed under which the
implication of what was recalled contradicted the implications of the subjectively
experienced ease or di culty with which it came to mind, disentangling the impact
of both sources of information. In 3 experiments, subjects attributed themselves
higher assertiveness after recalling 6 rather than 12 examples of assertive behaviour
and lower assertiveness after recalling 6 rather than 12 examples of unassertive
behaviour. If their judgement was solely based on content of what they
recalled, their self-attributions should have been more extreme the more
examples they recalled, in particular because content analyses of the reported
20
fi
ff
fi
fi
fl
fi
ffi
ff
ffi
ffi
ff
fi
ffi
ffi
fi
ffi
examples provided no evidence that the larger number of examples requested
decreased their representativeness. A judgemental process that is based on
recalled content as the only source of information cannot account for the
observed results. 
 
Rather, present ndings indicate that people paid attention to the subjective
experience of ease or di culty of recall in drawing inferences from recalled
content. Subjects concluded that they cannot be that assertive (or unassertive)
if it is very di cult to recall the requested number of examples. Ratings of ease
of recall were positively correlated with their self-assessment of assertiveness
if they had to report examples of assertive behaviour but were negatively
correlated if they had to report examples of unassertive behaviour. 
 
Most importantly, discrediting the experienced ease of recall by misattribution
manipulations (Experiment 3 music) reversed the otherwise obtained pattern
of ndings. In this case, subjects reported higher assertiveness after recalling
12 rather than 6 examples of assertive behaviour (when the insecurity music
was playing), and lower assertiveness after recalling 12 rather than 6 examples
of unassertive behaviour (when the assertive music was playing). 
 
Findings show that people not only consider what they recall in making a
judgement but also use the ease or di culty with which that content comes to
mind as an additional source of information. They only rely on the content of
their recall if its implications are not called into question by the di culty that
they experience in bringing the relevant material to mind. 
Di culty in recall may decrease judgements of frequency, probability, or
typicality, much as ease of recall has been assumed to increase these
judgements. 
 
B. Informative Functions of Subjective Experiences 

People may use their perceived a ective state as a source of information, according
to a “how do I feel about it” heuristic. In doing so, people misinterpret their pre-
existing a ective state as a reaction to the object of judgement, resulting in more
favourable evaluations under elated than under depressed moods, unless the
diagnosticity of their feelings for the judgement at hand is called into question. 
People were found to rely on their feelings at the time of judgement only under
conditions in which they could assume those feelings to re ect their a ective
reaction to the object of judgement. 
With regards to the availability heuristic, it re ects the correct insight that is
easier to recall frequent rather than rare events. What renders this heuristic
error prone is that the experienced ease of retrieval may re ect the impact of
variables other than frequency, such as salience or vividness. Hence we may
conclude that inappropriate applications of the availability heuristic re ect a
21
ffi
fi
ff
ffi
fi
ffi
ff
ffi
fl
fl
fl
ffi
fl
ff
process of misattribution, people rely on their subjective experiences of ease
of retrieval to the extent that they misattribute it to frequency of occurrence,
rather than to the impact of other variables. 
I. PUTTING ADJUSTMENT BACK IN THE ANCHORING AND ADJUSTMENT

HEURISTIC: Di erential Processing of Self-Generated and Experimenter, Nicholas
Epley and Thomas Gilovich. (2001).
A. Abstract. People’s estimates of uncertain quantities are commonly in uenced by
irrelevant values. Anchoring e ects were originally explained as insu cient
adjustment away from an initial anchor value. The existing literature provides little
support for the postulated process of adjustment however, and a consensus that
none takes place seems to be emerging. This conclusion may be premature, and
evidence will be presented that insu cient adjustment produces anchoring e ects
when the anchors are self-generated. 
Study 1: Participants’ verbal reports made reference to adjustment only from self-
generated anchors. 
Study 2 & 3: Participants induced to accept values (by nodding heads) gave
answers that were closer to an anchor (i.e. they adjusted less) than participants
induced to deny values (by shaking heads), again, only when the anchor was self-
generated. 
Results: It may be time to reintroduce anchoring and adjustment as an explanation
for some judgements under uncertainty. 
 
Anchoring E ects have traditionally been interpreted as a result of insu cient
adjustment from an irrelevant value, but recent evidence casts doubt on this
account. Instead, anchoring e ects observed in the standard paradigm appear
to be produced by the increased accessibility of anchor-consistent
information. Some researchers conclude that anchoring occurs because of
biased retrieval of target features and not because of insu cient adjustment. 
 
This experiment aims to show that this conclusion is premature. Just as memory
research was sidetracked by an overly persistent analysis on people’s ability to
recall nonsense syllables, so too has anchoring research been sidetracked by an
overly persistent analysis of people’s responses in the standard anchoring
paradigm. Outside this paradigm, anchors are often self-generated, rather than
provided by an experimenter or other external sources.
B. STUDY 1
1. Participants verbalised their thoughts when answering questions involving self-
generated and experimenter provided anchors. Predicted that participants would
describe a process of anchoring and adjustment only when anchors were self-
generated. In these cases, we expected that the verbal reports would typically
begin with a reference to the anchor value, followed by a statement describing
adjustment away from it (e.g. The US declared its independence in 1776 and it
probably took a few years to elect a president, so Washington was elected in…
1779).  
In contrast, we expected experimenter provided anchors to produce little to no
22
ff
ff
ff
ff
ffi
ffi
ffi
fl
ffi
ff
mention of either the anchor or adjustment, consistent with selective-
accessibility account of anchoring e ects in the standard paradigm.
2. Method. 
50 Cornell Undergraduates were each asked 4 questions. 2 questions were ones
for which most participants could be counted on to generate a particular anchor
value (e.g. When did the 2nd European explorer, after Columbus, land in the
West Indies?—1492), and 2 involved anchors provided by experimenter (1 high
and 1 low value). 
Participants were asked to explain how they arrived at the answer to each
question.  
Participants were considered to have utilised anchoring and adjustment only if
their verbal reports referred to both the anchor and a process of adjustment.
3. Results and Discussion. 

Participants were more likely to describe a process of anchoring and adjustment
when the anchor values were self-generated than when they were provided by
the experimenter. Of those participants who appeared to know both self-
generated anchors, 94% made reference to anchoring and adjustment in
response to at least one of the self-generated items, and 65% did so in
response to both.  
In contrast, only 22% of the participants described anchoring and adjustment in
response to at least one of the experimenter provided anchors, and only 4% did
so in response to both. 
People were far more likely to report using anchoring and adjustment when
considering self-generated anchors, than when considering experimenter
provided anchors. 
Results indicate that self-generated anchors activate di erent mental
processes than experimenter provided anchors. One might be concerned
however, about relying on participants’ self-reports given the widespread
doubts about whether people can accurately report on their own mental
processes. 
Participants may have been also less likely to mention the experimenter
provided anchor value and how they adjusted from it because the anchor
value was already mentioned in the initial comparative question.
C. STUDY 2
1. When people adjust from self-generated anchors, they may do so in either 2
ways. 1 is that people slide along some mental scale, continuously testing until
they arrive at a satisfactory nal estimate. 2nd, is that they jump some amount
from the anchor, analogous to a saccade in reading, to a more reasonable value
and assess its plausibility. If the new value seems plausible, adjustment stops. If
it does not seem plausible, new jumps or saccades are made and the new value
is assessed and so on. 
Asking participants to nod their heads would make them more willing to accept
values that initially came to mind, and thus produce less adjustment from self-
generated anchors. Shaking their heads from side to side in contrast would
make participants more willing to reject values and thus produce more
adjustment from self-generated anchors. We also predicted that participants
23
fi
ff
ff
would generate an answer more quickly when nodding than when shaking their
heads.
2. Method. 
50 participants. Told that experiment was a study of product evaluations, and
that they would be asked to evaluate a set of headphones while moving their
heads from side to side or up and down in order to assess the headphones
under everyday use. All participants listened to a tape containing 16 anchoring
questions. To justify this procedure and minimise suspicion, experimenter
explained that she wished to examine implicit evaluations that people form
without conscious intention or e ort. Thus had to busy participants with another
task while they were evaluating the headphones, in this case by answering the
questions on the tape. Depending on random schedule, participants were then
asked to nod their head up and down, shake their head from side to side, or
hold their head still.
3. Results and Discussion. 

2 preconditions had to be met for an adequate test of the hypotheses about self-
generated anchors. First, participants had to know the self-generated anchor.
Second, they had to report considering the anchor when making their estimate.
Participants who did not meet these preconditions were excluded on an item by
item basis. On 3 questions, fewer than 30% of participants met both
preconditions, generally because they did not know the intended anchor value.
In some cases, this left no participants in one or more of the experimental
conditions. 
Participants head movements signi cantly in uenced their answers to the
items with self-generated anchors.  
Follow up contrast showed that participants provided answers closer to the
self-generated anchor when they were nodding their heads than when they
were shaking their heads.  
Participants gave responses in between those in these 2 conditions when
they were holding their heads still.  
Participants’ head movements also in uenced the speed with which they
generated their answers to the questions with self-generated anchors, and
they answered more quickly when nodding than when shaking their heads. 
D. STUDY 3
1. Used equal number of items with self-generated and experimenter provided
anchors, and counterbalanced the order in which these items were presented.
This allowed a direct statistical test of di erential e ect of head movements.
2. Results and Discussion 

Participants’ answers were more discrepant from a self-generated anchor
when they were shaking versus nodding their heads. 
Head movements did not in uence responses to items with experimenter
provided anchors. 
Participants were somewhat faster to provide answers to the items with
self-generated anchors when they were nodding their heads than when
they were shaking them. 
24
fl
ff
fi
ff
fl
fl
ff
Head movements had no in uence on reaction times to questions with
experimenter provided anchors. 
3. FINAL GENERAL DISCUSSION 

 
This research provides compelling evidence that anchoring e ects can be
produced by a process that includes adjustment. 
 
When questions activate self-generated anchors, people adjust from those
anchors to arrive at nal estimates. This process di ers considerably from
the processes involved when anchors are provided by an external source,
demonstrating that there are distinct anchoring e ects produced by
di erent mechanisms. 
 
Adjustments in numerical estimates also tend to be insu cient. People
tend to systematically fall short of the actual answer when adjusting from
self-generated anchors.

Ease of Retrieval E ects in Persuasion: A Self Validation Analysis. Zakary L.
Tormala. (2004)
I. Experiment Background 
-People sometimes base judgements on the subjective experience of ease with which
certain information comes to mind has received considerable attention recently. 
-Schwarz et al. 1991, found that people experienced greater di culty in generating 12
rather than 6 examples of assertive behaviours, and presumable interpreted this
di culty as indicating that assertive behaviours were low in frequency or likelihood, thus
inferring that they must not be very assertive. 
Retrieval e ect in play, sparking new research such as ease of retrieval impacting
probability estimates, stereotyping, ingroup vs. outgroup judgements, health risk
assessments, attitudes, and attitude strength. 
-In most work on ease of retrieval, e ects have been assumed to operate according to
the availability heuristic. It has been argued that subjective experience of ease or
di culty a ects inferences about the amount of information available, which in turn
25
ffi
ffi
ff
ff
ff
fi
ff
fl
ff
ff
ff
ffi
ffi
ff
serves as a judgemental heuristic. 
-This present research examines an alternative mechanism to account for ease of
retrieval e ects. We agree that when an individual is asked to generate a given
number of thoughts, they can experience greater di culty to the extent that many
thoughts are required. 
We di er by arguing that the ease or di culty experienced can impact the
con dence the individual has in their thoughts and that this con dence in
thoughts determine the extent to which people rely on them. 
The easier it is to generate a list of thoughts (due to low number required), the
more con dence an individual should have in these thoughts. The more di cult it
is to generate a list of thoughts (due to high number required), the less con dence
an individual should have in them. 
The feeling of con dence stems not from the perceived number of thoughts
available (e.g. when I ask you to name all cat breeds) but rather, more directly from
the feeling of ease itself. 
This experiment aims to provide the rst meditational evidence for the role of
thought con dence in ease of retrieval e ects. 
II. Possible Moderation of Ease of Retrieval E ects 

Availability heuristic and thought con dence explanations make di erent predictions as
to when ease of retrieval e ects should occur.  
According to previous research, ease of retrieval e ects should be most powerful when
conditions limit people’s motivation or ability to think about issue relevant information.  
However, according to the thought con dence explanation, however, ease of retrieval
e ects should be more likely to to occur when people’s motivation and ability to
process are relatively high.  
Assessments of thought con dence are more impactful under high, rather than low
elaboration conditions 
E.g. unless someone was actively generating thoughts, assessments of these thoughts,
such as how easy or di cult it was to generate them, should not matter.  
Under high elaboration conditions, people have increased motivation or ability to
engage in this higher order kind of processing, attending to their thoughts and gauging
their feelings about them.  
Petty et al. (2002). Measured and manipulated the con dence participants had in their
own cognitive responses to a message. Increased con dence in thoughts was
associated with increased persuasion when the thoughts were favourable, but
increased con dence in thoughts was associated with decreased persuasion when
thoughts were unfavourable. Furthermore, these e ects were con ned to situations in
which participants were likely to be high in their extent of thinking. This suggests that
people are more attuned to subjective assessments of their own thoughts when
they were processing relatively extensively. Therefore, if ease of retrieval a ects
con dents in thoughts, its impact on attitude change should be greatest under
high-elaboration conditions.
III. Empirical Evidence for Moderation of Ease of Retrieval E ects
A. Evidence Favouring Low Elaboration Conditions. 
Rothman & Schwarz (1998) Study assuming that ease of retrieval e ects is most
26
ff
fi
fi
ff
fi
ff
fi
fi
fi
ffi
ff
fi
fi
fi
fi
ffi
ff
ff
ff
ff
fi
fi
ffi
ff
fi
ff
ff
fi
ffi
ff
fi
prominent under low elaboration conditions (operationalised as low personal
relevance). They included low, moderate, to high relevance conditions as part of the
experimental design. 
However, we are suggesting another explanation for these results. Rather than
eliminate the low relevance conditions, it may be more appropriate to eliminate the
high-relevance condition and focus on the comparison of low with moderate
relevance. This is because one of the personal relevance variables Rothman &
Schwarz (1998) examined, family history, was an individual di erence variable.
Individuals with a family history of heart disease are likely to be more acutely aware
of their own personal risk and possibly have more knowledge about risk factors than
those without a family history. Consequently, those with family histories could have
been more adept at generating a list of risk factors. 
If the high-relevance participants are eliminated and the focus of comparison is on
moderate versus low relevance participants, then the direction of e ect is that
increasing elaboration likelihood increases the ease of retrieval e ect, consistent
with the self validation hypothesis
B. Evidence Favouring High Elaboration Conditions. 

Wänke and Bless (2000) departed from prior research and suggested that ease
e ects should be more likely to occur when processing motivation is relatively high.
Participants were presented with a strong or weak persuasive message containing
information about a co ee maker and asked them to recall information from the
message. They found that ease of retrieval enhanced persuasion when accuracy
motivation was high but not when it was low and that this occurred in both the
strong and weak message conditions. 
However, we disagree that ease invariably enhances persuasion. If weak arguments
are those that elicit primarily unfavourable thoughts, then easy retrieval should
reduce persuasion. 
Second concern is that Wänke and Bless’s experiment is that their manipulation of
elaboration came after the persuasive message had been presented. Thus, this
manipulation likely a ected the e ort participants put into recalling information from
the message but could not have a ected their extent of processing the message or
the attitude object during message presentation. 
They also used an unusual manipulation of ease of retrieval, asking participants to
recall information from a persuasive message using retrieval cues that were either
helpful (easy) or unhelpful (di cult). This was okay but it would have been better to
show that high elaboration participants relied on subjective experience instead of
the actual amount of information, as is customary in the ease of retrieval literature. 
IV. Overview of Present Research 

Most of the previous literature on ease of retrieval clearly suggests that subjective
experience should function as a heuristic in the judgemental process when cognitive
processing is relatively low. However, relatively little empirical evidence has been
gathered to examine this notion. Attempts at clarifying the role of elaboration have
produced ambiguous ndings. 
In our 3 studies, participants were asked to generate either a small or large number of
favourable or unfavourable thoughts in response to a persuasive message and then to
report their attitudes toward the issue raised in the message. Expected that high
27
ff
ff
fi
ff
ffi
ff
ff
ff
ff
ff
elaboration participants would be more in uenced by the subjective experience of
generating these thoughts (e.g. favouring the issue more after listing 2 than 10
favourable thoughts), whereas low elaboration participants would be relatively more
in uenced by the actual number they generated (e.g. favouring the issue more after
listing 10 than 2 thoughts). 
In Study 3, we sought to provide evidence for the meditation of ease of retrieval e ects.
We predicted that ease e ects would be mediated by the con dence individuals had in
the thoughts they listed. Attitudes would be more consistent with thoughts when
participants were con dent in them, when they generated an easy (small) number rather
than a di cult (large) number of thoughts. 
A. STUDY 1 
Used a stable individual di erent assessment of elaboration. The need of cognition
(NC), referring to the tendency to engage in and enjoy e ortful thought. Individuals
high in NC consistently have been found to engage in greater elaboration of
persuasive messages than those low in NC.  
Participants were asked to generate counterarguments to a persuasive
communication. Ease of retrieval e ect would be present to the extent that
individuals opposed the persuasive message more after generating a low number of
counterarguments than a high number of counterarguments. 
For low NC individuals, we expected a heuristic or cue e ect, whereby they would
oppose the message more after generating many rather than few counterarguments. 
 
Method 
59 undergraduates from Ohio State University participated in a partial ful lment of a
course requirement in their introductory psychology course. Participants completed
the NC scale and were randomly assigned to generate either 2 or 8
counterarguments against a persuasive message. 
Participants read that a specially appointed committee at their university had
recently submitted a proposal to implement senior comprehensive exams as a
graduation requirement beginning in the next 2 years. Failure to pass these exams
they were told, would result in remedial work to be completed before a degree could
be conferred. Participants were led to believe that before implementing this policy,
their Board of Trustees wanted to assess students’ reactions and so they would be
asked to read about the policy and answer several questions. After reading, these
participants were presented with 4 arguments in favour of implementing the senior
comprehensive exam policy. After reading this message, participants listed a high or
low number of arguments against it and completed attitude items and the NC scale. 
Independent variables included number of counterarguments, and need for
cognition. 
 
Results and Discussion 
High NC individuals opposed the exams more after listing 2 rather than 8
arguments against the exams, and Low NC individuals demonstrated attitudes
in the reverse direction, opposing the exams more after listing 8 rather than 2
counterarguments. 
This could be because there were di erences in the quality of
28
fl
ffi
fi
ff
ff
ff
ff
fl
ff
ff
fi
fi
ff
counterarguments listed across the 2 and 8 arguments conditions. Participants
might have generated counterarguments that were lower in quality when they
were required to generate many of them. Consistent with previous research on
the elaboration likelihood model, high NCs would be expected to be
particularly sensitive to these quality di erences, whereas low NCs would be
expected to rely more on the number of arguments. 
High NCs appeared to base their attitudes more on the subjective experience
of argument generation, whereas individuals relatively lower in NC did not. 
Also found that ease is not inevitably translated into more favourable attitudes
but can produce less favourable attitudes when thoughts are negative (e.g.
counterarguments). This reverse direction of means for low NCs means that
these individuals based their attitudes on the actual number of
counterarguments generated rather than subjective experiences associated
with them, using the number of counterarguments as a heuristic or simple
peripheral cue. 
B. STUDY 2 
Replicating and extending the ndings of rst study, 2 primary changes were made.
1st, because we argued that extent of thinking plays an important role in the ease of
retrieval e ect, we manipulated it to permit more causal conclusions in this regard.
2nd, we felt it was also important to demonstrate that the ndings from Study 1
were not limited to the precise nature or content of the thoughts we asked
participants to generate. Therefore, participants in Study 2 were asked to list
positive thoughts in response to the message. 
In this study, we expected that individuals in high elaboration conditions would
demonstrate the ease e ect, such that they would have more favourable attitudes
after generating a few rather than many positive thoughts. Individuals in low
elaboration conditions on the other hand, were expected to demonstrate the reverse
e ect for the number of thoughts listed. (No more NCs). 
 
Participants in the high-elaboration condition reported thinking more deeply
than those in low-elaboration condition. 
Individuals in high elaboration conditions favoured the exam policy
signi cantly more after listing 2 instead of 10. On the other hand, individuals in
low elaboration conditions tended to favour the exams more after listing 10
instead of 2 positive thoughts. 
High elaboration participants were sensitive to di erences in the quality of
thoughts listed across the 2 and 10 thoughts conditions, whereas low
elaboration participants were not.  
High elaboration participants relied more on their subjective experience of
listing their thoughts, reporting attitudes that were more in line with their
thoughts after listing a small rather than large number of them. 
Low elaboration participants relied more on the actual number of thoughts
listed, reporting attitudes that were more in line with these thoughts (not
29
ff
fi
ff
ff
fi
fi
ff
ff
fi
necessarily their own POV) after listing a high rather than low number. 
C. STUDY 3 
Provide evidence that ease e ects in persuasion were mediated by thought
con dence under high elaboration circumstances. Predicted that for participants
who were motivated to think, generating a high number of positive thoughts would
be di cult and the e ect of this di culty on attitudes would be mediated by the
con dence participants had in the thoughts they listed (having less con dence when
it is more di cult to generate the thoughts). Also expect this to rule out availability
heuristic interpretation of ease e ects. This means that we expected that
participants’ estimates of the number of positive thoughts they actually had (i.e. how
many were available) would not mediate the ease e ect.  
Procedure Measures 
Attitudes, immediately following pro-exam message and thought listing task,
attitudes toward exam policy were assessed using the same 4 items as in the rst 2
studies. 
 
Con dence in thoughts. To assess thought con dence, participants were asked how
much con dence did they have in the positive thoughts they listed. 
 
Perceived di culty. A single question was used to assess di culty participants
experienced generating the requested number of thoughts. “How di cult did you
nd it to generate the required number of positive thoughts?” 
 
Estimate number of thoughts. One item assessed the number of positive thoughts
participants believed they had toward comprehensive exams. “Type in the actual
number of positive thoughts you would estimate you had on your own.”. 
 
 
Number of positive thoughts they were requested to generate had a
signi cantly direct e ect on perceived di culty, participants asked to generate
10 thoughts found it more di cult than participants asked to generate 2
thoughts.  
Perceived di culty had an inverse e ect on con dence in thoughts such that
the more di cult participants perceived the thought generation task to be, the
less con dence they had in their thoughts. 
Thought con dence had a direct e ect on attitudes, which is to be expected
when thoughts are favourable. This pattern supports the self-validation
perspective.  
Participants in the 10 thoughts conditions thought they had more positive
thoughts in general than participants in the 2 thoughts condition. 
Perceived di culty also had a marginal negative e ect on perceived number of
thoughts, consistent with the availability notion that people infer number from
ease of retrieval, but perceived number of thoughts did not ultimately a ect
attitudes. 
30
fi
fi
fi
fi
ffi
fi
fi
fi
ffi
ffi
ffi
ffi
ffi
fi
ff
ff
ff
ffi
ff
ffi
ff
ff
ffi
fi
ff
fi
ff
ffi
ffi
fi
ff
fi
In short, participants assigned to high elaboration conditions experienced
greater di culty in generating 10 than 2 positive thoughts. Impact of perceived
di culty on attitude ratings was mediated by the con dence people had in the
thoughts they generated.  
When it is easy to generate the thoughts (only 2 were requested), participants
had more con dence in their thoughts and thus formed attitudes that were
more consistent with them and their on POV (i.e. favourable). When it was
more di cult to generate the thoughts (because 10 were requested), thought
con dence decreased, making it less likely that they would form attitudes that
were consistent with them.  
So previous work was correct in concluding that generating high number of
thoughts were perceived as more di cult than a lower number. However in
alignment with the self-validation hypothesis, the e ect of this di culty on
attitudes was mediated by the con dence individuals reported having in the
thoughts they generated, not participants’ inferences about how many positive
thoughts they must have had. 
 
Discussion and Implications 
Subjective experiences associated with thoughts are taken into account
primarily under high-elaboration circumstances. Primarily under high-
elaboration circumstances, individuals have the motivation to attend to their
cognitive phenomenology. 
Ease of retrieval paradigm experiment showed that high elaboration
participants opposed a new policy more after generating few rather than many
arguments against a message supporting the policy, whereas low elaboration
participants showed the opposite e ect. 
Experienced ease of generating arguments can decrease persuasion under
high elaboration conditions. 
Study 2 showed that high elaboration participants favoured a new policy more
after generating few rather than many positive thoughts in response to a
message supporting the policy, and low elaboration participants tend to show
opposite results, reporting more thought congruent attitudes after generating
a high number of such thoughts, treating NUMBER as a SIMPLE CUE. 
 
Applications? 
Dijksterhuis et al. (1999). Stereotyping can be in uenced by ease of retrieval of
stereotypic traits of an out-group, but only for individuals low in prejudice. 
For high prejudiced individuals, it is easy to recall both a low and high number
of stereotypical traits, and their unfavourable attitudes toward the out-group
are too strong to be in uenced by momentary and extraneous factors. 
Petty, Fleming and White (1999). People low in prejudice are also are more
motivated to thoughtfully process information from a stigmatised source to
prevent biases from in uencing their judgements, and it is possible that for the
Dijksterhuis et al. experiment, low prejudiced individuals were motivated to
31
ffi
fi
ffi
ffi
fi
fl
fl
fi
ff
ffi
fl
ff
fi
ffi
process information carefully to make unbiased judgements. This heightened
level of elaboration might have turned some of their attention to their
subjective experience, increasing or decreasing their con dence in the
stereotypic traits they were listing, and a ecting the stereo-typicality of their
judgements accordingly. 
 
Conclusions 
Aspects of subjective experience (e.g. ease of retrieval) can in uence the
con dence people have in their thoughts, at least for individuals attuned to
that experience (e.g. those under high elaboration). Ease of Retrieval is more
likely to occur under high thinking due to thought con dence. 
Traditional ease of retrieval e ect may not be heuristic in nature at all. Instead,
subjective experience of ease or di culty can have a more complex e ect on
judgements through individuals’ personal assessments of the con dence they
have in their own thinking and their subsequent willingness to rely on this
thinking in forming judgements. 
Under extremely high elaboration conditions, people might consider both
subjective experience of ease and the actual number/quality of thoughts
generated before making their nal judgement. 
I. Elaboration and Consequences of anchored estimates, An attitudinal perspective on

numerical anchoring. Blankenship, K. L., Wegener, D. T., Petty, R. E., Detweiler-Bedell,
B., & Macy, C. L. (2008). [TUTORIAL 1 - Week 4] 
 
Abstract 
Theories of attitude change suggest that the same judgements can result from relatively
thoughtful or non-thoughtful processes, with more thoughtful processing resulting in
judgements that last longer over time and better resist future attempts at change. 4
studies manipulated participants’ level of cognitive load to produce relatively high vs.
low levels of thinking guided by an attitudinal approach to anchoring. 
 
Anchoring can occur under both high and low cognitive load conditions,
anchoring based on a higher level of thinking involves greater use of judgement,
relevant background knowledge, persists longer over time, is more resistant to
subsequent attempts at social in uence, and is less likely to result from direct
numeric priming. 
 
Purpose of Article and why this line of research is important 
 
Processes and consequences of anchoring and attitude change are rather similar. 
 
Purpose of this article is to examine the implications of a general attitudinal approach to
conceptualising anchoring e ects.  
 
This line of research is important because it is possible that the dominant view of
32
fi
ff
ff
fi
fl
ffi
ff
fi
fi
fl
fi
ff
anchoring is incomplete and that anchors can serve in multiple roles (e.g. serve as a
simple cue to in uence judgement, or serve to bias e ortful processing). 
 
Experiment 1 
 
Variables 
Anchor: between subjects 2 levels (high/low), only for the 4 critical targets which
appeared in the same order for all participants, participants received high (low) anchors,
e.g. Neil Armstrong was 48 (23 years old) when he walked the moon. 
Cognitive load: Between subjects factor with 2 levels (high/low).  
Background Information: Within subjects factor with 2 levels (anchor consistent/
anchor inconsistent). E.g. participants who received high anchors were given 2
background knowledge paragraphs that suggested the respective answers were also
high (anchor consistent) in addition to the other 2 paragraphs that suggested the
respective answers were low (anchor-inconsistent) 
 
Procedure 
Participants read 8 background knowledge passages (4 related to subsequent
anchoring task, while other 4 were unrelated ller). 
Engaged in unrelated impression formation task 
Completed series of MCQs to assess comprehension ability 
Engaged in standard anchoring task (4 ller items appeared rst followed by target
items in an alternating sequence (i.e. LHLH order or HLHL order). 
 
Results  
Anchoring occurred regardless of level of cognitive load. 
Low cognitive load, the e ect of anchors are greatest when background
knowledge is consistent with anchor. 
High cognitive load, the e ect of background knowledge have little e ect on
estimates. 
Participants think more deeply and used background knowledge to a greater
extent when cognitive load was low rather than high. 
Numerical anchors (like elaboration in attitude change) can serve multiple roles in
producing judgement.). 
 
Experiment 2 
Do anchored estimates have longer lasting impact when they are based on relatively
high rather than low levels of elaboration (thinking)? 
 
Variables 
Anchors: Similar to experiment 1, but either HHLLHHLL order or LLHHLLHH order. 
Cognitive Load: Similar to experiment 1, but either high cognitive load during rst 4
items or during last 4 items. 
Time: Results taken immediately or after 1 week delay. 
 
Procedure 
33
fl
ff
ff
fi
fi
ff
fi
ff
fi
Same type of standard anchoring task in Experiment 1. One week later returned to the
lab to make the same estimates (without the anchors). 
 
Results 
No di erence in initial anchoring as a function of cognitive load manipulation. 
Estimates were more in uenced by anchors when initial anchored judgements
were made under a low level of cognitive load. 
Anchored judgements persist longer when they are based on relatively thoughtful
rather than non-thoughtful processes. 
Anchored perceptions can last longer over time if they were formulated in
thoughtful ways (space given, self-generated meanings) rather than non-
thoughtful ways. Support for the attitudinal view of anchoring. 
 
Experiment 3 
Do anchored judgements better resist social in uence when formed under conditions of
low rather than high cognitive load? 
 
Variables 
Same as Experiment 2, but Time was replaced by Judgement. 
 
Procedure 
Similar procedure to Experiment 2. Di erences included that participants did not return
for a second session, and after participants completed the ller task after the initial
anchoring task, their initial estimates were attacked, as they were told that previous
participants had provided estimates that were quite di erent. After which participants
were given the chance to provide another estimate of target value. 
 
Results 
Anchoring remained stronger (i.e. resisted attack to a greater extent) if anchored
judgements initially occurred with low cognitive load. 
Anchored estimates better resist attempts at social in uence if the anchored
estimates were formed on the basis of high, rather than low-elaboration
processes. 
Thoughtful anchors are more resistant to future attempts at change as compared
to less thoughtful anchors, consistent with previous evidence of multiple roles in
which thoughtful e ects of stereotypes on judgements were more resistant to
social in uence than non-thoughtful e ects of stereotypes on judgements. 
 
Experiment 4 
Is numeric priming a relatively non-thoughtful process that is likely to create anchoring
e ects when the ability to think about target judgements is relatively low rather than
high? 
 
Variables 
Anchor and Cognitive Load manipulations used in previous experiments were also used
34
ff
ff
fl
ff
fl
ff
ff
fl
ff
fl
fi
here. Anchors were counterbalanced across participants (i.e. HLHLHLHL or
LHLHLHLH), Cognitive Load was either high or low (secondary task present or absent). 
 
Procedure 
Participants engaged in a similar anchoring paradigm that was used in previous
experiments. They were rst asked a comparative question (i.e. anchor), and then asked
to provide an estimate of a real target value. All comparative questions were unrelated
to the target. 
 
Results 
Under low cognitive load, no anchoring e ects were found. 
Under high cognitive load, signi cant anchoring e ects were found. 
Numeric priming e ects are stronger when ability to think carefully about the
target judgement is decreased (i.e. moderation). 
The results from this experiment can help explain why certain studies have found
robust numeric priming e ects while others have found weak numeric priming
e ects.

Determining What is, What was, and What Might Have Been: Hypothesis Testing,
Covariation Detection, and Counterfactual Thinking. 
Kunda, Z. (1999). Social Cognition: Making sense of People. MIT Press: London,
England.
I. Positive Test Strategy 

Klayman and Ha (1987). This is a one sided approach to hypothesis testing.
When you use this, you test a hypothesis by seeking cases that match it. To
determine whether your friend is extraverted, you search for examples of her
extraversion. If you can nd such evidence, you conclude that the hypothesis is
true (she is indeed, extraverted). If you cannot nd such evidence, you conclude
35
ff
ff
fi
fi
ff
fi
ff
fi
ff
the hypothesis is false. 
This serves us well in cases where evidence is clear cut, like if your friend is
uniformly extraverted or uniformly introverted, it will not matter that your search
for evidence is one sided, even a more balanced search would yield identical
results. 
But if the relevant evidence is mixed, the positive test strategy would bias your
evaluations. 
People utilise the positive test strategy to evaluate hypotheses both when
searching through their memories for preexisting knowledge and when
searching the external world for new evidence that could bear on their
hypotheses. This is likely to lead to hypothesis con rmation when applied in the
social domain because people’s social behaviour varies from one situation to
another.
II. Hypothesis Con rmation in Evaluating Others
A. Biased Memory Search 

People often nd themselves confronted with new questions about people
we know well. (E.g. will my girlfriend make a good roommate? Will my child
thrive in the neighbourhood school? Will my brother succeed as a lawyer?). 
Mark Snyder & Nancy Cantor (1979). Wanted to determine whether people
test such hypotheses about familiar others by conducting one-sided
searches through their relevant memories, searching for those memories that
t their hypotheses. They rst provided participants a set of mixed memories
about a person, Jane. They read a lengthy description of a week in Jane’s
life, which portrayed her as behaving in an extraverted manner on some
occasions and in an introverted manner on others. (E.g. she engaged in
animated conversation at the doctor’s but refrained from socialising during
her co ee break, spoke to strangers while jogging but acted shyly at the
supermarket). When participants returned to the lab 2 days later, they were
asked to determine Jane’s suitability for the job of real estate agent. This is
was explicitly stated, but rather implied, that the person needed to have an
extraverted personality. The remaining half were asked to determine Jane’s
suitability for the job of librarian, which is typically stereotyped as requiring
an introverted personality. 
Results 
People were engaged in positive test strategy.  
Those testing Jane’s suitability to the extraverted profession of real
estate recalled more extraverted than introverted facts about her. 
Those testing Jane’s suitability to the introverted profession of librarian
recalled more introverted than extraverted facts about her. 
36
fi
ff
fi
fi
fi
fi
People appeared to be asking themselves “what do I know about Jane
that is consistent with her being suitable for this job?”, and not “what
do I know that is inconsistent?”. 
Biased recruitment of relevant facts about Jane resulted in hypothesis
con rmation. Those testing Jane’s suitability for real-estate agent
thought she would make a better real-estate agent than librarian. Those
testing Jane’s suitability for librarian thought she would make a better
librarian than real-estate agent. When we have mixed evidence about a
person, merely entertaining a one-sided hypothesis about this person
can increase our con dence in this hypothesis.
B. Biased Evidence Seeking 

People often nd ourselves testing one-sided hypotheses about newly
encountered, unfamiliar others. These can be derived from prior expectations
(John’s teachers say he is pretty smart, isn’t he?), from stereotypes (lawyers
tend to be aggressive, is this lawyer aggressive?), or from our goals (I want to
hire a warm and caring person as a baby-sitter, is this person I am
interviewing warm and caring?). 
Our reliance on the positive test strategy when seeking the evidence we
need to evaluate such hypotheses about others will often bias us toward
hypothesis con rmation, because people’s behaviour on any personality
dimension is rarely uniform. 
Mark Snyder, William Swann et al. (1978). Each participants was asked to
evaluate the personality of a stranger waiting in another room. Half were told
their job was to nd out whether this person’s behaviour and experiences
matched those of the typical extravert. They were given a brief pro le of
extraverts, which included attributes such as outgoing, con dent, and
enthusiastic. The remaining half were told to nd out whether this person
matched the pro le of the typical introvert, which included attributes such as
shy, quiet, and retiring. All participants were given 26 questions from which
they were to choose 12 that they would pose to the other person so as to
nd out whether that person matched the pro le. List included 2 kinds of 1
sided questions, 11 extraverted questions asked for examples of extraverted
behaviour, and 11 introverted questions asked for examples of introverted
behaviour, the remaining 5 questions were neutral (e.g. what are your career
goals). 
Participants favoured questions that matched the hypothesis they were
asked to test, those assessing whether the person was extraverted chose
more extraverted than introverted questions. But those assessing introverted
person showed preference for more introverted than extraverted questions.
Participants’ choice of questions re ected reliance on the positive test
37
fi
fi
fi
fi
fi
fi
fi
fl
fi
fi
fi
fi
strategy. 
Snyder and Swann reasoned that a list that includes mostly extraverted
questions gives respondents many opportunities to reveal their
extraverted side, but few opportunities to reveal their introverted side.
Respondents might convey an image of themselves that is biased
toward extraversion. Similarly, the converse is true for the introverted
question-dominated set.  
Trope and Bassok (1982, 1983). Pointed out that Snyder and Swann’s
studies were not given an opportunity to ask truly diagnostic questions,
e.g. questions that could actually reveal whether or not someone was
extraverted. When one is asked “What would you do to liven up a
party?”, one is essentially compelled to respond as an extravert, the
only way one could reveal introversion would be to reject the
assumption that one even livens up parties in the rst place, something
that is di cult to do in polite conversation. Thus, this questions cannot
really distinguish between extraverts and introverts. 
 
Trope and Bassok (1983) replicated Snyder and Swann’s original study but
using truly diagnostic questions (e.g. do you like loud parties? Do you shy
away from social interactions?”). Also varied question diagnosticity so that
some of the questions, like the ones above, were highly diagnostic of
extraversion and introversion, and some were only somewhat diagnostic (e.g.
do you talk loudly?). Participants were found to favour highly diagnostic
questions over less diagnostic ones, and diagnosticity was the most
important determinant of their choice of questions.  
Participants even preferred highly diagnostic questions that did not match
their hypotheses over less diagnostic ones that did. 
Participants also relied on positive test strategy, preferring to ask questions
that matched their hypotheses. 
Ultimately, people do not rely on the positive test strategy to the point
of ignoring important information about diagnosticity; we do try to ask
the most informative questions we can. We also rely on the positive test
strategy when seeking evidence needed to test our hypotheses, and
this can bias our judgements. Even judgements about the self can be
a ected by reliance on the positive test strategy.
C. Hypothesis Con rmation in Evaluating Oneself 

We need to answer questions about ourself daily. (E.g. do I support vegans?
Will I do well on a canoeing trip?, etc.). Sometimes we will have a ready,
prestored summary answer (e.g. we know we are vegan and sporty as a
38
ff
ffi
fi
fi
central part of our self-image). But more often than not, we need to construct
our answer on the spot, based on what we know about our related
behaviours, thoughts and feelings. In these cases, if we rely on the positive
test strategy, we may attempt to answer one-sided questions by selectively
recruiting information about ourselves that matches them. 
Therefore, people asked whether they are assertive will likely view
themselves as more assertive than people asked whether they are
unassertive. 
Kunda et al. (1993) study where half the participants were asked to list
examples of their past thoughts, feelings and behaviours that came to mind
as they tried to answer if they were happy with their social life. The remaining
half were asked to do the same for the opposite questions if they were
unhappy with their social life. Participants engaged in the positive test
strategy, those asked whether they were happy recruited more happy
thoughts, fewer unhappy thoughts than those asked whether they were
unhappy.  
Participants asked whether they were happy rated themselves as
happier with their social lives than did participants asked whether they
were unhappy. 
Hypothesis con rmation also took place. Those that were asked if they
were happy rated themselves as happier with their social lives than
participants who were asked if they were unhappy. 
Further studies revealed that one-sided questions about the self result
in hypothesis con rmation only when the relevant knowledge base is
mixed and so capable of supporting opposite hypotheses. Tendency to
con rm hypotheses was eliminated when people’s knowledge base was
relatively uniform, the case where questions focused on relatively
consistent domains of the self or when questions were addressed to
individuals whose social behaviour was relatively consistent. In these
cases, one sided questions’ directions had no impact on self-views.
D. Choosing vs. Rejecting 

Sha r (1993). Parents A and B. A (Average income, average working hours,
reasonable rapport with child, relatively stable social life). B (Above average
income, very close relationship with the child, extremely active social life, lots
of work related travel, minor health problems). 
When asked who was the better parent, people would zoom into the good
and say parent B. When asked who was the worse parent to put the child
with, people would zoom into the bad and say parent B too. 
Suggests that when we are choosing winners, we will favour complex
39
fi
fi
fi
fi
people who have striking strengths and weaknesses over more bland
and unremarkable ones, because we have more reasons for choosing
said complex individuals.  
But when we are weeding out losers, we will also choose to reject
complex people over the more bland ones, because we have more
reasons for rejecting the complex ones.
E. Analysing Reasons for Predictions 

When people start o underestimating the likelihood of behaviour,
analysing reasons can boost accuracy by increasing their likelihood
estimates. 
We often expect that thinking hard about an issue will improve our
reasoning. However, careful re ection leads us to rely on heuristics
such as the positive test strategy that can increase bias.
III. Covariation Detection 

Do computer hackers tend to lack interpersonal skills? Do better scientists make
better teachers? Are children who are unruly at home also disruptive at school?
These questions concerns a covariation between 2 variables. 
Much of our social knowledge is based on beliefs about such covariation, and it
is important to determine just how good we are at assessing covariation.
A. Assessing Covariation from 2 x 2 Tables 

Suppose you had to determine whether there was a relationship between
being a professor and being absentminded. 
Research has been done, summarised information in a table, Absentminded/
Professor? (Yes, Yes, 600), (Yes, No, 300), (No, No, 200), (No, Yes, 400).
Examining this information, are professors particularly likely to be
absentminded? 
Correct answer is to nd out whether ratio of professors who are
absentminded to professors who are not (600 to 400, 3:2) is greater than the
same ratio for people who are not professors (300 to 200, 3:2 too). So
technically there is no conclusion that professors are absentminded.  
But people do not realise all 4 cells are relevant and tend to focus on just 1
or 2 of these cells, many focus on the yes-yes cell and conclude that
professors are absentminded. 
Reliance on positive test strategy, looking for cases that match the
hypothesis. People have di culty with more naturalistic covariation
detection tasks.
B. Illusory Correlation 
Redelmeier & Tversky (1996), suspected that faulty covariation detection
40
fi
ff
ffi
fl
contributes to belief that arthritis pain is related to weather. Long story short,
there was no signi cant correlation in the experiment that weather was
related to pain. 
People’s prior theories lead us to expect a correlation, and we will pay
special attention to cases that embody this correlation. 
Chapman & Chapman (1967, 1969) study on Rorschach test and what
homosexual/heterosexual people perceived in the blots showed that  
Perception of illusory correlations are is driven by prior expectations,
and can persist even in the fact of data in which these correlations are
nonexistent, or even completely opposite to what they believe in.
C. Stereotype Formation Through Illusory Correlation 

We are likely to overestimate the frequency of rare behaviours among
members of relatively small groups. When both the group and behaviour
are rare, their co-occurrence (i.e. a group member who performed the
behaviour) will be even rarer. 
Distinctive individuals will be especially noticeable and memorable, and
so will give rise to an illusory correlation between group membership
and the behaviour. 
Hamilton and Gi ord (1976). “Bruce” and “Joe” experiment about their group
membership and behaviour. Ratio of positive to negative behaviours for both
were identical. Despite lack of actual correlation, participants saw an illusory
correlation between group membership and positivity of behaviour,
overestimating the frequency with which members of smaller Group B
performed the rarer negative behaviours. They viewed group B more
negatively than group A. 
We pay more attention to members of small groups who perform rare
behaviours, and are particularly likely to remember them can explain
why we come to view small groups as especially likely to perform rare
behaviours. 
If we expect a minority group to be especially likely to produce
criminals, we will pay special attention to criminal members of that
group not only because they are distinctive but also because they
con rm our hypotheses. We may see support for our negative
stereotypes where none exists.
D. Failure to Detect Actual Correlations 

In the absence of prior theories, people may sometimes fail to see
unexpected correlations that do in fact exist. 
Chapman and Chapman (1969) found that participants failed to detect
41
fi
ff
fi
unexpected correlations for real correlations between homosexuality and
some counterintuitive but valid signs of homosexuality. 
Jennings, Amabile, and L. Ross (1982) study showing men of varying heights
holding walking sticks of varying heights and asking participants to estimate
the strength of relationship between them found that… People are usually
able to detect extremely strong correlations (0.8 or above), and rated
their magnitude as high. But they rated even quite strong correlations
(0.7) as moderate at best, and often failed to detect more modest
correlations (0.2-0.4), rating their magnitude close to zero. 
People’s inability to detect such modest correlations even from simple
and sanitised data sets implies that we will likely also fail to detect
many real world correlations when these are not predicted by our
theories.
E. Accuracy or Inaccuracy in Covariation Detection 

We may be capable of remarkable accuracy when assessing some everyday
correlations.  
Kunda and Nisbett (1986), series of studies compared people's assessments
of correlations in several domains that varied in familiarity and codability to
the actual correlations in those domains. 1st study examined perceptions of
degree of agreement among people on evaluations that varied in familiarity.  
-People were remarkably accurate in their estimates of covariation in
familiar domains. One such domain concerned the agreement among
people in their assessments of others’ personalities. 
-People spend a great deal of time discussing with other people and are
highly familiar with the extent to which they agree with others when
evaluating di erent aspects of personality. (E.g. people know, that
others are far more likely to agree with one another when evaluating
their acquaintances’ talkativeness than when evaluating their likability). 
-People may be quite inaccurate of their assessments of unfamiliar
correlations but can be quite accurate in assessments of highly familiar
correlations. 
-Familiarity however, can only be an asset if the data are also codeable.
Some familiar behaviours are very di cult to code, such as social
behaviour, as compared to athleticism and IQ. 
-Social data changes over time, hard to keep track between relationship
of 2 variables. 
-Inconsistencies in memory due to the lack of salient characteristics.
42
ff
ffi
F. Explanation Based Judgement 
We make judgements by recruiting independent pieces of data, then
somehow add up or average the implications of the recruited instances
by assessing their relative frequency or the ease with which they were
recruited to come up with our nal judgement. 
However, many judgements require more complex causal reasoning and
draws on extensive world knowledge to relate di erent pieces of
evidence to each other and to likely outcomes.
G. Story Models in Jury Decision Making (Explanation Based Judgement) 

Pennington and Hastie (1992) presented a criminal case in di erent orders
that a ected the ease and di culty of constructing the prosecution and the
defence stories. The story would be easy to construct when evidence
pertaining to it was presented in a typical “story order” which preserved
causal and temporal sequence of events, but di cult to construct when
evidence was presented in a jumbled order that did not preserve the original
sequence of events. 
Found that Altering order of evidence presentation without altering its
content can more than double the likelihood that the defendant will be
found guilty. People do not merely add up independent implications of
each piece of evidence to arrive at a verdict, but rather we arrive at a
verdict by constructing a story, factors that a ect the relative ease of
constructing alternative stories will also a ect our verdicts.
H. Coherence of Explanations. 
Our faith in a hypothesis increases with its explanatory breadth, its
simplicity, and the extent to which it, in turn, can be explained by other
information. As our faith in one hypothesis increases, our faith in
competing hypotheses decreases. 
We prefer hypotheses that can be explained with other information.
IV. Counterfactual Thinking 

Many events lead people to generate such counterfactual outcomes, e.g.
outcomes that run contrary to what had actually happened, outcomes that
might have, could have, perhaps should have happened but did not.
Counterfactuals that we generate in response to an event can in uence
our understanding of that event and our emotional reactions to it.
A. Event Normality 
Kahneman and Tversky (1982). Slot machine phenomenon. You wasted
money on a slot machine for an unlucky streak. You leave, and someone
43
ff
ffi
fi
ff
ffi
ff
ff
ff
fl
plays on that machine, and wins 10,000 on his rst try. You will feel more
frustrated and disappointed than any of the nonwinners present because it is
easier for you to imagine that you had hit the jackpot, if only you had inserted
one more coin. 
Daniel Kahneman and Dale Miller (1986) termed events that can easily be
imagined otherwise abnormal. 
Kahneman and Tversky (1982). Missed ight scenario. 
The more abnormal an event seems, the stronger one’s emotional
reaction to it. 
Anticipatory regret can lead us to avoid actions that might lead to such
events. 
Roese and Olson (1995), Turnbull (1981), Miller and McFarland (1986). These
studies show that… People who have experienced or observed identical
negative outcomes feel worse if they can generate positive
counterfactuals more readily. More surprisingly, people who are
objectively better o than others can feel worse if they can imagine a
still more positive outcome more readily. 
Gavanski and Wells (1989) Tendency to undo exceptional rather than
routine events is particularly strong when the events lead to exceptional
outcomes. 
Miller and McFarland (1986) People nd exceptional events easier to undo
and therefore view them as more regrettable. 
Miller, Taylor, and Buck (1991) People nd actions of exceptional
individuals easier to undo than those of more “normal” individuals. 
Kahneman and Tversky (1982), Davis et al. (1995) People attempt to undo
actions that are under control of the individual they are focusing on.
When focusing on the victim of a crime, we tend to presuppose the
actions of the perpetrator, which become part of the immutable
background, and we attempt to undo the victim’s actions. 
Miller and Taylor (1995) People regret and anticipate regretting actions
more than inactions.  
However, Gilovich and Medvec (1994, 1995) observed that the very
opposite pattern is obtained when people look back on their lives. The
pattern of regret we experience depends on the passage of time, in the
short term our actions seem most regrettable, but in the long run our
failures to act cause us the most grief. 
We may nd failures to act particularly disturbing as they recede into
the past because often in retrospect, we are unable to understand why
we had failed to act. Obstacles that loomed large at the time no longer
44
fi
ff
fi
fi
fl
fi
seems so insurmountable when we look back at our lives. We think we
could have handled with ease tasks that in retrospect seemed all but
impossible. As time passes, regret over inaction becomes intensi ed
because failure to act becomes inexplicable, today we cannot
understand what stopped us back then from approaching that
attractive person, what prevented us from getting a better education. 
Also, inactions become more regrettable than actions as time passes
because it is easier to determine and deal with the consequences of our
actions, we know where the road we have chosen has led us. And if
those actions brought about bad consequences, we have often dealt
with these through corrective action or thought.
B. Ease of Replicating Events Mentally 

Miller, Turnbull, and McFarland (1989), Miller et al. (1989). Cookie and child
experiment, and discrimination against women in terms of exam and grades. 
Ease of replicating events mentally can also a ect judgements of
fairness and of foul play in situations that do not involve random draws.
C. Functions of Counterfactual Thoughts 

Wells and Gavansky (1989), thought of “if only I had left home earlier I would
have made my ight”. People think about how misfortunate might have
been avoided, and this may identify the causes and circumstances that
led to the misfortune and we would be better prepared to deal with
such circumstances in the future. 
Gleicher et al. (1995), Roese (1997), Roese and Olson (1995, 1997).
Counterfactual thoughts are particularly likely to be triggered when we
are feeling badly; they may provoke us into action that will ameliorate
the negative a ect as well as action that will prevent the recurrence of
the misfortune that had given rise to that a ect. 
Roese (1994) students recalling exam they had done poorly for. Generating
counterfactuals about how to do better made students see themselves
as more likely to engage in behaviours that could improve their
academic performance, such as studying their notes frequently and
attending all lectures.
D. Final Summary 
People rely on positive test strategies which entail seeking cases that
match the hypothesis. They use this strategy both when searching their
memories for pre-existing knowledge and when searching the external
world for evidence that bears on the hypothesis.  
45
fl
ff
ff
ff
fi
When evidence base being searched is mixed, containing information
that con rms hypothesis as well as discon rming information, use of
the one sided positive test strategy can bias people toward con rming
their hypotheses. Social world is complex and inconsistent, we may often
be biased toward con rming any hypothesis we entertain about our own or
other people’s attributes. 
 
People rely on positive test strategies when assessing covariation
among variables. Paying special attention to those cases that match our
hypotheses about what correlates with what, and overestimate the
magnitude of the hypothesised covariation. When we have a strong
prior theory that one variable is correlated with another, we may see
illusory correlations that do not actually exist in the data, because we
are likely to notice and remember cases that embody the theorised
correlation. Salience, memorability and distinctiveness factors can give
rise to illusory correlations. Members of minority groups who perform
unusual behaviours are especially distinctive and people are prone to forming
illusory correlations between group membership and unusual behaviour 
 
People can actually detect many everyday correlations with impressive
accuracy. We are especially likely to be accurate about correlations in
domains such as sports that are both highly familiar and easy to code.
However, social behaviour is di cult to code and we may therefore be
especially prone to seeing illusory correlations and failing to see actual
ones in the social domain. 
 
People follow reasonable principles when evaluating the coherence of
accounts, preferring those that explain more evidence, those that
require fewer additional assumptions, and those that can be explained
by other information. However, irrelevant factors such as order with
which we encounter information can in uence the ease with which we
can construct a particular account can also in uence our judgements. 
 
Events lead us to generate counterfactual outcomes that can in uence
our understanding of the event and our emotional reactions to it.  
 
Events that can be easily imagined otherwise seem abnormal, and the
more abnormal they seem, the stronger our emotional reactions to
them. 
46
fi
fi
ffi
fl
fi
fl
fi
fl
 
Event seems more abnormal if our model if it is very close to the model
of an alternative, counterfactual event, if only a few things had been
di erent, the outcomes would have been quite di erent. When the close
counterfactual is more positive than what actually happened we may
feel especially upset about having failed to achieve it, and when the
close counterfactual is more negative than what actually happened, we
may feel especially grati ed about having narrowly missed it. 
 
When people contemplate events that resulted from action or inaction,
our emotional reactions depend on the passage of time. Actions
provoke greater regret in the short term, but inaction provokes greater
regret in the long run.
 
From What Might Have Been to What Must Have Been: Counterfactual Thinking
Creates Meaning 
Neal J Roese et al. (2010 
I. 4 Experiments  
Explored whether 2 uniquely human characteristics, counterfactual thinking
(imagining alternatives to the past) and the fundamental drive to create meaning
in life, are causally related. 
Experimenters hypothesised and found that counterfactual thinking heightens
the meaningfulness of key life experiences. Re ecting on alternative pathways
to pivotal turning points even produced greater meaning than directly re ecting
on the meaning of the event itself. Fate perceptions and bene t nding were
independent causal links between counterfactual thinking and the construction
of meaning.
II. Experiment 1 
Tested hypothesis that constructing counterfactual worlds for an event
enhances the meaningfulness of the event in one’s life.  
 
Method 
32 people, 5 male, 15 female, 12 information unavailable, racially diverse
undergraduates. 2 between subject conditions, counterfactual and baseline. 
 
Procedure 
Participants wrote short essay about why they ended up at their current
47
ff
fi
fl
ff
fi
fi
fl
university. They were instructed to think about how they decided where to go for
college. How did they end up coming to Northwestern? 
After completing the essay, half of the participants engaged in counterfactual
thinking, asking them to describe the ways that things could have turned out
di erently. The other half comprising the baseline, did not receive this prompt
and proceeded to the meaning questionnaire. 
 
People who had considered (downward) counterfactual alternatives
reported that their college choice was more meaningful and signi cant.
Imagining the consequences of having chosen a di erent path caused
participants to endorse more strongly the view that their ultimate college
choice was a de ning moment in their life.
III. Experiment 2 
Explored whether counterfactual thoughts impart greater meaning to
relationships. Understanding the impact of counterfactual thoughts on how
relationships are perceived by looking beyond romantic relationships to
friendships in general. Ruled out the e ect of counterfactuals in Experiment 1,
as it could be due to greater cognitive demands of the counterfactual task that
caused the e ect. 
 
Method 
Same. 
 
Procedure 
All participants wrote a short essay recounting how they met a close friend.
Asked to describe the various factors and sequence of events or circumstances
that led to your meeting. 
Participants in counterfactual condition were instructed to describe all the
possible ways that they might not have met the person and how things could
have turned out di erently.  
Those in factual condition were prompted to describe any other details about
the way they met that determined how things ultimately turned out. 
The 2 tasks were designed to be comparable in cognitive demand and
consequential thinking. 
 
Counterfactual re ection (downward) led the close friend relationship to be
seen as more meaningful and self-de ning than did the factual re ection
(how we might not have met vs. how we met). 
48
ff
ff
fi
ff
fl
ff
fi
ff
fi
fl
Although close friends are by de nition special, recognising the
contingencies underlying the relationship accentuated the relationship’s
signi cance. 
 
Counterfactual re ection (downward) endows both major life experiences
and relationships with greater meaning. Mentally subtracting a signi cant
other from one’s life increases satisfaction with the relationship. By
identifying ways in which their lives would be worse without existence of a
focal relationship, participants imbued the relationship with meaning.
IV. Experiment 3 
Fate and Bene t Finding Mediate the Link Between Counterfactual Thinking and
Meaning. Connecting fate perceptions to the creation of meaning in life
narratives. 
Also tested an alternative explanation for why counterfactual thinking enhanced
fate perceptions in the previous experiment. 
Also tested to determine whether link between counterfactual re ection and
meaning is limited to experiences that were positive to begin with, as positive
a ect is associated with experience of meaning in contexts. 
 
Method 
Participants asked to write a series of discrete statements about the turning
point. 
Participants were instructed to describe the signi cance of the turning point
incident in detail. 
 
Counterfactual re ection heightens perceptions of fate. By contrasting
reality with what might have been, people gained clarity about the positive
consequences of their turning point. 
 
Mentally undoing turning points lead people to derive more meaning from
the turning point than did directly pondering the meaning of the event.
Contrast between reality and what might have been seems to crystallise an
event’s signi cance in a way that direct attempts to generate its meaning
cannot. 
 
Counterfactual re ection predominantly involved the spontaneous
generation of downward counterfactuals, or considerations of worse
possible worlds if the turning point had not occurred. 
49
ff
fi
fi
fi
fl
fl
fl
fi
fi
fl
fi
 
Generation of downward counterfactuals fully mediated the relationship
between how turning points were re ected upon and the sense that they
were fated. On the other hand, indulging in upward counterfactuals may be
associated with dysfunctional consequences. (Envy, dissatisfaction?). 
 
Counterfactual thinking, the pondering of what might have been, brings us,
subjectively, closer to answering our understanding of life. Through
mentally veering o the path of reality imaginatively, we forge key
connections between what might have been and what was meant to be,
thereby injecting our experiences and relationships with deeper meaning
and signi cance. 

MISINFORMATION
Planting Misinformation in the human mind, a 30 year investigation of the malleability
of memory. Elizabeth F Lotftus.
I. Background 
Misinformation e ect refers to the impairment in memory for the past that arises
after exposure to misleading information. The phenomenon has been
50
fi
ff
ff
fl
investigated for at least 30 years, as investigators have addressed a number of
issues. These include conditions under which people are especially susceptible
to the negative impact of misinformation, and conversely when are they
resistant. Warnings about the potential for misinformation sometimes work to
inhibit its damaging e ects, but only under limited circumstances. 
Some people are more susceptible to misinformation than others.
II. The Big Questions
A. Under what conditions are people particularly susceptible to the

negative impact of misinformation? (WHEN)
1. People are particularly prone to having their memories a ected by

misinformation when it is introduced after the passage of time has
allowed the original event memory to fade. (Loftus et al. 1978). 
With the passage of time, event memory is weakened, and there is
less likelihood that a discrepancy is noticed while the misinformation
is being processed, and the mind would readily embrace the
misinformation.
2. Discrepancy Detection Principle (Tousignant et al. 1986). Recollections

are more likely to change if a person does not immediately detect
discrepancies between misinformation and memory for the original
event.
3. A short interval between misinformation and a subsequent test

causes subjects to be less likely to claim that the misinformation
item was in the event only, and more likely to claim that the item was
in both parts. (Higham 1998).
4. Temporarily changing someone’s state can increase misinformation

e ects. If people are led to believe that they have drunk alcohol, they
are more susceptible (Asse and Garry 2002), and when people are
hypnotised, they are more susceptible (Scoboria et al. 2002). These
states have the e ect of disrupting ability of subjects to detect
discrepancies between misinformation and what remains of their
original memory.
B. Can people be warned about misinformation, and successfully resist its

damaging in uence?
1. In accordance with the Discrepancy Detection Principle, if people are

warned prior to reading post-event information that the information
51
ff
fl
ff
ff
fi
ff
might be misleading, they can better resist its in uence, by
increasing likelihood that the person scrutinises the post event
information for discrepancies.
2. Warnings given after misinformation has been processed did not

improve abilities to resist damaging e ects. (Greene et al. 1982).
3. Warning people that they may have in the past been exposed to
misinformation (post misinformation warnings) may have some
success, but in limited circumstances. 
Highly accessible misinformation causes immediate post-
misinformation warnings not to work at all. (Accessibility of
misinformation is higher when it is presented multiple times, vs. a
single time).
4. Warnings need not be item speci c. They can also be general. (Eakin
et al. 2003). Favoured suppression hypothesis, when people get a
warning, they suppress the misinformation and it has less ability to
interfere with answering on the nal test.
5. Highly accessible misinformation might distract subject from

thinking to scrutinise the misinformation for discrepancies from sone
presumably overwhelmed original event memory.
C. Are some types of people particularly susceptible? (WHO)
1. Young children are more susceptible to misinformation than are

older children and adults. (Ceci and Bruck 1993). Suggests that age-
induced distortion in memory is a phenomenon that occurs with
people of all ages, but more so when attentional resources are
limited, especially in young children and the elderly.
2. Personality traits such as empathy, absorption, and self-monitoring

have greater susceptibility to misinformation. The more one has self-
reported lapses in memory and attention, the more susceptible one
is to misinformation e ects. (Wright and Livingston-Raper 2002).
3. Pigeons, who have good memory, can be disrupted by

misinformation e ects when exposed to post-event misinformation.
(Red light peck, other lights not to peck to get food). (Harper and
Garry 2000; M. Garry and D.N. Harper, in prep.). 
Pigeons are more susceptible, like humans, to misinformation if it
52
ff
ff
fi
fi
ff
fl
occurs later in the original- nal test interval than if it occurs early in
that interval.
4. Retrograde interference (forget previous thing when learning/

memorising new thing) is a mere disruption in performance, not a
biasing e ect. It typically makes memory worse, but does not pull for
any particular wrong answer.
D. When misinformation has been embraced by individuals, what happens

to their original memory?
1. Misinformation probably has little e ect on original memory, but this

is still not con rmed. (McCloskey and Zaragoza 1985).
2. Misinformation items get slotted into memory because people have

no original memory due to it having never been stored, or have faded
over time. Sometimes this happens because of deliberation, and
sometimes it appears as if the original event memories have been
impaired in the process of contemplating misinformation. (Wagenaar
and Boer 1987; BElli 1989; Tversky and Tuchin 1989).
E. What is the nature of misinformation memories?
1. Misinformation memories’ verbal descriptions (unreal memories)

were longer, contained more verbal hedges (I think I saw…), more
references to cognitive operations (After seeing the sign the answer I
gave was more of an immediate impression…), and fewer sensory
details. (Schooler et al. 1986).
2. Many of the unreal memory descriptions contained verbal hedges

and sensory detail, making it extremely di cult to take a single
memory report and reliably classify it as real or unreal.
F. How far can you go with people in terms of the misinformation you can
plant in memory?
1. False memories of familial memories planted by relatives of test

subjects. (E.g. Getting lost in a mall for rather long, accident at
family wedding, victim of animal attack, nearly drowned and
rescued, etc.) (Loftus 1993; Loftus and Pickrell 1995), (Hyman Jr. et al.
1995), (Porter et al. 1999), (Heaps and Nash 2001).
2. Sometimes subjects will start with very little memory, but after
several suggestive interviews lled with misinformation they will
53
ff
fi
fi
fi
ff
ffi
recall the false events in quite a bit of detail. (Low blood sugar
experiment) (Ost et al. 2005).
3. Power of suggestion is strong. This can lead people to believe or

even remember in details that did not happen, that were completely
manufactured with the help of family members, and that would have
been traumatic had they actually happened.
4. Guided imagination (Libby 2003), Suggestive dream interpretation

and exposure to doctored photographs have also led subjects to
believe falsely that they experienced events in their distant and even
recent past (Loftus 2003).
5. Visual aids produced more false memories than verbal aids. (Bugs
Bunny Experiment) (Braun et al. 2002).
G. SUMMARY
1. Distortions occur in the real world not only from explicit

misinformation. As we retrieve and reconstruct memories,
distortions can creep in and these will eventually solidify into
misinformation.
True Photographs and False Memories, D. Stephen Lindsay, Lisa Hagen, J. Don
Read, Kimberley A. Wade, and Maryanne Garry, 2004.
I. Background 
Old photos might cue long-forgotten memories, but when combined with other
suggestive in uences they might also contribute to false memories. 
Not surprising that doctored photographs are powerfully suggestive, as people
perceive photographs as compelling evidence that depicted events really
occurred, and photos provide a rich source of information regarding the
perceptual details of suggested events.
II. Experiment Subjects 

45 undergraduates (36 women, 9 men) who volunteered to participate.
III. Procedure 
Each subjects’ parents provided brief narratives describing 2 unique, school
related events experienced by their child, one event in Grade 5/6 and the other
in 3/4. 
Parents asked to only give rarer family stories. 
All parents con rmed their child have never experienced experiment’s target
54
fl
fi
pseudoevent, (putting Slime, a brightly coloured gelatinous compound
manufactured by Mattel as a toy, in the teacher’s desk in Grade 1/2). 
Parents also provided child’s class photo for each of their school years
corresponding to target events, name and gender of their child’s Grade 1/2
teacher. 
1-1 interview, experimenter read each narrative aloud and asked the subject to
recall it, starting with 5/6 Grade and working back in time to Grade 1/2. 
Random assignment, 23 subjects were given photocopy of their school’s class
group photo for each year before corresponding narrative was read to them. 
Subject was egged on to recall as much as possible about each event, and had
to rate extent to which the memory experience resembled reliving the event,
extent to which the subject felt he or she was remembering the event, and their
con dence that the event had occurred. 
End of session, subjects were told to spend some time each day over the week
working at remembering more about the event, and given a printed copy of the
narrative and for the subjects in the photo condition, additionally a copy of the
class photo. 
Week after initial interview, subjects returned to the lab and had to recall.
IV. Results 
False memory reports were more common for photo conditions than for
no-photo condition, but DID NOT APPROACH STATISTICAL SIGNIFICANCE. 
However, subject ratings for memories of suggested event were
signi cantly higher in the photo than no-photo condition. 
Strong convergence between judges’ categorisations and subjects’ self-
ratings (e.g. subjects who were judged to have neither images nor
memories indeed selected ratings near the bottom of the scale on each
measure. 
Ratings of memories of the pseudoevent by subjects categorised as having
false memories were equivalent to and sometimes directionally greater
than ratings of memories of the true events, indicating that subjects’ false
memories were as compelling as memories of true events, at least on
these dimensions. 
Ratings of subjects classi ed as reporting images but not memories were
more similar to ratings of subjects classi ed as having neither memories
nor images than they were to ratings of subjects classi ed as having
memories. Suggests that images but no memories category should not be
considered tantamount to false memories.
V. Discussion
55
fi
fi
fi
fi
fi
A. Supplementing other suggestive in uences with a photo associated
with (but not depicting) the suggested pseudo-event doubled the rate of
false memory reports, yielding a substantially higher rate of false
memory reports than any prior study.
B. Photos may add to the authoritativeness of the suggestive narrative,

increasing con dence in that false memory/event.
C. The lack of visual aid may cause people to feel more di cult to enter
into speculations on a false memory/event because of the inability to
recall relevant details, such as appearances of people in that said
memory.
D. People categorised as experiencing images but not memories often

appeared to be speculating about, rather than remembering, the false
memory/event.
E. Many research journals pertaining to false-memory induction

sometimes collapse across partial and complete false-memory reports,
either by not distinguishing the 2 categories at all or by emphasising the
sum of both when summarising the false memory rate. 
This should not be appropriate, because self-ratings of subjects
classi ed as having images but not memories (analogous to terms of
partial false memories) more closely resembled the self-ratings of
subjects judged to have neither memories nor images than they
resembled the ratings of subjects judged to have false memories. 

Kunda Chapter 5 Memory
I. State of Mind When Encoding Events 

Our state of mind at the time we observe events, as we encode them and
store them in memory, may a ect how we later recall these events.
II. Expectancies and Interpretation
56
fi
fi
ff
fl
ffi
A. We often approach people and events with prior expectancies, these
may arise from concepts such as stereotypes. They can also come from
other sources such as external information/priming.
B. Expectancies can guide the interpretation of behaviour (e.g. a smile

when you know Laura is happy vs. a smile when you know Laura is
arrogant). 
Our expectancies may determine the very meaning we ascribe to
someone’s behaviours as we observe them, especially if these
behaviours are ambiguous and can be understood in more than one
way. 
Over time, we may forget speci c behavioural details we have observed
pertaining to particular person, and we only recall the gist, the meaning
that we had imposed on them.
III. Expectancies and Attention
A. Expectancy-Congruent Information: We may be particularly likely to

notice information that is congruent with our expectations, if we expect
Laura to be warm, we may pay special attention to her warm
behaviours, and as a result we will be especially likely to recall these in
the future.
1. We may be most vulnerable to expectancy-congruent information

when stressed or aroused.
2. Reliance on the positive test strategy can cause people to pay

special attention to events that match their hypotheses and
expectations and therefore increase memorability of events.
3. Increased memorability of congruent events can lead people to

overestimate the extent to which the information they have observed
supports their hypotheses.
4. People may believe illusory correlations because they pay special

attention to cases that embody those correlations that they expect
to see. (E.g. Asian people are good at math, and you may be likely to
pay attention to any Asian American math whiz you encounter.).  
This leads to overestimation of frequency.
5. People are more likely to recall information that was consistent of

their stereotype of a person/object than they are to recall the
stereotypically-inconsistent information.
57
fi
6. Congruent events may be memorable not only because we pay more
attention to them but also because they are more strongly related to
our existing beliefs.
B. Expectancy-Incongruent Information
1. People are particularly likely to notice and recall information that

violates our expectations.
2. People are particularly likely to pay attention to members of minority

groups performing a rare behaviour because the co-occurrence of a
member of a scarce group performing an uncommon behaviour is so
unusual, and they subsequently overestimate the frequency of such
distinctive individuals, and form a mistaken belief that minority
members are especially likely to perform that uncommon behaviour.
3. People are particularly likely to recall information that violates their

expectancies. 
This is because they capture our attention but also our minds work
hard to reconcile these incongruent events with our expectancies. 
Similarly, cognitive structures that lend coherence to behaviours that
are consistent with expectations may also serve to group and
impose meaning on inconsistent behaviours, and so increase their
memorability.
4. Incongruent information will be relatively better recalled when

people are MOTIVATED to form an accurate impression that takes all
relevant information into account.
5. Incongruent information will be relatively poorly recalled when they

are NOT MOTIVATED to form an accurate impression and are not
concerned with investing e ort in reconciling incongruent
information with their expectancies/incapable of such e ort due to
being preoccupied with other cognitive tasks.
IV. Goals
A. The goal of forming an impression of another can lead you to pay more
attention to information about that person and invest greater e ort in
making sense of that information.
1. People organise the information in a meaningful way to nd ways of

relating one piece of information to another.
58
ff
ff
fi
ff
2. Increased attention and organisation allows for better recall.
B. Personal signi cance a ects the way we process the event and its
resulting memorability.
V. Event Signi cance
A. Flashbulb memories are an example of the importance of the personal

signi cance of events to a memory. (Brown & Kulik, 1977).
1. Personally signi cant memories may be better recalled because of

their emotional signi cance, BUT there is little direct evidence that
emotional intensity increases the memorability of dramatic events.
2. Stress hormone release during the witness of dramatic events can

contribute to the memorability.
3. Flashbulb events are memorable because of how dramatic and

di erent they are to mundane events, and people may work harder
to make sense of it/circumstances surrounding them.
4. Flashbulb events, while memorable, are by no means absolutely

accurate. Susceptible to repeated rehearsal and retrieval (share
these stories and become inaccurate over time), vivid imagery
causes us to be overcon dent that our information is correct, and we
underestimate the extent to which we use pre-existing schemas to
ll in the blanks. (Neisser & Harsch, 1992, Students describe
Challenger explosion).
VI. Expectancies and Retrieval
A. Even if people try to encode events in a relatively unbiased manner,

recollection of these events may still be biased by their expectancies at
the time of retrieval.
B. Expectancies that derive from a person’s own behaviour can bias

memories about this person in a similar manner.
1. Expectancies lead not only to selective retrieval of a biased subset of

observed behaviours, but also to memory distortion, transforming
one’s recollections of particular behaviours.
2. Even if expectancies are subtle, they can a ect our reconstruction of

events/impressions of others.
59
fi
ff
fi
fi
fi
fi
fi
ff
fi
ff
C. Implicit Theories construct personal history, as we assess our current
standing, consulting implicit theories about the relation between the
past and the present. (Ross, 1989).
1. If the person's goal was to perceive that they improved, they prefer
to remember in the past that they were horrible and then now they
are doing so much better. Perceiving their current standing in a
positive fashion, and to satisfy the implicit theory of the ability for
human change. (Conway & Ross, 1984, Study Skills Course, where
people had no objective improvement in academic performance, but
participants who attended the useless study course recalled earlier
study skills as worse than present, whereas participants who were
assigned to a waiting list control condition did not).
VII. Theories about Stability and Change
A. People tend to believe that they are stable on many dimensions (e.g. I
am the same as I was 10 years ago). This is called Theory of Stability,
and can lead us to conclude that because we like (or dislike) a particular
person today, you must have always felt the same way.
B. People can use mistaken theories (did not factor that they have
changed) to reconstruct our past state of mind, and these will be
systematically biased.
C. Recollections about our past behaviours may also be biased by our

current attitudes.
VIII. The Hindsight Bias
A. People tend to exaggerate the extent to which they could have

predicted currently known outcomes due to hindsight bias, as they try
to reconstruct their past state of knowledge.
B. Knowledge of outcomes can bias our estimates of what we would have

predicted without outcome knowledge.
1. Outcome knowledge can also bias our recollections of what we had

actually known before the fact.
C. Hindsight bias appears for many events, historical, scienti c, sports,

games, elections, medical, legal and psychiatric cases, accidents, etc.
60
fi
1. Preceding events take on new meaning and importance as they are
made to cohere with the known outcome.
2. People’s inability to reconstruct earlier state of knowledge correctly

may lead them to be overly harsh towards them and themselves
following decisions that turned out badly.
IX. Goals
A. Goals may bias our recollections by leading us to pose one-sided

questions. (E.g. due to positive test strategy).
B. Memories associated with the desired trait had become especially

accessible. Memory of how we have behaved in the past can be
coloured by how we wish we had behaved.
C. Motivation may not only in uence which incidents we bring to mind but
also in uence how we reconstruct and distort our memories of
particular incidents.
X. Mood
A. Mood Congruent Memory: People are particularly likely to recall

memories that are congruent with their current mood.
B. Mood Dependent Memory: Mood can in uence memory in another way.

Events that are encoded in a certain mood may be best recalled when
we are again in that same mood. This is termed mood-dependent
memory, focuses on t between mood at retrieval and mood at
encoding.
1. People are especially likely to show mood-dependent memory when

the recalled events are internal (thoughts, feelings, imaginations)
rather than external, when we are not explicitly reminded of these
events at the time of recollection, and when our moods at encoding
and at retrieval are su ciently intense.
2. When people are happy, they may be particularly to recall happy

events (mood-congruent memory), and may be particularly likely to
recall events experienced on previous happy occasions (mood-
dependent memory). Similar as sadness and anxiety.
3. Emotions can be represented as nodes in a network, e.g. happy

events linked to happy node, and sad events to sad nodes.  
61
fl
fi
ffi
fl
fl
Events experienced when one is happy or sad are also linked to
these emotion nodes. This increased activation increases the
likelihood that one will retrieve these associates.
4. Emotions not only activate congruent material but also inhibit

incongruent material and incongruent emotions.
5. Our recollections can be coloured by our moods.
6. People’s memories of past events are reconstructed, in uenced by

expectancies, beliefs, goals, and feelings that were on our mind
when we rst experienced the events as well as by those that are on
our mind at the time of recollection. Bias our memories in systematic
ways. Beliefs about sources of our memories are also reconstructed,
and subject to bias.
XI. Distinguishing Reality from Imagination
A. Memories of actual events have more temporal and spatial attributes

than do memories of imagined events, more sensory attributes, more
detailed and speci c information and more emotional information.
B. If a memory includes logically impossible events (e.g. you jumping o

the cli and surviving), you may conclude that this event must have
been imagined. Conversely, if recalled events seem coherent and
logically interrelated, you may view the memory as more accurate. 
Imagined events that are rich and detailed are more likely to get
mistaken for being real.
C. Stereotypes can indeed lead us to imagine stereotypic behaviour which

we later mistake for real behaviour.
D. People are more likely to create false memories if they are also asked to
imagine the fabricated events and if they are good imagers.
E. IMPORTANT. People may come to hold false memories of abuse as well,

and some therapists may have provoked the creation of false memories
by their clients.
1. Implications include it now being di cult or nearly impossible for us

to determine what had really transpired in an abuse case.
2. Recovered memories of abuse are not su cient anymore for courts

to convict anyone, with the lack of additional evidence.
62
ff
fi
fi
ffi
ffi
fl
ff
XII. Monitoring External Sources
A. External source monitoring is more likely to break down as the similarity

among di erent sources increase.
B. People are particularly likely to misattribute one person’s statements to

another if the 2 belong to the same social group.
C. People who view race as more meaningful for categorisation may be

more likely to fall prey to within race confusions. 
This does not hold true for people who are extremely low in prejudice. 
Highly prejudiced people may be particularly likely to make within race
confusions because it is easier to mistake one for another.
D. Sleeper E ect. Phenomenon where attitude change provoked by a

discredited communication is greater after a delay than it is
immediately. 
Seen as a dissociation between the message and the information that
discredited it, due to a failure at source monitoring. 
E.g. smoking company representative says that there is evidence that
cigarette smoking is neither harmful nor addictive. On the spot you may
not believe him, but after a week or so when you recall the statement
without remembering exactly where you had heard it, you do not recall
the credibility.
1. Exact underlying mechanism remains unclear.
2. BUT, it is clear that the sleeper e ect results from a failure of source
monitoring.
63
ff
ff
ff
Week 8 Readings (STEREOTYPES)

Daniel T. Gilbert & J. Gregory Hixon 1991. The Trouble of Thinking: Activation and
Application of Stereotypes.
I. 2 Studies investigating the e ects of cognitive busyness on activation and

application of stereotypes.
A. Background 
Stereotypes are forms of information and as such, are thought to be stored in
memory in a dormant state until they are activated for use.  
Some mental operations require very little e ort or intent, e.g. word
meanings spring to mind when their written referents are encountered, and
even complex beliefs about others can be activated without one’s awareness
(Bargh & Pietromonaco, 1982; Brewer, 1989; Devine, 1989; Lewicki, 1985).
Despite the ease with which such phenomena seem to occur, none of these
operations is unconditionally automatic in that it can occur in the complete
absence of intention, volition, awareness, or processing resources (Bargh,
1989). 
Automaticity of stereotype activation is also conditional (Bargh, 1989), and
that mere exposure to a stereotype object is therefore insu cient to activate
the corresponding stereotype.
B. Experiment I
1. Subjects and Method 

-71 female students at University of Texas participated to ful ll
requirement in their introductory psychology course.  
Arrival at the laboratory, subjects were greeted by a male experimenter
who escorted them to an individual cubicle was equipped with a video
monitor and tape recorder. Experimenter explained that he was testing
hypothesis that people are capable of performing 2 tasks simultaneously
as long as 2 tasks involve di erent cerebral hemispheres.  
-Subjects in the busy condition were told that they would be asked to
perform simultaneously a verbal and nonverbal task, and the subjects in
the not busy condition were told that they had been assigned to a control
condition and would therefore be performing only the verbal task. 
-Word fragment completion test. Verbal task which all subjects had to
perform required the subject to observe a word fragment (e.g. P_ST) and
then to generate its completions, e.g. POST, PAST, and PEST. Subjects
were told they would see videotape in which a female assistant would
hold up a series of cards, on each of which would be printed a fragment.
Subject had to read the fragment, generate as many completions as
64
ff
ff
ff
ffi
fi
possible during the 15s where the card was displayed, and state each of
these completions into a tape recorded. 
-Of 19 word completion trials, 5 were critical and 14 were llers. 20
Caucasian students were asked to “list all the words that come quickly to
mind when you think about Asian American students”, and of the words
generated, 8 concepts (i.e. words and their close synonyms) were
independently generated by at least 33% of the Caucasian pretest
subjects. Each concept was designated by a single word e.g.
smart=intelligence.
2. Independent Manipulations 
Subjects saw a silent videotape in which female assistant turned over a
series of 19 cards, each of which bore a word fragment. Half the subjects
saw the card turner being Caucasian, and the remaining subjects saw a
videotape in which the assistant was Asian. Except for the assistant’s
ethnicity, both tapes were identical.  
-Some subjects were cognitively busy while they watched the videotape,
and others were not.
3. Results 
-Not-Busy subjects were more likely to generate stereotypic
completions when exposed to an Asian than a Caucasian assistant,
but busy subjects were not. 
-Busyness did not itself impair any obvious aspects of task
performance. 
-Busy and Not-Busy subjects showed equally good recall of the
assistant’s race, and of the colour in which the word fragments were
printed. This means that cognitive busyness did not prevent subjects
from performing well on the completion task or from noticing the
assistant’s race, but that it did inhibit the activation of their
stereotypes about Asians.
4. Caveat 
Busy subjects performed equally well as Not-Busy subjects on test
indices and made equally few errors on the digit rehearsal task itself. 
Osborne and Gilbert (1990) showed that subjects who were given 20s to
memorise an eight digit number responded more slowly to probes that
occurred over the following 2 minutes than did subjects who had not
been asked to memorise the number. This suggests that typical subject
does indeed rehearse the number, rather than merely store it in long
65
fi
term memory, and that this rehearsal does usurp processing
resources.
C. Experiment 2 
To nd out that… since busyness may decrease likelihood of locating the
tool (activation), it may increase the likelihood of using the tool once it has
been found (application).
1. Subjects and Procedure 

-111 female students from same University.  
-Experiment composed of 2 phases, an initial activation phase (in which
subjects’ stereotypes about Asians either were or were not activated) and
a subsequent application phase (in which subjects were given an
opportunity to use their stereotypes during impression formation). 
-Application phase, half of the subjects were assigned to be busy during
this phase. Told that while they listened to the assistant describe a typical
day in her life, one of four letters (R, S, T, or U) would appear at a random
location on the screen. Screen was divided into a 6x6 invisible grid, and a
letter appeared in one of the 36 sectors on a black background. Basically,
one quarter participants were assigned to always busy condition (digit
rehearsal task during activation phase and visual search task during
application phase), one quarter to the never busy condition, one quarter
to the early busy condition, and one quarter to the late busy condition.
2. Dependent Measures 
-Female narrator describe a rather mundane series of events. After
hearing description, subjects were given 90s to complete their ratings of
the assistant. Subjects rated the assistant on 9 trait dimensions, timid,
intelligent, calm, composed, aloof, sociable, friendly, happy, and
conversational. These trait terms were either synonyms or antonyms of
the 8 stereotypic concepts that were generated by at least 33% of pretest
subjects in Experiment 1. 
-First 5 of these words were synonyms and thus considered typical of
Asian American students and the last 4 were antonyms and thus
considered atypical of Asian-American students. Each trait was
represented on an 11 point scale anchored at the end points with phrases
not a very X person… 
-Subjects given 13 item recognition memory test. 5 items were taken
verbatim from the assistant’s description, and remaining 8 items were
foils created by altering key phrases contained in the assistant’s actual
statements. For e.g. if the assistant actually said “I live by myself in a
small campus apartment”, from the statement a foil item was created that
66
fi
read “I live in a private campus dormitory”. Every e ort was made to
create foils that could be easily confused with statements that the
assistant had actually made.
3. Results and Discussion
a) Activation Phase 
-Subjects who were not busy during the activation phase were
more likely to generate stereotypic completions when exposed to
an Asian than a Caucasian assistant. 
 
-However, subjects who were busy during the activation phase
generated equivalent numbers of stereotypic completions
regardless of the race of the assistant to whom they were
exposed. 
 
-Busyness itself did not alter general task performance in any
meaningful way. Subjects who were busy and who were not busy
during activation phase generated equal numbers of correct
completions across all trials, and across critical trials. 
 
-Busy and Not-Busy subjects were equally adept at recalling the
assistant’s race, and at recalling the colour in which the
fragments were printed. 
 
-Although busy subjects did make marginally more common
words across all trials than did not-busy subjects, the 2 groups
made equally common words on the critical trials.
b) Application Phase 
-Subjects who were not busy during the activation phase but who
were busy during the application phase made more stereotypic
ratings of the Asian assistant than of the Caucasian assistant. 
 
-Busyness during the application phase increased subjects’
tendency to view the Asian assistant in stereotypic terms, but
only if the corresponding stereotypes had been activated in the
rst phase. 
 
-Busyness during the application phase had no discernible e ect
67
fi
ff
ff
on subjects whose stereotypes were not activated earlier (always
busy subjects). 
 
-Stereotype activation had no discernible e ect on subjects who
were not busy during the application phase. 
 
-Possible that the 3 groups of subjects who showed no evidence
of stereotype application may have done so for very di erent
reasons. Early-busy and always-busy subjects should not have
applied their stereotypes because busyness during the rst phase
should have kept those stereotypes from being activated. Never-
busy subjects on the other hand, should have had their
stereotypes activated, and evidence from the word fragment
completion test suggests that they did, yet they showed no
evidence of stereotype application. 
 
-2 explanations for the aforementioned nding: the behavioural
suppression account suggests that stereotype application did
occur, and the individuation account suggests that it did not.
Devine (1989) argued that both activation and application of
stereotypes are automatic, but overt responses are not. Now, we
know no mental process is unconditionally automatic Bargh
(1989), Devine’s point about greater controllability of stereotypic
behaviour is an important one. 
 
-Individuals may suppress stereotypic responses either because
they consider such responses immoral, or because they wish to
manage their impressions. Suggests that never-busy subjects
may well have achieved stereotypic impressions of the assistant,
but the lack of busyness during the application phase allowed
them to adjust their responses so as not to be or appear
prejudiced. 
 
-Alternatively, it is possible that activated stereotypes of never-
busy subjects did not a ect their judgements because the surfeit
of cognitive resources during application phase enabled them to
individuate (rather than stereotype) the assistant (Fiske &
Neuberg, 1990). 
68
ff
fi
ff
ff
fi
 
-Principle of cognitive economy suggests that stereotypes are
applied in order to ease the burden of information processing,
and thus one should not expect to observe such application when
the information processing task is not particularly taxing. 
 
-2 groups of subjects had their stereotypes of Asian Americans
activated (i.e. the never busy and late busy subjects who were
exposed to an Asian assistant). Those who were busy during the
application phase (i.e. late busy subjects) showed a marginally
reliable correlation between degree of stereotype activation and
application. Those activated subjects who were not busy during
application phase (i.e. never busy subjects) showed no such
correlation. 
 
-We are unable to ascertain whether the never busy subjects
failed to achieve stereotypic impressions (deindividuation
account) or simply failed to announce them (the behavioural
suppression account).
4. General Discussion.
a) Mental experience seems inexorably to involve the admixture of

old and new information. People rely strongly on prior information
to ease the burden of ongoing perception that some of the
greatest thinkers (e.g. Kant, 1781/1965) have doubted whether
perception could occur otherwise.
b) Social Interaction and Stereotypy
(1) Stereotypes are psychologically fundamental, but they may

also be socially pernicious, and this is a dilemma yet to be
solved. One method has been to encourage people to spend
time and e ort necessary to individuate others rather than
allowing their preconceptions to dominate their judgements
(Fiske, 1989). Second remedy has been to increase the
accuracy of the preconceptions upon which people rely.
(2) When people interact with members of out groups, 2 things

may happen. Firstly, they may gain information that increases
accuracy of their preconceptions (e.g. an unmusical Black, a
69
ff
generous Jew, or sober Irishman may undermine the certainty
with which people embrace such racial caricatures (Taylor,
1981; Weber & Crocker, 1983)). 
Secondly, social interaction raises practical cost of inaccurate
beliefs (Swann, 1984). It is one thing to misconstrue a famous
Black politician or feminist opinion leader, and quite another to
misconstrue one’s dentist, student, or daughter in law. People
are more accountable for and thus may craft more carefully,
judgements about those with whom they have true and
enduring commerce. (Tetlock & Kim, 1987).
(3) The current experiment suggest a 3rd way in which social

interaction may a ect stereotype use. Social interaction
requires one to consciously regulate one’s own actions at the
same time that one draws inferences about others (Gilbert,
Krull, & Pelham, 1988). This may cause interactants to become
cognitively busy and thus reduce the likelihood that their
stereotypes about each other will be activated. 
If stereotypes are activated prior to a resource consuming
social interaction, then the interactants may be especially
likely to view each other in stereotypic terms. Timing of the
onset of busyness would appear to be critical in determining if
and when social interaction will ameliorate or exacerbate
stereotypy.
c) Inevitability of Stereotypy
(1) Activation of racial stereotypes is not an unconditionally

automatic consequence of exposure to a person. Rather, a
perceiver must have adequate processing resources for such
stereotypes to be activated. 
Along with (Brewer, 1989), the researchers believe that mere
presentation of a stimulus person does initiate certain
classi cation processes; however, we do not believe that
these classi cations are inevitably along certain dimensions or
that they are una ected by the perceiver’s goals. In fact,
busyness may exert its e ect on stereotype activation by
causing subjects to classify others only along those
dimensions that are directly relevant to their current
information processing goals. 
70
fi
fi
ff
ff
ff
When the person’s race is wholly inconsequential (as it was in
this study), busy perceivers may not have the luxury of
activating pre-existing information that is, in fact, irrelevant to
their concerns.
(2) Studies may not be accurate or obscure these issues by

exposing subjects to words that represent stereotype objects
rather than stereotype objects alone. (Devine, 1989).  
May be virtually impossible for a literate adult to read the
phrase Black reman without experiencing activation of both
the racial and occupational constructs. (Logan, 1980; Stroop,
1935; cf. Kahneman & Treisman, 1984).  
It may also be entirely possible for a literate adult to encounter
a Black reman and given the appropriate information
processing goal (e.g. to nd quickly someone who will enter a
burning building to save a child), to construe the Black reman
only in terms of his occupation (Taylor, 1981). 
Words and phrases contain implicit categorisations of the
objects they describe, they e ectively demand that associates
of these categories be activated. 
Linguistic descriptions (e.g. a black person), force a single
categorisation, in contrast to a real person, who is not only
black, but perhaps young, male, well-dressed, tall, speaking
with a Southern accent and so on. (Zarate and Smith, 1990). 
 
 
STEREOTYPES (II)
C. Neil Macrae, Alan B. Milne, & Galen V. Bodenhausen The Dissection of

Selection in Person Perception: Inhibitory Processes in Social Stereotyping
1995
I. Background Information
A. Basic Mechanisms in Category Activation. 

People are the most complex stimuli we encounter, in part because they
belong to multiple social categories at the same time. These given social
categorisations facilitate the investigation of stereotyping in a controlled and
rigorous manner, failing to speak to the more fundamental question of how
perceivers deal with a target when competing categorisations are available.  
 
71
fi
fi
fi
ff
fi
Contemporary views of human cognition have characterised the social
perceiver as a cognitive miser, and have linked the tendency to engage in
social stereotyping to precisely those conditions in which perceivers need to
or want to economise mental e ort. This experiment hence assumes that
simple category dominance is a default preference of the information
processing system. 
 
Several factors can be identi ed that determine which of the possible
competing categories will come to dominate social impressions. These are:
relative accessibility or salience of a particular categorisation, perceivers’
current processing objectives, and perceivers’ levels of prejudice toward
particular social groups. These factors and others, come together to drive
the categorisation process, determining when one category rather than
another assumes dominance in mental life. What existing research fails to
detail is the cognitive mechanisms through which these categorisation
e ects are realised.
B. Inhibitory Mechanisms in Mental Life 

The concept of selectivity answers how humans are able to form coherent
representations of the stimuli we encounter in our daily interactions with a
complex world. Equally important, are the processes through which other
stimuli are denied access to awareness or higher cognitive processing. 
 
In this experiment, an inhibitory framework was utilised in an attempt to
inform our understanding of the processes involved in stereotype activation. 
E.g. consider the case of social perceiver who encounters a Chinese woman
(a multiply categorisable target, “Chinese” and/or “woman”). When a
perceiver encounters a target, both applicable categories are activated in
parallel and that a competition for mental dominance ensues. Category
salience, accessibility, and goal relevance are all factors that would confer an
activational advantage in such a competition. Once a given categorisation
(e.g. Chinese) achieves su cient activation to win the race, a critical
question concerns what happens to the loser during older views of selective
attention, is that following initial activation of both categories, the loser is
simply ignored, with construct activation gradually decaying to its baseline,
or resting state. Inhibition models make quite di erent claims. In the race for
category dominance, these models suggest that the loser is not simply
neglected, but actively dampened through a spreading inhibition process,
even to levels below resting state. The advantage of active inhibition of
alternative categorisations lie in the suppression of potentially distracting
(hence disruptive) mental representations. 
72
ff
ffi
fi
ff
ff
 
The experiment explored the extent to which inhibitory mechanisms in
selective attention may contribute to our understanding of the process
of stereotype activation.
II. Study I
A. Method 
Participants performed 3 ostensibly unrelated experiments. First was billed
as a vigilance task, reporting whether a rapidly presented ash appeared to
the left or right of a xation point in the centre of a computer screen. In
reality, this was a parafoveal priming task, in which either the category
woman or the category Chinese was primed. 
Next phase of study, with the instruction to check the edit quality of a
videotape, participants watched a short videotape of a Chinese woman
reading a book. It was anticipated that prior priming experience would
in uence categorisations of the target during this task. Participants who were
previously primed with the category of woman would activate the category
woman and inhibit the category Chinese, those primed with the category
Chinese would activate the category Chinese and inhibit the category
woman. 
Third phase of study, in an ostensibly unrelated word identi cation task,
these predictions were tested by measuring the accessibility of applicable
stereotypes about the target (i.e. woman and Chinese).
B. Participants and Design 

36 female undergraduates at Cardi University, 3 x 2 mixed design with
repeated measures on second factor. (prime: woman or Chinese or control)
and (trait type: woman or Chinese).
C. Procedure and Stimulus Materials 

Participants were greeted by female experimenter and randomly assigned to
one of the priming conditions (i.e. woman, Chinese or control). Experimenter
then explained that she required the participant’s assistance in the piloting of
a number of tasks for future departmental research projects. Students in
category priming conditions were seated facing the monitor of an Apple
Macintosh micro-computer and instructed that they would be performing a
vigilance task. It was in reality a parafoveal priming manipulation that was
based on the procedure used by Bargh and his colleagues. During this task,
students were instructed to direct their gaze at all times to a small xation
point that was located in the centre of the computer screen. Task was simply
to indicate, by means of a keypress, whether ashes on the computer screen
appeared to the left or right of the central xation point. Participants were not
73
fl
fi
ff
fi
fl
fl
fi
fi
informed that ashes were actually words presented for 75ms. 
 
Critical element of this parafoveal priming task is that participants are unable
to report the identity of the priming stimuli, and this was facilitated through a
number of task related procedural features. First, timing and location of the
priming stimuli are made unpredictable to preclude anticipation of the
appearance of the next stimulus presentation. Second, priming stimuli are
presented very brie y, then immediately overwritten by a consonant mask.
Third, stimuli are presented outside the foveal eld.  
 
10 students were presented with 48 stimulus presentations (i.e. 5-7 letter
words). Students were informed that ashes on computer screen were
words. Their task was to simply guess the identity of each word following its
presentation. Participants asked to maintain their gaze on xation spot, given
the unpredictable location of each word presentation. Participants were told
that they must make a guess for each word presented, even a blind guess if
necessary. 
 
After completing vigilance task, students in the category priming conditions
performed an ostensibly unrelated activity. Experimenter explained that
psychology department was compiling videotaped materials for use in future
research projects, and student’s task was simply to check the edit quality of
a short extract from one of the tapes. Extract was a 15 second clip depicting
a Chinese woman sitting in a chair reading a book. Chinese woman was
used as the experimental target for a number of reasons. This satis ed the
requirement that she be multiply categorisable. Second, in the student
sample, there was no well de ned subtype of Chinese women.  
 
Following videotape, previously primed participants performed an ostensibly
unrelated lexical decision task. Each student was seated facing the screen of
the copmuter and infromed that on the presentation of a letter string in the
centre of the screen, she had to indicate whether it was a word or nonword.
Of 16 words and 16 nonwords, 8 were ller words, 4 were traits stereotypic
with respect to women but not to Chinese, and 4 were stereotypic with
respect to Chinese but not with women.
D. RESULTS AND DISCUSSION
1. Participants primed with the category Chinese responded more

quickly to traits associated with the Chinese than with women.  
Participants primed with the category women responded more
quickly to traits associated with women than with the Chinese.
74
fl
fl
fi
fl
fi
fi
fi
fi
2. Strong evidence for operation of facilitatory and inhibitory
mechanisms in the categorisation process. 
Simple e ects analyses revealed an e ect of the priming category on
both sets of traits. 
On Chinese traits, relative to control condition, participants’
responses were facilitated when the previously primed category was
Chinese but inhibited when it was woman. Relative the control
condition, participants’ responses were facilitated when previously
primed category was woman but inhibited when it was Chinese.
3. The operation of both excitatory and inhibitory mechanisms in

categorisation process exist. Categorisation of the target was
directed by the previous priming experience.
4. Supports the dual process interpretation. Competing representation

is actively suppressed instead of passively ignored.
5. It is conceivable that results may simply derive from participants

parafoveal priming experiences and not interplay of excitatory and
inhibitory processes during target presentation phase. 
No way of knowing whether it was priming manipulation alone that
produced observed e ects, rather than the hypothesised
competition for selection during the video monitoring task.
III. Study II
A. Method 
Same general procedure used in Study I. All that di ered was the content of
the videotape used in edit quality task. On this occasion, all students
observed the 15s clip from a wildlife show was used in the control condition
in Study 1.
B. Results and Discussion
1. Participants primed with the category Chinese responded more

quickly to traits associated with the Chinese than with women. 
Participants primed with the category woman responded more
2. On Chinese traits, participants’ responses were facilitated when the

previously primed category was Chinese, response times in this
condition was signi cantly faster than in either of the other two
conditions. However, response times in the control and woman
75
ff
fi
ff
ff
ff
prime conditions did not di er signi cantly from one another, thereby
revealing no evidence for the inhibition of the unprimed category. 
Similar pattern emerged for woman traits.
3. E ects observed in prior study not simply the consequence of

participants parafoveal priming experiences. When students were
primed with the category Chinese or woman but were not given an
opportunity to observe the target (i.e. the Chinese woman), quite
di erent e ects emerged. Straightforward facilitatory priming e ects
were observed, such that participants’ responses were enhanced
when the words in the lexical decision task were stereotypic with
respect to the previously primed category.
4. Excitation and inhibition, cognitive mechanisms, exist. When

perceivers are confronted with a multiply categorisable target, the
outcome of the categorisation processes is determined by the
interplay of both excitatory and inhibitory attentional processes.
IV. Study 3 
Establish the generality of e ects observed in Study 1 by manipulating category
dominance in a more naturalistic manner. To achieve this, we have decided to
manipulate category dominance in a more naturalistic manner.
A. Method 
Undergraduates performed 2 ostensibly unrelated experiments. First, was a
monitoring task in which they checked the edit quality of a short videotape,
depicting a Chinese woman either eating noodles from a bowl with a pair of
chopsticks or putting on makeup by a mirror.  
In the second phase, an ostensibly unrelated word identi cation task,
measured the accessibility of applicable stereotypes about the target.
B. Participants 
24 female undergraduates
C. Procedure and Stimulus Materials 

Participants seat facing a TV monitor and instructed to check the edit quality
of a short extract from a videotape. Chinese woman eating noodles from
bowl with chopsticks or applying makeup by a mirror. 
Hypothesised that stereotypic congruence of target’s behaviour would drive
categorisation process. Chopsticks = Chinese, Makeup = Woman.
D. Results and Discussion
76
ff
ff
ff
ff
ff
fi
fi
ff
1. Participants who watched videotape of target eating with chopsticks
responded more quickly to traits associated with the Chinese than
with women. 
Participants who watched the target eating with chopsticks
responded more quickly to traits associated with Chinese than with
women. 
Participants who watched target applying makeup responded more
2. Firm evidence for operation of facilitatory and inhibitory processes in

category activation.
3. On the traits characteristic of women, relative to control condition,

participants responses were facilitated when the previous videotape
depicted the target applying makeup but were inhibited when she
was shown eating noodles with chopsticks. 
Similar pattern emerged on traits characteristic of the Chinese.
4. Category activation and inhibition were driven by more naturalistic

forces, namely, the target’s expressed behaviour. Stereotypic
congruence of the target’s behaviour is used as a cue to guide the
category selection process. When the behaviour could readily be
identi ed with one of the target’s categories, construct excitation
was ampli ed, thereby increasing the accessibility of stereotypic
information associated with that category. 
 
Converse occurred for the unprimed or competing category (i.e.
woman). In that case, category was actively suppressed, resulting in
a signi cant reduction in the accessibility of stereotypic material.
Inhibitory mechanisms serve a central regulatory function in the
category activation process, at least when competing
categorisations are readily available.
V. ACTIVATING SOCIAL STEREOTYPES
1. Everyday interaction demands a great deal of attentional selectivity.

A dual mechanism attentional system facilitates this end.
2. Simultaneous operation of excitatory and inhibitory processes

enables perceivers to amplify a target signal while suppressing a
distractor, thereby enhancing their ability to pull the competing
representations (i.e. signal and noise) apart. 
77
fi
fi
fi
The goal of an inhibitory mechanism, is one of noise reduction,
eliminating or reducing interference from competing or distracting
representations. Through this, single action systems acquire mental
dominance and perceivers achieve a unitary focus of attention.
3. Humans prefer social categorisation as a way to navigate capacity

limited processors in a world of overwhelming complexity.
4. This experiment showed clear results and evidence for the operation
of both excitatory and inhibitory mechanisms in stereotype
activation. These suggest the desirability of extending contemporary
models of social stereotyping, in particular their treatment of the
initial categorisation process.
VI. CONJUNCTIVE STEREOTYPING AND SUBTYPING
1. Our impressions of others sometimes re ect an awareness (and

in uence) of more than on categorisation cue. For example,
reactions to an adolescent Latino may be quite di erent from
reactions to an elder Latino, an adolescent Latina, or an adolescent
male of Anglo Saxon heritage. The conjunction of social categories
is crucial more so than a di erential emphasis on age, gender, or
ethnicity cues singularly.
2. When particular grouping combinations are frequently encountered,

distinctive stereotypes concerning such subtypes can develop.
Subtypes of this sort, however, do not simply re ect the joint
activation of more than one social category. (Smith & Osherson, 1984).
The use of a subtype still a ords the social perceiver a relatively
streamlined, economical impression without the necessity of
activating multiple, diverse sets of stereotypes. The principal
di erence, in this case, is simply that the activated subtype applies
to a more circumscribed population than do more encompassing
global categorisations.
VII. CONCLUSIONS
1. When perceivers identify a target, they identify several competing

superordinate categories to which he or she belongs. As a
consequence of repeated exposure to these categories, moreover,
we consider this process to be automatic and cognitively
impenetrable. The next stage, target categorisation, we believe to be
78
fl
ff
ff
ff
fl
fl
ff
considerably more exible. The experimenters accept that the
process often proceeds in a largely automatic manner. It can also,
under certain circumstances, be under perceivers’ executive control. 
Default categorisation may invariably be an automatic one, where
excitatory and inhibitory components of the categorisation process
may be triggered by factors residing outside of perceivers’
awareness.
2. In the same way that automatic inhibitory processes contribute to

our capacity to see complex patterns of motion, initiate complex
actions, comprehend written text, and select relevant rather than
irrelevant objects from a stimulus array, they may also facilitate our
ability to categorise people when competing classi cations are
available. This ensures exibility when engaging in social
interactions. 
Dissection of stereotype selection reveals operations of 2
fundamental processes, excitation and inhibition.
3. Inhibitory processes in stereotype activation may play a bigger role

than excitatory processes than we previously believed.
STEREOTYPES (III)
Jennifer Crocker, Kristin Voelkl, Maria Testa, and Brenda Major - Social Stigma: The
A ective Consequences of Attributional Ambiguity (1991)
I. Background Information 
-2 experiments investigated the hypothesis that the stigmatised can protect
their self-esteem by attributing negative feedback to prejudice.  
-Many social groups are stigmatised in America. E.g. Blacks, women, etc.
Although most people do not characterise these groups in uniformly negative
terms. 
-It is well documented that members of these groups are relatively
disadvantaged in American society in economic or interpersonal outcomes. 
Concern in this article is with the consequences of prejudice and discrimination
for the global personal self-esteem of members of these groups. This is because
self-esteem is widely recognised as a central aspect of psychological
functioning and is strongly related to many other variables, including general
satisfaction with one’s life (Diener, 1984).  
79
ff
fl
fl
fi
-Several social psychological theories predict that prejudice and discrimination
against members of stigmatised groups will result in lowered self-esteem and
diminished self-concept for the stigmatised. (cf. Cooley, 1956; Mead, 1934). 
-Research and theory on self-ful lling prophecies are also consistent with
predictions that social stigma lead to low self-esteem. These occur when a
perceiver acts on initially false beliefs in such a way that those beliefs come to
be con rmed by the behaviour of the target. (Merton, 1948). 
-However, evidence that members of stigmatised groups have lower self-esteem
than the nonstigmatised in society is remarkably scarce. Blacks have levels of
global personal self-esteem equal to or higher than that of Whites (Hoelter, 1982;
Porter & Washington, 1979; Rosenberg, 1979; Rosenberg & Simmons, 1972;
and Wylie, 1979). Women also do not have self-esteem lower than men.
(Maccoby & Jacklin, 1974; Wylie, 1979). 
-Several mechanisms bu er the self-esteem of members of stigmatised or
oppressed groups from the prejudice of others. (Crocker & Major, 1989). One
such mechanism is attributing negative feedback or relatively poor outcomes to
the prejudiced attitudes of others toward their group. (Wright, 1960; Kleck and
Strenta, 1980).
II. Experiment 1
A. Method 
59 women from introductory psychology classes.
B. Procedure 
Individual participation of subjects. Instructed by means of tape that the
experiment involved attitude assessment and that they were interested in
what kinds of attitudes college students hold in how they respond to other
people’s attitudes. 
Subjects were led to believe that they would be writing an essay expressing
their opinion on a current topic, which would be critiqued by a second
subject, seated in a di erent room in the laboratory. 
Subjects were rst asked to complete an attitude survey, 25 item survey
asking about subjects’ attitudes towards a variety of current issues, including
drugs, national defence and homosexuality. 5 items on the survey assessed
subjects’ attitudes regarding women’s roles in society. 
Essay afterwards to be written, to convince others of the correctness of their
views. 
Subjects told to evaluate other people’s essays objectively rather than to
criticise the other person’s point of view. Led to believe that each of them
would have the chance to critique the other’s essay, but, for the rst session,
the other ctional subject would evaluate the actual subject’s essay. 
80
fi
fi
fi
ff
ff
fi
fi
Subjects then shown completed survey by ctitious men on di erent
conditions. Prejudiced vs. non-prejudiced evaluator conditions. 
Afterwards, subjects heard the other subjects’ voices (prerecorded) over the
audio system. In the positive feedback conditions, the other subject
indicated that he thought the essay was clear and persuasive and that the
arguments were well presented. In the negative feedback condition, the other
subject stated that essay was weak, arguments were ignored, and that it was
not persuasive. 
Subjects had to complete the Rosenberg Self-Esteem Scale, short mood
scale, and rated the other subject on 6 x 7 point scales.
C. RESULTS
1. Subjects in positive feedback condition thought their essay had been

evaluated more positively than did negative feedback subjects.
2. Subjects in negative feedback conditions, the evaluation was

signi cantly more likely to be attributed to the evaluator’s attitudes
toward women if he had expressed attitudes against women’s
changing roles in society than if he had expressed liberal attitudes
for women’s changing roles in society.
3. Subjects in positive feedback conditions, the evaluation was equally

likely to be attributed to his attitudes toward women, regardless of
whether the evaluator expressed favourable or unfavourable
attitudes toward women’s changing roles.
4. Subjects who received negative feedback experienced more

depressed a ect if the evaluator had favourable attitudes toward
women’s nontraditional roles than if he had unfavourable attitudes.
5. Subjects who received positive feedback experienced the same level

of depressive a ect regardless of whether the evaluator had
favourable or unfavourable attitudes towards women’s changing
roles.
6. Main e ect on feedback on hostility. Subjects who received negative

feedback were more hostile than were those who received positive
feedback.
7. Self-esteem tended to drop whenever negative feedback was

received from a non-prejudiced evaluator and rose in all other
conditions.
81
fi
ff
ff
ff
fi
ff
8. The evaluator was rated more negatively in the negative feedback
than in positive feedback conditions.
D. DISCUSSION
1. Attributing negative outcomes to the prejudice of others has self-

protective consequences. Subjects appeared to use the discounting
principle and attributed negative feedback from a prejudiced
evaluator to the evaluator’s prejudice. Subjects also experienced
less depressed a ect after receipt of negative feedback if they had
previously learned that the evaluator had negative attitudes towards
women.
2. E ects of evaluator prejudice on mood appeared to be speci c to

depression.
3. All subjects who received negative feedback apparently felt angry,

but only those who could not attribute the feedback to prejudice
internalised the feedback.
4. Although non-signi cant, trend was that subjects had non-

signi cantly lower depressed a ect and higher self-esteem when
they received positive feedback from evaluators who were not
prejudiced, compared with evaluators who were prejudiced.
III. Experiment 2
A. Similar to prior study this time done on Black college students receiving
either positive or negative feedback from a White peer.
B. Method 
38 Black and 45 White students.
C. Procedure 
Individual participation. Similar setup to prior experiment.  
Asked to ll out Rosenberg Self-Esteem Scale, self-description form that
included questions about the subjects’ likes and dislikes, strengths and
weaknesses, personal qualities. Subjects were also told that their answers to
these questions would be shown to the other student, who would use this
information to determine whether the 2 of them could become friends. 
Random assignment to one of two conditions, either blinded or non blinded
conditions. 
Subject shown either favourable or very unfavourable response from the
other subject. 
82
ff
fi
fi
ff
fi
ff
fi
Subject completed more measures of positive and negative mood after
receiving the feedback, and asked to indicate the degree to which various
factors had in uenced the other student’s response to them.
D. RESULTS
1. Subjects in positive feedback conditions thought the evaluator liked

them more than did subjects in the negative feedback conditions.
2. Black students attributed the feedback more to prejudice than did

White students.
3. Subjects made greater attributions to prejudice when they received

negative feedback than when they received positive feedback.
(Quali ed by interaction with race).
4. Subjects made greater attributions to prejudice when they thought

the other student could see them than when they thought the other
student could not see them. (Quali ed by non-signi cant interaction
with race).
5. Subjects indicated feedback was due to their own personality more

when the feedback was positive than when it was negative.
6. Black subjects were more likely to attribute the feedback to their

personality when they could not be seen than when they could be
seen. This e ect was nonsigni cant and in the opposite direction for
White subjects. Black subjects were more likely to attribute both
positive and negative feedback to their personality when the blinds
were down.
7. For White subjects, the FEEDBACK x VISIBILITY interaction was not

signi cant, but for Black subjects, the interaction was signi cant.
Black subjects have their self-esteem protected when the blinds
were up from negative feedback, in relation to the blinds down
conditions.
8. All subjects rated the other subject signi cantly more negatively
after negative feedback than after positive feedback, this e ect was
more pronounced for Black subjects than for White subjects.
9. Black subjects thought the other subject was more racist than did
White subjects.
83
fi
fi
ff
fl
fi
fi
fi
fi
ff
fi
10. Black students were signi cantly more likely to rate the evaluator as
racist after negative feedback than positive feedback, whereas White
students were less likely to show this e ect for positive and negative
feedback.
11. Subjects in positive feedback conditions experienced more positive

mood than did subjects in the negative feedback conditions.
12. Subjects experienced more negative mood after negative feedback

than after positive feedback.
E. DISCUSSION
1. Black subjects were more likely than White subjects to attribute the
feedback they received to prejudice when they received negative
rather than positive feedback and when the evaluator could see
them, hence was aware of their race, than when they were not seen.
2. Black students were somewhat less likely to believe that the

feedback was due to their personality if they could be seen by the
evaluator than if the evaluator was blind to their race. Black students
apparently tend to discount interpersonal feedback from White
evaluators, especially when they know that the White evaluator can
see them, hence is aware of their race. 
In contrast, White subjects were not signi cantly a ected by either
the valence of the feedback or whether the evaluator could see
them.
3. Data supports hypothesis that Black students exist in a state of

attributional ambiguity regarding causes of feedback, both positive
and negative that they receive from their White peers. Only when
positive feedback was received from a White evaluator who was
unaware of their race did Black students indicate that the feedback
was not at all due to prejudice.
4. Receiving negative feedback or being evaluated when one’s race is

know to the evaluator apparently raises the possibility that prejudice
was a factor in the evaluation but by no means conclusively
demonstrates that it was a factor in the view of our subjects.
5. Attributional ambiguity appears to have self-protective

consequences for Blacks who received negative feedback. They
reported more positive a ect after negative feedback than did White
84
ff
fi
ff
fi
ff
subjects. Self-esteem was also protected by this attributional
strategy. When Black students received negative feedback and could
be seen by the evaluator, they were most likely to attribute the
feedback to prejudice, and their self-esteem did not su er. In
contrast when the evaluator could not see them, the Black students
were somewhat less likely to attribute the feedback to prejudice, and
their self-esteem tended to drop after negative feedback.
6. Black subjects discounted positive feedback only when they knew

the evaluator was aware of their race. When the blinds were up,
Black subjects believed the positive feedback had nothing to do with
their race, racism, or discrimination. Self-esteem fell when they
received positive feedback with blinds up but increased with blinds
down.
7. Blacks may be aware that Whites often respond more positively

toward them because they are Black.
8. Positive feedback for the stigmatised may only have positive

consequences for self-esteem when they are certain the feedback
re ects their deservingness, rather than some special consideration
or a fear of appearing prejudiced on the part of the non-stigmatised.
Blacks may sometimes prefer social transactions in which their race
is not known to others.
9. The valence of feedback received from another student and the

visibility of the subjects to the evaluator had stronger e ects on the
Black students than on White students.
F. GENERAL DISCUSSION
1. The stigmatised will attribute negative outcomes to prejudice against

their group when such an attribution is plausible (e.g. when they
know that the evaluator has negative attitudes toward the group and
when the evaluator is aware of their group membership). Discounting
of negative feedback appears to have self-protective consequences
for the stigmatised.
2. Neither studies provide evidence that stigmatised tend to augment

positive feedback from a prejudiced evaluator. Stigmatised may be
suspicious of the motives underlying positive feedback and that if
they believe such feedback is due to racism, their self-esteem will
85
fl
ff
ff
su er.  
This is inconsistent with work of Sigall and Michela (1976) and Major et
al. (1984), which showed that unattractive subjects tended to
augment positive feedback from opposite sex evaluator. Same
underlying process, evaluations from others are discounted when
one has reason to suspect the evaluator has ulterior motives.
3. Positive feedback may raise and negative feedback may lower self-
esteem only when one believes an evaluator has no ulterior motives
for giving feedbacks. Black students are in double jeopardy as White
students could have ulterior motives in giving them both positive and
negative feedback.
4. Blacks may have good reason to suspect positive feedback from

Whites is not genuine and re ects self-presentational concerns.
Women on the other hand, may not be as mistrustful of positive
feedback from men.
5. Members of stigmatised groups who generally believe that they are

discriminated against or that others are racist should be more likely
to attribute negative feedback to prejudice and therefore may be
higher in self-esteem. This is contrary to predictions of looking glass
self-perspective, suggesting that believing that others are prejudiced
against one’s group should result in lower self-esteem. However,
belief that one is discriminated against may undermine the value of
positive feedback if positive feedback is also discounted.
6. Self protective consequences of believing that others are prejudiced

may depend on whether one receives predominantly negative or
positive outcomes from the non-stigmatised.
7. Stigmatised individuals may sometimes attempt to manage their

interactions to encourage or discourage particular attributions for
their outcomes.
8. Attributional ambiguity in which the stigmatised exist may have

unanticipated negative consequences. Attributional ambiguity should
make it relatively di cult for the stigmatised to predict their future
outcomes and select tasks of appropriate di culty.
86
ff
ffi
fl
ffi
Week 9 Readings (AUTOMATIC

PROCESSING I)
AUTOMATIC PROCESSING I (I)
Kurt Hugenberg and Galen V. Bodenhausen. FACING PREJUDICE: Implicit

Prejudice and the Perception of Facial Threat 2003
A. Researchers proposed that social attitudes, in particular implicit prejudice,

bias people’s perceptions of the facial emotion displayed by others. To test
this hypothesis, the researchers employed a facial emotion displayed by
others, where European American participants detected the o set (Study 1)
or onset (Study 2) of facial anger in both Black and White targets.
B. Higher implicit, but not explicit prejudice was associated with greater
readiness to perceive anger perceptions regarding similar White faces. This
pattern indicates that European Americans high in implicit racial prejudice
are biased to perceive threatening a ect in Black but not White faces,
suggesting that the deleterious e ects of stereotypes may take hold
extremely early in social interaction.
C. The human face is central to social interactions and of primary importance in

social perception. Little research has investigated the in uence of social
attitudes on the decoding of facial a ect.
D. Stereotypes clearly in uence how people interpret the behaviour of others. 

Notable experiments include:
1. Duncan (1976) showed that ambiguous behaviour was interpreted more

negatively when performed by a Black actor than White actor.
2. Sagar and Scho eld (1980) found that ambiguously hostile behaviours
were rated as more hostile when performed by Black than White actors.
E. Most research involving facial a ect has used unambiguously emotional

faces that are rarely found in everyday interaction.
1. Ekman, Sorenson, & Friesen (1969).
2. Wehrle & Kaiser (2000).
II. Study 1
A. Method and Design
87
fi
fl
ff
ff
ff
ff
fl
ff
1. Participants and Design 
24 European American university students were used. Both implicit and
explicit measures of prejudice served as predictors of hostility o set in
White and Black targets, target race was manipulated on within subjects
basis.
2. Materials and procedure 

Stimuli consisting of faces generated using Poser 4TM 3D character
animation software which a orded control over each target’s facial
structure, expression, skin tone, and hair style and colour were used. This
permitted Black and White target faces to be matched precisely for both
facial structure and expression. 
Participants seated at computers at individual cubicles and instructed to
watch each movie and press the spacebar when they saw that the target
face was no longer expressing its initial emotion. Order of 4 target movies
(2 Black and 2 White) were randomised for each participant. 
Participants had to indicate how warmly or coldly they felt about each
group on a scale from 1-100, with higher responses indicating more
warmth. 
Participants performed the IAT, which was described as an ostensibly
unrelated word categorisation task. IAT consisted of 5 trial. blocks. First 2
were practice where participants learned to map White names to one
response key and Black names to another (the rst block) and to map
pleasant and unpleasant words to those same response keys (the second
block). Third block were compatible trials, where White names and
pleasant words were mapped to same response and Black names and
unpleasant words were mapped to another key. Fourth block was
learning new mapping. Last block was incompatible trials, where White
names and unpleasant words were mapped to the same key and Black
name and pleasant words were both mapped to another key. Implicit
prejudice was indicated to the extent to which performance on the
incompatible trials (i.e. Black-good, White-bad) were impaired relative to
performance on the compatible trials.
B. Results and Discussion (Study 1)
1. Longer response latencies indicated lingering perceptions of anger

in a particular face. Neither explicit prejudice nor its interaction with
implicit prejudice predicted any signi cant variance in the latency of
responses to the face stimuli.
2. Revealed that implicit prejudice scores were positively related to

response times for Black faces. Participants higher in implicit
88
ff
fi
fi
ff
prejudice indicated that hostility o set occurred later for Black faces
than did lower prejudice participants.  
However, implicit prejudice was unrelated to response times for
White faces.
III. Study 2
A. Background 
First study showed that anger was perceived to linger longer in Black faces
to the extent that viewers possessed greater levels of implicit prejudice.
However, it might be the case where slower response times were a result of
greater indecision or inhibited perceptual processing of Black faces among
more prejudiced persons. Experiment 2 examines perceptions of anger
onset. 
If perceptions of high prejudice respondents are coloured by social
stereotypes, then compared with low prejudice respondents, high prejudice
respondents should see anger emerge more quickly on Black faces and
therefore respond more quickly to the onset of anger in Black faces. If high
prejudice viewers are simply indecisive or do not process Black faces as
e ciently as low prejudice viewers, then they should be slower than low
prejudice viewers to respond to Black faces. 
Method of Experiment 2 virtually identical to Study 1, save that participants
were required to detect the onset of hostility in Black and White faces. 
If response times of high prejudice were faster than response times of low
prejudice respondents, this would suggest that Study 1 results were due to
in uence of social stereotypes in high prejudice participants. However, if
response times in high prejudice participants were slower than response
times of low prejudice participants, this would suggest that Study 1 results
were due to indecision or slower processing among low prejudice
participants.
B. Method, Participants, Design, Materials and Procedure. 

24 European American students, same materials and procedure.
C. Results and Discussion (Study 2)
1. Neither explicit prejudice nor interaction with implicit prejudice was

reliably associated with response latencies.
2. Implicit prejudice scores were inversely related to response times for

Black faces. Individuals high in implicit prejudice perceived onset of
hostility much earlier for Black faces than did low prejudice
89
ffi
fl
ff
participants. However, response times for White faces were
unrelated to implicit prejudice scores.
D. GENERAL DISCUSSION
1. Compared with individuals low in implicit prejudice, those high in

implicit prejudice saw hostility as lingering longer and appearing
more quickly on the faces of African Americans. Hence, stereotypic
expectancies appear to penetrate a fundamental aspect of on-line
person perception.
a) These ndings add to the evidence from social cognition research

showing that relatively low level cognitive processes such as
attention and encoding are subject to the e ects of stereotypes
and prejudice. (Macrae, Milne, & Bodenhausen, 1994; Sherman,
2001; von Hippel et al., 1995).
2. Explicit prejudice did not predict when Whites saw threatening a ect
in Black faces. Past research has shown that implicit measures
predict relatively automatic aspects of behaviour that occur outside
of conscious control, including spontaneous nonverbal behaviour. IN
contrast, explicit measures better predict more consciously
controlled behaviours such as speech. (Asendorpf, Banse, & Mücke,
2002; Dovidio et al., 2002; McConnell & Leibold, 2001).
3. Although current task involved decoding rather than encoding of

nonverbal signals, it nevertheless involved an on-line, dynamic
perceptual judgement that required a rapid assessment of a
changing stimulus. Therefore not surprising that a measure of the
implicit, relatively automatic aspects of prejudice predicted
performance better than the measure of explicit prejudice did.
Present ndings add to growing evidence that implicit measures
such as IAT have predictive validity in consequential domains of
social cognition.
4. Finding a relationship between implicit prejudice and decoding of

facial a ect not only is important but also holds powerful
implications for interracial interactions.
a) Perceptions of hostility may determine not only the perceiver’s

behaviour toward another person, but also, in the manner of the
90
fi
ff
fi
ff
ff
classic self-ful lling prophecy, the target’s behaviour toward the
perceiver.
b) Perceived hostility will at best promote avoidance, or worse, may

foster reciprocation.
c) High prejudice perceivers may be particularly likely to adopt

negative orientations in interactions in which a ective ambiguity
is present.
d) Further research is required to determine how perceptual biases

in the rst moments of contact might play themselves out over
the course of social interaction.
AUTOMATIC PROCESSING I (II)
John A. Bargh, Mark Chen, and Lara Burrows. New York University. Automaticity of
Social Behaviour: Direct e ects of trait construction and stereotype activation
on action 1996
A. Experiments 1, 2, & 3 
Prior research has shown that trait concepts and stereotypes become active
automatically in the presence of relevant behaviour or stereotyped group
features. Through the use of the same priming procedures as in previous
impression formation research, Experiment 1 showed that participants
whose concept of rudeness was primed interrupted the experimenter more
quickly and frequently than did participants primed with polite related stimuli.
In Experiment 2, participants for whom an elderly stereotype was primed
walked more slowly down the hallway when leaving the experiment than did
control participants, consistent with the content of that stereotype. In
Experiment 3, participants for whom the African American stereotype was
primed subliminally reacted with more hostility to a vexatious request of the
experimenter.  
Implications of this automatic behaviour priming e ect for self-ful lling
prophecies are discussed, as is whether social behaviour is necessarily
mediated by conscious choice processes.
B. Automaticity in Attitudes and Social Cognition 

Extent to which one’s own thought and behaviour are or are not under one’s
91
fi
fi
ff
ff
ff
fi
own intentional control is a fundamental existential question. 
Researchers in the area of attitudes and social cognition have documented
that many of the phenomena they study are unintentional or automatic in
nature (Bargh, 1994; Smith, 1994; Wegner & Bargh). 
Attitudes are discovered to become activated automatically on the mere
presence of the attitude object, without conscious intention or awareness
(i.e., preconsciously; Bargh, 1989) to then exert their in uence on thought
and behaviour (Bargh, Chaiken, Govender, & Pratto, 1992; Bargh, Chaiken,
Raymond, & Hymes, 1996; Fazio, Sanbonmatsu, Powell, & Kardes, 1986). 
Self-concept is shown to become active automatically on the presence of
self relevant stimuli to a ect self perception and emotions. 
Stereotypes can also become active automatically on the mere presence of
physical features associated with the stereotyped group and categorising
behaviour in terms of personality traits, and then making dispositional
attributions about personality have both been shown to occur automatically
to some extent.  
Although it is quite reasonable to assume that attitudes and social perceptual
processes exist in the service of guiding behaviour, this does not require the
assumption that behavioural responses always require such services.
C. Case for Automatic Social Behaviour
1. Behavioural responses can be associated with situational features 

Social behavioural responses are represented mentally just as are trait
concepts and attitudes. Thus, they should be capable of becoming
activated automatically on the mere presence of relevant features in the
environment by the same principles that produce automatic trait
categorisation and automatic attitude activation. 
Behavioural responses can be activated immediately by situational
context, and there is no theoretical or conceptual reason why the e ects
of preconscious, automatic activation should be limited to perception and
evaluation. 
Preconscious activation of mental representations develops from their
frequent and consistent activation in the presence of a given stimulus
event in the environment. (Bargh, 1989; Shi rin & Schneider, 1977).
2. Principle of Ideomotor Action 

William James believed that the mere act of thinking about a behaviour
increased the tendency to engage in that behaviour, which is known as
the principle of ideomotor action.  
Lashley (1951), believed that priming was the preparatory function of
thought. Argued for the necessity of a direct connection between thought
92
ff
ff
fl
ff
and behavioural representations. 
Automaticity of ideomotor action e ect, merely thinking about a
behaviour makes it more likely to occur, even if unintended, has been
demonstrated in a series of studies by Wegner and his colleagues. (1994).
Ironic process model contends that acts of intentional control over
thought and behaviour involve an automatic monitoring of the presence
of the unwanted state.  
Irony of this e ect is that likelihood of this occurrence under attentional
load is actually greater than if the person had not tried to stop that
response.
3. Perception Behaviour Link 

The likelihood of a behavioural response may increase from thinking
about that behaviour, regardless of the source of that thought. Cognitions
about a type of behaviour can come not only from internal sources, but
also from external sources, such as perceiving that type of behaviour
enacted by others. 
People implicitly mimic facial expressions of others, such that when this
subtle imitation is prevented, memory for the faces are impaired.  
Hypothetically, activation will spread automatically from the interpretive to
the behavioural schema, so that perceiving another person’s hostility
increases the likelihood that one will behave in a hostile manner oneself. 
Activation of the concept of hostility had the simultaneous e ect of
making the participant both more likely to perceive hostility in another
person and to behave in a hostile manner him or herself.
II. Experiment 1 : Behavioural Consequences of Trait Construct Priming
A. Method and Participants 

34 students at NYU. 
Priming manipulation took form of Scrambled Sentence Test, presented to
participants as test of language ability. For each of 30 items, participants are
to use the 5 words listed to construct a grammatically correct 4 word
sentence as quickly as possible. 5 words presented in a given test item are in
scrambled order, such as “he it hides nds instantly”. 
3 versions of scrambled sentence test were constructed, one was intended
to prime construct rude, and the other to construct polite, and the last to
prime neither trait.
B. Procedure 
Participants took part one at a time.  
Handed participant an envelope that contained one of the 3 forms of the
scrambled sentence test. Randomly determined type of test version. 
93
ff
ff
fi
ff
Participants told to complete and then nd experimenter at a di erent room
around the corner. 
Confederate posed as another participants who was apparently having
di culty understanding how to complete a task. Confederate and
experimenter would hence be talking. Participant would be waiting for
experimenter to acknowledge his or her presence and give them the next
experimental task to complete. 
Dependent measure was the amount of time the participant would wait until
interrupting the conversation between experimenter and confederate and ask
to be given the next experimental task. 
Time taken was recorded.
C. RESULT
1. Participants in the rude priming condition interrupted signi cantly

faster than did participants in the neutral priming conditions.
2. Participants whose concept of rudeness had been surreptitiously

activated in the scrambled sentence test subsequently were more
likely to interrupt the conversation between the experimenter and
confederate than were the other participants, and those who
concept of politeness had been activated were the least likely to
interrupt.
3. It is possible that rude primed participants were more likely to

perceive the experimenter as rude because he was ignoring them
and attending to the confederate, whereas polite primed participants
were more likely to perceive the experimenter as polite because he
so patiently dealt with the confederate’s questions.
4. The fact that the experimenter essentially ignored participants

(focusing his attention on the confederate and her questions) while
they waited in the hall led all of them to perceive him as equivalently
non polite (but not impolite either).
5. Strong e ects of priming manipulation, e ect on judgement measure

was nonexistent. Not the case that priming manipulation a ected
consciously made judgements about the experimenter, which then
determined behavioural responses to him. Results point to direct
e ect on behaviour that is not mediated by conscious perceptual or
judgemental processes. In fact, the present priming e ect in
94
ff
ffi
ff
fi
ff
ff
ff
fi
ff
impression formation research (cf. Bargh, et al., 1986; Srull & Wyer,
1979).
D. DISCUSSION
1. Participants primed with rudeness related stimuli in an ostensibly

unrelated rst experiment interrupted a conversation reliably faster
and as a group, more frequently than did other participants.
2. Those whose concept of politeness was surreptitiously activated

interrupted the least often.
3. Behaviour in social interaction, like social perception and evaluation,

apparently can be driven directly by environmental stimuli, pre
consciously and automatically.
4. Passive, automatic activation of a trait concept results in trait-like

behaviour by the individual. But there are 2 ways in which trait
concepts can be activated directly by the environment. 
One is a direct activation by the presence in the environment of trait-
relevant behaviour. Several lines of research show that behaviour
relevant to a trait automatically activates that trait concept (Newman
& Uleman, 1989; Srull & Wyer, 1979). 
Another way that a given trait concept can be activated
automatically is by its membership in a larger schema, such as a
stereotype.
5. Stereotypes of social groups consist in part of constellations of

interrelated trait concepts (e.g. Brewer, 1988; Devine, 1989; Fiske &
Neuberg,1990) that become active in an all or none fashion in the
presence of the features of a group member.
6. Priming or automatic activation of both single trait concepts (e.g.

Bargh & Pietromonaco, 1982; Srull & Wyer, 1979) and stereotypes
(Devine, 1989) in uences social perception without the individual
being aware of or intending this in uence.  
Experiment 1 shows that priming a single trait concept in uences
subsequent social behaviour as well. Deduction that follows naturally
from this is that priming or automatic activation of stereotypes
should make the perceiver themselves more likely to act in
accordance with the trait concepts that participate in that
stereotype.
95
fi
fl
fl
fl
III. Experiments 2a and 2b : Behavioural E ects of Activating the Elderly
Stereotype
A. Methods, Participants, and Procedure 

Method and procedure are identical to Experiment 1. 
Priming condition was changed to slowness (elderly) and age non-speci c
words. 
Experimenter recorded the time the participant took to walk down the
corridor after exiting the laboratory room. 
30 male and female NYU undergraduates. 
After participant completed priming task, the confederate of experimenter
sitting in a chair apparently waiting to talk to a professor in a nearby o ce
recorded the amount of time in seconds that the participant spent walking a
length of corridor starting from the doorway of the experimental room and
ending at a broad strip of silver carpet tape on the oor 9.75m away.
B. RESULTS
1. Participants in the elderly priming condition had a slower walking

speed compared to participants in the neutral priming condition as
predicted.
2. Analyses revealed that participants in elderly priming condition again

had a slower walking speed compared to participants in the neutral
priming condition.
3. Passively activating the elderly stereotype resulted in a slower

walking speed.
C. DISCUSSION
1. Exposing individuals to a series of words linked to a particular

stereotype in uences behaviour non-consciously.  
How the activated stereotype in uences behaviour depends on the
content of the activated stereotype itself, not the stimulus words
actually presented.
2. No allusions to time or speed in the stimulus materials, results of the

study suggest that elderly priming stimuli activated the elderly
stereotype in people’s memories, and participants subsequently
acted in ways consistent with that activated stereotype.
IV. AWARENESS CHECK STUDY
96
fl
fl
ff
fl
ffi
fi
A. Rationale 
Need to check that automatic e ects on behaviour derives from the
perceiver’s lack of awareness of the in uence of the words. Conclusions in
terms of automatic social behaviour depended on the participants not being
aware of this in uence.
B. Method and Participants 

19 male and female students at NYU. 
One at a time. Informed that purpose of study was to investigate language
pro ciency and they would complete a scrambled sentence task. 
Participants randomly administered either version of the task containing
words relevant to elderly stereotype or the neutral version containing no
stereotype relevant words. 
After completion, participants were asked to complete a version of the
contingency awareness debrie ng
C. RESULTS 
Only 1 of 19 participants showed awareness of relationship between
stimulus words and elderly stereotype. Even this participant could not
predict in what form or direction their behaviour might have been
in uenced had such an in uence occurred. 
Safe to conclude that e ect of elderly priming manipulation on walking
speed occurred nonconsciously.
V. MOOD MEDIATION ON EFFECT OF PRIMING ON WALKING SPEED STUDY
A. Rationale 
Alternative explanation for e ect of elderly stereotype related stimuli on
walking speed is that in general, words relating to elderly are more likely than
control words to induced in participants a sad mood, which might then be
the reason why they walked more slowly.
B. Method and Participants 

33 undergraduates from NYU. 
SST again, prime containing elderly stereotype relevant words and neutral
version. 
Modi ed version of A ect-Arousal Scale (Salovey & Birnbaum, 1989) was
adminstered to each participant to show how they felt at that moment.
C. RESULTS 
Main e ect of priming condition was not signi cant across the 2
dependent measures. 
Participants in the priming condition were in a more positive mood than
97
fl
fi
fi
ff
fl
ff
ff
fl
ff
fi
ff
fl
fi
control participants and were more aroused or energetic as well.  
Participants exposed to the elderly stereotype stimuli were not more
likely to be sad or less aroused compared to participants who were not
exposed to the stereotype related stimuli.
VI. Experiment 3 : Behavioural E ects of the African American Stereotype
A. Background 
Experiment 2 shows the automatic activation of elderly stereotype has direct
and nonconscious e ects on behaviour in line with the content of the
stereotype. 
Experiment 3 was intended to assess the generality of these results to an
entirely di erent stereotype, that for African Americans. 
Priming stimuli was now presented subliminally to rule out experimenter
demand or other explanations of results in terms of conscious, strategic,
processes.
B. Method, Participants and Procedure. 

Participants worked on computerised visual task that pretesting had shown
to be very boring and tedious. 
Before each trial, computer ashed a subliminal picture of a young African
American male face or a picture of a young Caucasian male face. On the
130th trial, computer alerted the participant of an ostensible data saving
failure and also informed the participant that he or she would have to do the
entire computer task again.  
Hidden video camera was placed in lab room to capture participants’ facial
reactions to computer error and news that the task would have to be redone. 
41 non African American undergraduate students from NYU. 
Participant had to do a task that made them determine whether or not a
picture contained an odd or even number of circles. Asked to make the most
accurate judgement possible. 
Experimenter left the room. 
Participant would encounter error on the 130th trial and participant would
have to wait for experimenter to come back and ddle with the computer
before saying they might have to redo the experiment again. 
Reactions were recorded.  
Subsequently, participants were asked to ll out 2 questionnaires, Racial
Ambivalence Scale (Katz & Hass, 1988), and the Modern Racism Scale
(McConahay, 1986). 
Probed for suspicion and debriefed.
C. RESULTS
98
ff
ff
fl
ff
fi
fi
1. MANOVA revealed signi cant e ect of priming condition across both
indexes of hostility such that participants primed with photographs
of African American faces behaved in a more hostile fashion
compared to participants primed with Caucasian faces.
2. No interactions between rating source and priming condition, and

codings of hostility did not signi cantly di er by source of rating.
3. Participants who were low in racist attitudes toward African

Americans were just as likely to behave in a hostile manner as
participants who were high in racist attitudes, regardless of priming
condition.
1. Activation of a trait construct or a stereotype in one context resulted

in behaviour consistent with it in a subsequent unrelated context.
Participants were not aware of the in uence or potential in uence of
the priming events on their behaviour.  
Same priming techniques that have been shown in prior research to
in uence impression formation produce similar e ects when the
dependent measure is switched to social behaviour.
2. A person has behavioural responses (e.g. assertiveness, anger,

patience) that are associatively linked to particular situations (e.g.
being made to wait by another person, losing one’s work because of
another’s mistake). This is congruent with the proposal made by
Lewin (1943), Mischel (1973), Berkowitz (1984), Higgins (1987), and
others that immediate psychological reactions to the environment
are not only cognitive or a ective in nature but also include
motivational and behavioural responses.
3. Behavioural responses become automatically linked to

representations of social situations just as previous research has
found perceptual trait constructs, stereotypes, and attitudes to
become automatically activated.
4. Identical priming manipulations result simultaneously in automatic

perceptual and behavioural e ects.
5. According to the behavioural schema model (Carver et al., 1983), the

perceptual and actional representations of the same type of
behaviour share many features in common and thus develop strong
99
fl
fi
ff
ff
ff
fi
fl
ff
ff
fl
connections. 
If one has just perceived another person acting in a generous or an
aggressive way, one’s behavioural schema for generosity or
aggression is activated and accessible, and so one is more likely to
behave that way oneself in subsequent situations for which
generosity or aggression is a relevant response.
E. UNDER WHAT CONDITIONS WILL BEHAVIOUR BE AUTOMATIC?
1. Priming stimuli directly activated trait concepts, directly in

Experiment 1 or by virtue of their inclusion in all or none stereotypic
representations, as in Experiments 2 and 3. These representations
contain the knowledge of what it means to act in the trait-like
manners, as well as the mechanisms for producing that kind of
behaviour. (Carver et al., 1983; Prinz, 1990).
2. Primed behaviour was relevant and appropriate for the experimental

situation into which we placed the participant. In accessibility logic,
the presentation of that behaviour was applicable to the situational
information (Higgins, inpress).
3. Automatic activation can occur only if the individual has that

behavioural representation available in the rst place.
4. Automatic social behaviour may occur only if the behavioural

representation that is activated is already associated with that
situation by the individual.  
In the present experiments, it is likely that all primed behaviours
were in the participants repertoire for those situations. 
Lewin’s 1943 eld theory holds that although you might be able to
a ect a person’s behaviour by making some motivations more
salient than others, you cannot give the person motivation that he or
she does not already have and make him or her do something for
which he or she has no motive base (Cartwright, 1959).
F. RELATIVE STRENGTH OF PRIMING EFFECTS ON BEHAVIOURS VERSUS

IMPRESSIONS
1. One might think that because judgements and perceptions mediate

behaviour, impression e ects should be stronger, but the
experiments show that obtaining priming di erences on behavioural
measures were easy.
100
ff
fi
ff
fi
ff
2. Some di erences that are likely to produce greater or lesser priming
e ects lie in that public behaviour measured in this experiment is
more constrained than are subjective judgements along personality
scales, there are less variability, hence e ects are easier to detect.
G. DOES SOCIAL PERCEPTION ALWAYS MEDIATE SOCIAL BEHAVIOUR?
1. The greater variability and freedom that people have in perceiving

others’ behaviours also suggest that social perception will be less of
a mediating in uence on behaviour than many might assume. 
 
Perceptions may play a role, but they are one of the several
in uences, and it is clear from the present results that behaviours
can be triggered directly without being mediated by impressions or
judgements of the person with whom one is interacting.
2. We should hold the assumption of mediation to the same standards

of proof and evidence to which we hold the hypothesis of direct,
non-mediated behaviour instead of assuming mediation by default.
H. CAN AUTOMATIC BEHAVIOUR BE CONTROLLED?
1. Control over automatic in uences require three things. 

Awareness of the in uence or at least the possibility of the in uence 
Motivation to exert the control 
Enough attentional capacity or lack of distractions at the time to
engage in the control process.
2. Awareness of the automatic e ect is necessary for the motivation to

be engaged, and for the motivation to operate to control the
automatic impulse it must be supported by su cient processing
capacity.
3. Even with the best of underlying intentions, one cannot control an

in uence if one is not aware of its operation or potential for
operating (Devine, 1989 and Experiment 2). 
Even if one is aware of the in uence it is possible to slip up if one is
not paying enough attention, as in ironic failures of control (Wegner,
1994) and action slips (Norman, 1981).
4. The individual’s lack of awareness of the source of automatic

behaviour impulses usually translates into a lack of monitoring or
attempt to control them, as well as a tendency to misattribute them
101
ff
fl
fl
ff
fl
fl
fl
fl
ff
ff
ffi
fl
to possible and justi able causes of which the individual is aware
(e.g. Nisbett & Wilson, 1977), which will also increase the likelihood
that the activation of automatic behaviour responses will nd
expression.
I. IMPLICATIONS FOR STEREOTYPE CONFIRMATION AND EMPATHIC

REACTIONS
1. There may be an automatic, non-conscious basis for self-ful lling

prophecy e ects (e.g. Snyder, Tanke, & Berscheid, 1977). If the
automatic activation of a stereotype by the physical features
(including speech accent, skin colour, gender, and age related
features) of another person causes the perceiver to behave in line
with the stereotype rst, the perceiver’s initial behaviour to the target
could well produce similar behaviour in the stereotyped individual. 
(E.g. the non-conscious in uence of the African American stereotype
could cause the perceiver to behave in a hostile manner to African
Americans they encounter and produce behavioural con rmation of
the stereotype, such as a hostile response in reaction.).
2. The perceiver’s subjective, phenomenal experience and hence

memory of the event, would be of the stereotyped group member’s
unprovoked initial hostility.
3. The perception behaviour link may be an important ingredient in the

“glue” that binds 2 (or more) interaction partners, keeps them on the
same wavelength, and helps to bring each partner a sense of
validation by others of their experience.
4. Automatic e ects of perception on behaviour is not necessarily

entirely bene cial for social interaction. 
Automatic e ect of priming on behaviour that has been documented
is further reason to believe that aggression in the mass media does
produce aggressive tendencies in the viewer (Berkowitz, 1984)
perhaps even more insidiously and pervasively than previously
believed.
J. CONCLUSIONS
1. Social behaviour can be triggered automatically by features of the

environment. 
Experiment 1 showed same trait priming manipulations that have
102
ff
ff
ff
fi
fi
fi
fl
fi
fi
fi
exerted a non-conscious in uence over social perceptual processes
in previous research were shown to produce trait-like behaviour as
well. 
Experiments 2 and 3 showed that trait-like behaviour is also
produced via automatic stereotype activation if that trait participates
in the stereotype.
2. The experiment reveals the apparent degree to which social

behaviour occurs unintentionally and without conscious involvement
in the production of that behaviour.
3. Possibility that the automatic activation of one’s stereotypes of

social groups, by the mere presence of group features, can cause
one to behave in line with that stereotype without realising it.
4. After eliciting that type of behaviour in response, because one is not

aware of one’s own role in provoking it, one may attribute it to the
stereotyped group member and hence the group.
5. Social behaviour is like any other psychological reaction to a social

situation, capable of occurring in the absence of any conscious
involvement or intervention. 
AUTOMATIC PROCESSING I (III)
Joseph Cesario, E. Tory Higgins, Jason E. Plaks, Columbia University. Automatic

Social Behaviour as Motivated Preparation to Interact. 2006.
A. Authors propose that automatic social behaviour may result from perceivers
preparing to interact with primed social group members. Authors suggest
stored knowledge about a social group (e.g. stereotype) can provide useful
information for interacting e ectively with group members, and automatic
social behaviour that arises from the activation of a social category can be
the result of perceivers preparing to interact with a member of the target
category, that is, when a social group category is activated, perceivers
engage in a motivated preparation to interact with a group member. This
preparation to interact, in turn results in systematic and measurable
automatic behaviour.
103
ff
fl
B. In Study 1, participants primed with a disliked outgroup (gay men) showed
evidence of interaction preparation (aggression) rather than direct stereotypic
trait expression (passivity).
C. In Study 2, participants with implicit positive attitudes toward the elderly

walked more slowly after elderly priming, but participants with negative
attitudes walked more quickly, results consistent with a preparatory account;
the reverse was found priming “youth”.
D. Study 3 demonstrated that the accessibility of a primed category follows a

pattern more consistent with that of goal related constructs including post
goal ful lment inhibition than that of semantically primed constructs.
Implications for the function of stored knowledge are discussed.
II. STUDY 1. Testing Divergent Predictions of our Motivated Preparation Account

and an Account Based only on the Direct Expression of Activated Traits.
A. Method, Participants, and Design 

71 participants from Columbia University in exchange for 5 dollars. Gender
had no main e ect or interactions for any dependent variable.
B. Procedure 
Participants completed experiment individually in a small room with a table
and two soundproof booths. Male confederate was ostensibly lling out
questionnaires at the table, with his back to the door when participants
entered. They were told they would complete 2 unrelated studies that had
been combined to save time and money, the rst investigating quick visual
judgements and the second testing “several short, new questionnaires”. 
Participants were seated in the rst soundproof booth. The position of the
confederate allowed him visual and auditory access to the participant with a
slight tilt of his head.  
First study, the priming task, was then begun. Participants were told they
would be asked to judge, as quickly and accurately as possible, whether an
odd or even number of circles appeared on the screen at any given time.
When the experimenter began the computer task, it randomly assigned the
participant to the no-prime control, straight prime, or gay prime condition.
Both the experimenter and the confederate were blind to condition. 
Program designed to present subliminal prime pictures while participants
judged whether an odd or even number of coloured circles were presented
on the screen. A prime picture (or a blank screen in control condition) was
presented for 11ms. Prime consisted of a black and white head and
shoulders photograph of a white adult man with either the word GAY or
STRAIGHT in black 48 point font written beneath his chin. Picture was
immediately followed by 2 pattern masks for 21 ms each, rst contained
104
fi
ff
fi
fi
fi
fi
black diagonal crosshatches, with the letters XVXVXV in the same position as
the prime word, and second contained six circles of di erent size and
position on a gray background.  
Finally, target picture was presented for 2000, 2500, or 3000 ms, followed by
a 1500 ms intertrial interval. Target was 1 of 14 randomly selected pictures
with blue and purple circles on a gray background, Participants pressed the
1 key for odd responses and 2 key for even responses. 
After 130 trials of this tedious task, computer supposedly crashed. Asked to
get the experimenter. Experimenter would ask the participant to start over.
Confederate and experimenter would note down the participants reaction,
then afterwards Experimenter would tell the participant that the data has
been recovered. 
Second Study, participants lled out following measures in order, 2 self
regulation measures related to the current study, Herek’s (1988) Attitudes
toward lesbians and gay men scale ATLG, (Free response measure eliciting
the stereotype content of gay men), and a funnelled debrie ng form.
C. Dependent Measures
1. Hostility Rating 
Experimenter and Confederate both rated participants’ reactions on an
11 point scale anchored at 0 = not hostile at all and 10 being extremely
hostile. Average ratings for each participant served as the primary
dependent variable.
2. ATGL 
20 item scale designed to assess explicit attitudes toward both lesbians
(ATL Subscale) and gay men (ATG subscale). Because primes involved
gay and straight men, only the ATG scale was relevant. Participants rated
on 9 point scales from 1 = strongly disagree to 9 = strongly agree, their
agreement with a variety of questions concerning gay men (“If a man has
homosexual feelings, he should do everything he can to overcome
them”).
3. Stereotype Content Assessment Measure. 

Modelled after Devine’s (1989) free response task, to assess participants’
knowledge of the cultural stereotype of gay men and to ensure that no
knowledge of the cultural stereotype of gay men and to ensure that no
participant believed gay men to be hostile in any way. Participants were
instructed to list the typical characteristics of 3 social groups: African
American men, Elderly women, and Gay men.
4. Demographics 
One item in the demographics form asked participants to indicate their
105
fi
ff
fi
sexual orientation by choosing either “heterosexual”, “homosexual”,
“bisexual”, or “I prefer not to answer”.
D. Results
1. Analysis 1: Hostility Ratings 

Near-signi cant e ect of prime condition, planned contrasts
revealed that the control prime and straight prime conditions did not
di er from each other.  
Participants in the gay prime condition however, had signi cantly
higher hostility ratings than both the control prime and the straight
prime conditions. 
Inclusion of the no prime control condition allows for the conclusion
that this latter di erence is due to an increased hostility in the gay
prime condition and not to a decreased hostility in the straight prime
condition. 
Participants responded with more hostility following priming of a gay
man than of a straight man.
2. Analysis 2: Explicit Attitude Measure 

No relation between explicit measure of attitudes toward gay men
and hostility ratings.
3. Analysis 3: Stereotype Content 

Important to show that no participant believed the stereotype of gay
men to include hostility or related traits. If anything, stereotype of
gay men is e eminate and weak, which is the exact opposite of
hostile. 
No participant listed any hostility related work as part of gay male
stereotype, and a full 34 participants 68% listed e eminate, weak,
feminine, or some form of these words in their free response
descriptions.  
Replicates previous research demonstrating gay men to be
perceived as relatively passive or e eminate and rule out a trait
expression explanation of these results. 
Gay men are not viewed as hostile or aggressive but rather, its
opposite.
E. DISCUSSION (STUDY 1)
1. Provided evidence that priming social categories can lead to

behaviours opposite of those traits associated with the stereotype.
106
ff
fi
ff
ff
ff
ff
ff
fi
Automatic behaviour may result in part from preparation to interact
with a member of a primed social group, rather than only from direct
expression of activated stereotypic traits.
2. If automatic behaviours were due solely to activating stereotypic

traits and the ideo-motor (ideomotor) process, then participants
primed with gay men should, if anything, have behaved with less
hostility than those primed with straight men. In contrast, preparing
to interact with a negative outgroup member would produce greater
rejection related hostility for the gay male prime, as found in Study 1.
3. Limitation included that no implicit measure of attitudes toward gay

men was collected. Although we found the predicted null e ect of
explicit attitudes on automatic behaviour, it would be useful to show
that an implicit measure does predict behaviour.  
Study 2 addressed this issue by collecting implicit as well as explicit
attitude measures toward target social group.
III. STUDY 2. ROLE OF ATTITUDES AND IMPLICIT MEASURE OF ATTITUDES
A. Background (2 Major Purposes)
1. First purpose was to test the role of attitudes when measured implicitly as
a motivational underpinning in automatic behaviours, as predicted by
their impact on preparing action toward a social category member.
2. Second purpose was to test hypothesis that implicit measure of attitudes

will relate to automatic behaviours even though an explicit measure does
not. Thus, 2 explicit measures of attitudes toward the elderly were
assessed at Time 1, allowing for comparison of explicit and implicit
attitude measures in their ability to predict automatic behaviours.
B. Method, Participants, and Design 

80 Columbia University students. No participant from Study 1 was included
in this study.  
Primary dependent variable was time it took participants to walk out of the
experiment room after priming task.  
Independent variables were prime condition (elderly, youth, no prime control),
4 orthogonal implicit attitude scales (positivity toward elderly, negativity
toward elderly, positivity toward youth, negativity toward youth), and 2
explicit attitude toward elderly scales (Semantic Di erential Scale; Osgood,
Suci, and Tanenbaum, 1957; and Kogan Attitude Toward Old People Scale;
Kogan, 1961).
107
ff
ff
C. PROCEDURE 
To conceal true nature of study, informed participants that they would
participate in 2 studies investigating “perceptual and motor activity, how they
relate, and how they change over time”. 
First session involved in following order, 3 self-regulation measures unrelated
to the current study, the implicit attitude measures (sequential priming task),
an approach/avoidance measure unrelated to the current study; and 2
explicit attitude measures (Semantic Di erential Scale and Kogan Attitude
Toward Old People Scale).  
Primary purpose of rst session was to assess implicit and explicit attitudes
toward elderly and youth and to separate these measures from the primary
dependent variable, so as to ensure no contamination of responses in the
priming task of Session 2. 
After completing these measures, participants signed up for second session
some time later. Length of this delay ensured that any potential in uence of
completing the attitude measures would decay by the second session. 
Session 2 took place in a long rectangular computer lab.  
A confederate with a concealed stopwatch sat at a large table in the centre
of the room, ostensibly working on an assignment. When participants arrived
(individually) at this session the experimenter met them at the door so as not
to walk with the participant. 
Confederate began the stopwatch as participants passed a piece of oor
tape (1.22m) from the door, and stopped it when participants reached a
piece of oor tape (1.22m) from the computer. This provided a prepriming
measure of the time it took the participant to walk across the set distance,
which was approximately 7.92m. 
Experimenter walked to the computer and told the participant that the
session would consist of a single 5 minute task during which he or she would
be asked to estimate the number of Xs present on the screen at any given
time. 
Judgement task allowed for subliminal presentation of elderly picture primes
or youth picture primes (or no primes in the control condition). 
Following the X-judgement task, participant were paid and dismissed; they
were told they would be sent a debrie ng form by email, so as to avoid
participants asking about the purpose of the study and delaying their exit. As
experimenter stayed at computer preparing it for the next participant, the
confederate recorded the participant’s walking time across the same 2
pieces of oor tape as he or she exited. After participant passed the second
piece of tape, the experimenter called out to him/her that he “forgot one last
thing”, then approached the participant and questioned about any possible
awareness of the prime during the judgement task, in a funnelled manner.
108
fl
fl
fi
fi
ff
fl
fl
D. MATERIALS
1. Implicit Attitude Measure 

Modelled after Fazio’s sequential priming task 1995, toward the “elderly”
and “youth”. Participants judged the valence of target adjectives
preceded by social category label primes. Relatively faster response
latencies in evaluating positive target adjectives when preceded by the
category prime is indicative of more positive attitudes toward that group,
and relatively faster response latencies in evaluating negative targets
when preceded by the category prime is indicative of more negative
attitudes toward that group. 
Check document for measure design*.
2. Semantic Di erence Scale (Explicit Attitude Measure 1). 

Rate the extent to which one of two opposing adjectives described the
social group elderly. 5 pairs of adjectives were given; bad and good,
pleasant and unpleasant, cold and warm, like and dislike, dirty and clean.
Each adjective was opposite end of 7 point scale, with midpoint
designated as neither adjective describing the social group better.
Responses were coded such that higher scores equaled more positive
responses.
3. Kogan Attitude Toward Old People Scale (Explicit Attitude Measure 2). 
Measures tapping attitudes toward elderly people and is conceptually
related to other explicit measures used in automaticity experiments.
Higher scores indicate more positive evaluations of elderly people.
4. X-judgement priming task. 

Expose participants to pictures of either elderly men (elderly prime
condition) or teenage boys (youth prime condition) or to no pictures (no
prime control condition) prior to measuring their exit walking time.
Participants had to estimate how many Xs were present on the screen at
any given time, with subliminal exposure to the primes preceding
presentation on each trial.
E. Dependent Measure
1. Time it took participants in the second session to walk out of the

experiment room following exposure to either elderly youth, or no picture
control primes.
F. RESULTS
1. Analysis 1: Replication of Previous Automatic Social Behaviour

E ects.
109
ff
ff
a) Participants in the elderly prime condition had the slowest mean
exit time, and participants in the youth prime condition had the
fastest mean exit time, with no prime control participants in
between.
b) Participants primed with elderly pictures took more time walking

to the exit than those primed with youth pictures. Mean exit time
for the no prime control condition was between the elderly and
youth prime conditions and was not signi cantly di erent from
either.
2. Analysis 2: Prediction of Elderly Prime Exit Walking Time by Implicit

Attitude Measure.
a) As positivity toward “elderly” increased (faster response

latencies), exit walking time signi cantly increased. (Walked
slower).
b) As negativity toward “elderly” increased (faster response

latencies), exit walking time signi cantly decreased. (Walked
faster).
c) Exit walking time was not determined simply by stereotype

activation associated with category priming but was more
completely explained by the strategic behaviours that would best
ful l attitudes toward the group members, walking slowly when
one likes the elderly and walking more quickly when one dislikes
the elderly.
3. Analysis 3: Prediction of Youth Prime Exit Walking Time by Implicit

Attitude Measure.
a) As positivity toward youth increased, exit time signi cantly

decreased. (Walked faster)
b) As negativity toward youth increased, exit time signi cantly

increased. (Walked slower)
c) Provide additional convergent evidence for the motivated

preparation hypothesis that attitudes toward a social category
inform the most e ective means of interacting with members of
that category.
110
fi
ff
fi
fi
fi
ff
fi
fi
d) In addition, the nding for participants with negative attitudes
toward youth in the youth prime condition rules out a possible
alternative explanation for the nding described above for
participants with negative attitudes toward the elderly in the
elderly prime condition. Elderly prime participants who dislike the
elderly may have walked more quickly simply because negative
attitudes induce greater arousal or vigour? If so, the youth prime
participants who dislike youth should have walked more quickly,
instead they walked more slowly.
4. Analysis 4: Full Regression model Predicting Exit Walking Time.
a) *Analysing present data alternative method: Multiple regression

on exit walking time with the full model implied by our study. One
could examine the ability of each implicit measure to predict exit
walking time uniquely for the relevant prime condition, while
simultaneously demonstrating that each implicit measure does
not predict exit time for the non relevant prime condition. If
implicit attitudes toward the elderly are in uencing walking
speeds because they imply e ective interaction behaviours
toward the elderly, then they should have no bearing on walking
speed after priming of youth. The same is true for youth attitude
measure.
b) 3 Priming Conditions, comparing elderly and control conditions

(PRIME 1), comparing elderly and youth conditions (PRIME 2), and
nally the relevant interactions among the variables were created,
Prime 1 x Elderly Attitude, Prime 2 x Elderly Attitude, Prime 1 x
Youth Attitude, and Prime 2 x Youth Attitude.
c) Prime 2 x Elderly Attitude Interaction were signi cant.
d) Those with positive attitudes walked more slowly following

priming of “elderly”, but those with negative attitudes walked
more quickly following priming. In elderly prime condition, the
implicit youth attitude measure did not predict exit time.
e) Youth prime condition, as attitudes became more positive, exit

time decreased. Participants with positive attitudes walked more
quickly following youth priming, but participants with negative
111
fi
fi
ff
fi
fl
fi
attitudes walked more slowly. In addition, elderly attitude
measure did not predict exist time in this condition.
f) In control condition, elderly attitude measure signi cantly

predicted walking speed, and the youth attitude measure, though
nonsigni cant, related to walking speed in the direction opposite
of the elderly measure.
5. Analysis 5: Prediction of Exit Walking Time by Explicit Attitude

Measures.
a) No relation between attitudes and exit walking time following

elderly primes.
6. Analysis 6: Comparison of Implicit Attitude Measures and Explicit

Attitude Measures.
a) E ects of implicit attitudes on exit walking time are independent

of participants’ general walking speed.
b) Participants had inherent, cognitive beliefs about the speeds of

old and young people. (Old means slow, young means fast).
c) Rules out explanation that participants who had positive or

negative implicit attitudes toward the elderly somehow di ered in
stereotype content.
G. DISCUSSION
1. Implicit attitude measures signi cantly predicted automatic

behaviour following priming of elderly and youth social categories in
a way consistent with our motivated preparation account.
2. Demonstrated the expected di erences in predictive validity for

implicit versus explicit attitude measures. (Elderly people walk
slowly, and youth walk quickly).
IV. STUDY 3: PROVIDING DIRECT EVIDENCE OF PRIMING THAT INDUCES

MOTIVATED BEHAVIOUR
A. Hypothesis
1. Predicts that priming “elderly” activates a goal to interact with the elderly.
Substitutable task of writing about interacting with an elderly man
represents ful lment of that goal, which should lead to inhibition or
reduction of the category accessibility. Therefore in this combination of
112
ff
fi
fi
ff
fi
fi
ff
conditions (prime/elderly writing task) the accessibility of “elderly” should
be reduced as a result of post-ful lment inhibition.  
Direct expression models however, should predict that this combination
of conditions would result in the highest level of accessibility of “elderly”.
B. METHODS, PARTICIPANTS AND DESIGN
1. 75 participants from Columbia University
C. PROCEDURES
1. Participants completed study individually in soundproof booths. Told that

the study concerned cognitive and perceptual performance and that they
would perform 2 tasks that measured this, with a ller task in between in
order to separate their performance on each task. Used this cover story
to avoid suspicion. 
Participants were then randomly assigned by the computer to either the
elderly prime condition or no prime control condition. For both groups, a
modi ed version of the X-judgement task from Study 2 was used, with
presentation of either pictures of elderly men or a blank space before
each trial. Following this task, participants were randomly assigned to
one of 2 ller task conditions. In the substitutable task condition,
participants wrote about what it would be like to interact with an elderly
man. In the irrelevant delay task, participants wrote about their daily
morning routine. After this, all participants completed a lexical decision
task designed to assess accessibility of elderly related constructs.
D. MATERIALS
1. X-judgement priming task. 

For participants in elderly prime condition, 4 pictures of elderly men were
subliminally presented at the beginning of each trial. For participants in
the control prime condition, a blank screen was presented instead.
2. Substitutable Interaction Elderly Writing Task 

Filler task to allow for activated goal of interacting with the elderly to be
completed through a substitutable task.  
Participants were asked to pretend you are interacting with an elderly
man, thinking about the sorts of things you would say if you were going
to talk with him, how you would act, the physical aspects of interacting
with him, etc. We are interested about anything that comes to mind when
you think about interacting with an elderly man.
3. Irrelevant Morning Routine Writing Task. 

Test general level of decay over time of elderly related words after priming
113
fi
fi
fi
fi
the elderly category (in the prime condition), as well as standard level of
accessibility of elderly related words when there was no previous priming
of the elderly construct (in the no prime condition). 
Asked to describe their morning routine in as much detail as possible.
E. DEPENDENT MEASURE 
Lexical decision task measured accessibility of category elderly by having
participants make word/nonword judgements for words related to that
category.
F. RESULTS
1. Primary Analysis
a) Di erence between to no prime/elderly writing condition was

signi cant. Decreased accessibility of elderly related words
following both elderly priming and elderly writing, as compared
with the no prime/elderly writing condition, can only be explained
if the priming task activated a goal, which was then ful lled by the
substitutable task.
b) Decrease in accessibility in the elderly prime/elderly writing

condition as compared with the elderly prime/irrelevant task
condition, though this di erence did not reach signi cance.
G. DISCUSSION
1. Accessibility of the primed social category (elderly) generally

followed a course that is characteristic of motivated processing.
After social category activation and subsequent (symbolic)
satisfaction of the goal, accessibility decreased, rather than
increased (compared with participants for whom the social category
was not activated).
2. A purely non-motivated account would suggest that the accessibility

of the primed social category should have been greatest following
priming and writing about the social category because such a
combination would have resulted in the most activation.
3. Priming a social category does activate interaction goals.
H. GENERAL DISCUSSION AND CONCLUSIONS.
114
ff
fi
ff
fi
fi
1. Study 1 used a social group that would elicit opposite behavioural
responses from a motivated preparation account versus a strict
direct expression account and found support for the former.
2. Study 2 demonstrates that for 2 di erent social groups: opposite

behavioural e ects for implicit positive attitudes versus implicit
negative attitudes toward each group, and results consistent with a
preparatory account.
3. Study 3 provided a more direct evidence of motivated processing in

that the accessibility of a primed category followed a time course
consistent with that of goal related constructs. When an interaction
with the primed social category took place through a substitutable
task, the accessibility of the primed category decreased.
I. FUNCTIONALITY OF STORED KNOWLEDGE: IS THINKING ONLY FOR

DOING?
1. Believe that thinking is not just for doing anything. Rather, thinking is
for doing that accomplishes goals. Because we are social beings,
our goals are often interpersonal in nature. Thus the processes
whereby stored knowledge is translated into behaviour should take a
social perspective, a purely cognitive approach may be insu cient. It
is di cult for perception and cognition to be of much assistance to
behaviour if they are not context sensitive and not in the service of
motivation.
2. Knowing that the elderly walk slowly is functional precisely because

it can be combined with other non-descriptive knowledge (one’s
motivations toward elderly; the current interaction situation) to
produce appropriate behaviour. Social interaction is not just doing; it
is doing for the current goal or end state present as a part of the
interaction or for the context in which the interaction takes place
(Plaks & Higgins, 2000).
3. If the social foundation of stored knowledge is taken seriously, it is

di cult to conceive of social behaviour as unfolding directly from
perceptual input.
4. The presence of a social group itself is enough to be considered an

interaction situation and su cient to activate relevant goals.
115
ffi
ffi
ff
ffi
ff
ffi
J. IMPLICATIONS FOR MIMICRY
1. Automatic behaviour phenomena may contain a motivated

component. The same may be true of behaviour mimicry “The
Chameleon E ect”.
2. Connection between perception and behaviour need not be

considered strictly a consequence of our perceptual and cognitive
apparatus but could be understood as part of a motivated drive to
experience a shared reality. Same applies to mimicry of gestures and
postures which occur in the service of increasing rapport. (Hardin &
Higgins, 1996), (Chartrand & Bargh, 1999; Lakin & Chartrand, 2003).
3. This research suggests a motivated alternative for these ndings.

Proposal is that when participants in mimicry experiments are placed
in an interaction situation, relevant interaction goals become
activated and the perceiver prepares to complete these goals in an
e ective manner. In mimicry experiments, when the interaction
target is similar to oneself, whether a stranger or not, an e ective
interaction is likely de ned as having a liation with the target of
having a smooth, cooperative interaction.
4. Because mimicry is one way to achieve success at a liation and to

ensure a positive interaction, behavioural mimicry will ensue.
5. What if the likability of the interaction partner were manipulated? Our

prediction was that people would evince less mimicry when paired
with a disliked partner, a hypothesis that we are currently testing.
6. Motivated processes may play an important, complementary role in

what is undoubtedly a complex, multifaceted phenomenon.
116
ff
ff
fi
ffi
ffi
fi
ff
Week 10 Readings (AUTOMATIC

PROCESSING II)
AUTOMATIC PROCESSING II
Galen V. Bodenhausen, Michigan State University Stereotypes as Judgemental

Heuristics: Evidence of Circadian Variations in Discrimination. 1990.
I. ABSTRACT
A. Question of when people rely on stereotypic preconceptions in judging

others was investigated in 2 studies.
B. As a person’s motivation or ability to process information systematically is

diminished, the person may rely to an increasing extent on stereotypes,
when available, as a way of simplifying the task of generating a response. It
was hypothesised that circadian variations in arousal levels would be related
to social perceivers’ propensity to stereotype others by virtue of their e ects
on motivation and processing capacity.
C. Subjects exhibited stereotypic biases in their judgements to a much greater

extent when the judgements were rendered at a non optimal time of the day
(i.e. in the morning for night people and in the evening for morning people.
D. Study 1 found this pattern in probability judgements concerning personal

characteristics. 
Study 2 found this pattern in perceptions of guilt in allegations of student
misbehaviour.
II. BACKGROUND INFORMATION
A. Bodenhausen and Wyer (1985) proposed that stereotypes can be viewed as

judgemental heuristics that are sometimes used to simplify the cognitive
tasks confronted by the social perceiver. Whether we respond to others
based on general beliefs about their group a thoughtful analysis of each
person’s unique attributes depends to a large extent on whether stereotypic
beliefs provide an easy, seemingly relevant basis for responding and whether
a quick or less e ortful response is necessary or desirable. (Chaiken,
Liberman, & Eagly, 1989).
B. Stereotypic responses may pre-dominate unless the social perceiver has

su cient momentary ability and motivation to engage in e ortful, systematic
thought. Motivational factors include personal involvement (Erber & Fiske,
1984), and incentives for accuracy (Neuberg & Fiske, 1987).
117
ffi
ff
ff
ff
C. Factors that limit ability to process information systematically include
distraction (Petty, Wells, & Brock, 1976), information overload (Rothbart,
Fulero, Jensen, Howard, & Birrell, 1978), and task complexity (Bodenhausen
& Lichenstein, 1987).
III. STUDY 1
A. BACKGROUND
1. Examined the impact of common social stereotypes on performance in a

probability estimation task. Materials used by Tversky and Kahneman
(1983) to demonstrate the conjunction fallacy in probability judgement
were adopted. Conjunction fallacy refers to erroneous belief that the joint
probability of two events is greater than the probability of either of the
constituent events separately. Theoretically linked to representativeness
heuristic (Kahneman and Tversky, 1972).
2. Study 1 tested students on probability estimation task at either 9AM or

8PM. Expected that morning people would be more prone to rely on
simple stereotypes and commit conjunction fallacy in the evening,
whereas evening people would be more likely to commit conjunction
fallacy in the morning.
B. Method
1. Subjects and Design 

59 undergraduate psychology students were recruited. Received course
credit.  
Randomly assigned to report to the testing location at either 9AM or 8PM
in groups of approximately 15.  
Subjects completed the MEQ and were categorised as either morning or
evening types.  
2 x 2 between subjects factorial design.
2. Materials and Procedure 

Subjects given a booklet in which they were asked to read a description
of 1 or 2 people, Bill or Linda, and make judgements about the person’s
characteristics. Descriptions were taken verbatim from Tversky and
Kahneman (1983). 
Bill was described in terms representative of common stereotypes about
accounts. 
Linda was described as possessing traits stereotypically ascribed to
feminists.  
After reading description, subjects were asked to choose which of the 2
statements about the target were more likely to be true. 
118
One of these statements consisted of conjunction of a representative
stereotypic label and an unrepresentative characteristics (e.g. Bill is an
accountant who plays jazz for a hobby) while the other was the
unrepresentative characteristic in isolation. Choice of conjunction
constitutes the conjunction fallacy.
C. RESULTS AND DISCUSSION
1. Conjunction fallacy occurred with high probability under all of the

experimental conditions.
2. 94% of subjects with morning personalities committed the

conjunction fallacy during an evening experimental session, while
only 71% did so in the morning.
3. 92% of subjects with evening personalities were more likely to

commit the conjunction fallacy during the morning experimental
session than during the evening (70%).
4. Results support the idea that people process information in a more

heuristic fashion during times of the day at which they are not at
their peak level of circadian arousal.
5. Subjects relied more on representative stereotypes about feminists

and accountants (rather than applying a straightforward rule of logic)
when there was a mismatch between their acrophase and the time of
testing.
IV. STUDY 2
A. BACKGROUND
1. Subjects asked to consider cases of alleged misbehaviour by college

students and to determine the probability of the accused students’ guilt.
Sometimes students were identi ed as members of particular social
groups, and they were accused of committing o ences that were
consistent with stereotypes of these groups. In other conditions, cases
involved students who had been accused of involvement in the exact
same o ences but who had not been identi ed as members of a
stereotyped group. Subjects were expected to rely more on guilt implying
stereotypic beliefs when asked to make guilt judgements at non-optimal
times of day.
B. METHOD
119
ff
fi
fi
ff
1. Subjects and Design 
189 undergraduate students recruited from psychology classes.  
Similar 2 x 2 between subjects factorial design.

Study was presented as investigation of legal socialisation in the college
environment. Subjects told they would read about the alleged misconduct
of other college students and would be asked to provide feedback about
these cases. 
Three di erent cases were constructed. 
One was about cheating in exam, one was about physical violence, and
the other was about sale of drugs.  
Each of these o ences matched to a stereotype group that had been
shown via pretesting in the same subject population to be associated,
however unfairly, with these o ences. 
Speci cally, athletes were more likely to be cheaters, Hispanics more
likely to be physically aggressive, and African American as more likely to
sell drugs. 
Small set of evidence was provided about each case in prose summary.
Identical for stereotypic and non-stereotypic o enders. 
Stereotype activation was accomplished in 2 di erent ways. For cheating
case, a sentence identifying the defendant as a well known athletic star
on campus was either included or was not. In other 2 cases, stereotypes
were activated by manipulating the defendant’s name. In the assault
case, the student defendant was named either Robert Garner or Roberto
Garcia. In the drug dealing case, student defendant was either Mark
Washburn or Marcus Washington.
1. Guilt was seen as most likely in the cheating case, moderately likely
in the assault case, and least likely in the drug dealing case.
2. Morning types perceived stereotyped targets to be more likely to be

guilty in the afternoon and evening than the morning.
3. Evening types perceived stereotyped targets to be signi cantly

greater in the morning than in the afternoon or evening.
4. Perceptions of non-stereotyped defendants were not a ected by

time of testing.
5. Stereotypes are more likely to be relied on in judgemental tasks that

occur at non-optimal times of the day.
120
fi
ff
ff
ff
ff
ff
ff
fi
1. Stereotypes function as judgemental heuristics and as such, are

likely to be more in uential under circumstances in which people are
less motivated or less able to engage in more systematic and careful
judgement strategies (Bodenhausen & Lichenstein, 1987; Chaiken et al.,
1989).
2. Morning people, who reach their functional peak early in the day,
were more likely to fall back on stereotypic responses in the
afternoon and evening, while evening people, who reach their
functional peak later in the day, showed greater tendency toward
stereotypic responses in the morning.
3. Regular variations in arousal levels may play a role in determining

the types of information processing strategies that are adopted by
social perceivers.
4. Highlights potential value that may accrue from considering the role
of time and temporal cycles in studies of social cognition. 
McGrath and Kelly (1986) argued that time has been a neglected
variable in social psychological research. 
Broadbent and Jones (1989) also resolved some con icting results in
research on basic attentional processes by examining the impact of
time of day on task performance.
5. Consideration of role of circadian variations on arousal levels on

social judgement is particularly interesting because it represents a
con ux of biological, cognitive, and social processes. May provide
one small step toward an integrated theoretical account of human
thought and action that exploits important developments in several
sub-disciplines.
AUTOMATIC PROCESSING II
Daniel T. Gilbert, Douglas S. Krull, and Brett W. Pelham. University of Texas at

Austin. Of Thoughts Unspoken: Social Inference and the Self-Regulation of
Behaviour. 1988.
I. ABSTRACT
121
fl
fl
fl
A. Question of how do attempts to regulate one’s own behaviour a ect the
inferences one draws about others? 
Authors suggest that perceivers draw dispositional inferences about targets
(characterisation) and then adjust those inferences with information about
the constraints on the targets’ behaviours (correction). 
Because correction is more e ortful than characterisation, perceivers who
devote cognitive resources to the regulation of their own behaviour should
be able to characterise targets but unable to correct those characterisations.
B. Experiment 1: Unregulated subjects incidentally ignored an irrelevant

stimulus while they observed a target’s behaviour, whereas self-regulated
subjects purposefully ignored the same irrelevant stimulus.
C. Experiment 2: Unregulated subjects expressed their sincere a ection toward

a target, whereas self-regulated subjects expressed false a ection.
D. In both experiments, self regulated subjects were less likely than unregulated
subjects to correct their characterisations of the target. The results suggest
that social interaction which generally requires the self regulation of ongoing
behaviour may profoundly a ect the way in which active perceivers process
information about others.
E. Social inference is not a process, but rather a consortium of conceptually

distinct processes (Gilbert, Pelham, & Krull, 1988; Quattrone, 1982; Trope,
1986). 
First, perceivers categorise or identify the actions of others (“Roger is
criticising Henry”), next they characterise or draw dispositional inferences
about the actor (“Roger is a critical person”), and nally they use information
about the situational constraints that may have conditioned the actor’s
performance to correct or adjust those characterisations (“But Roger has
been under a lot of stress lately, so perhaps he isn’t really such a critical
person after all”). 
These 3 sequential operations may di er in amount of cognitive resources
they require. Both categorisation and characterisation seem to be and
require little conscious attention (Gilbert & Krull, 1988; Johnson, Jemmott, &
Pettigrew, 1984; Kassin & Baron, 1985; Lowe & Kassin, 1980; McArthur &
Baron, 1983; Newtson, 1980; Winter & Uleman, 1984; Winter, Uleman, &
Cunni , 1985; cf. Bassili & Smith, 1986, and Higgins & Bargh, 1987). 
Correction on the other hand, seems to be an e ortful, deliberate form of
reasoning that requires a signi cant expenditure of cognitive resources.
F. Self Regulation of overt behaviour is like the rehearsal of word strings, a

resource consuming enterprise. In general, self-regulation requires that a
person envision the potential consequences of di erent courses of action
122
ff
ff
ff
fi
ff
ff
ff
fi
ff
ff
ff
and then inhibit those actions whose imagined e ects are undesirable. Even
if the person decides to perform no action at all, (e.g. when one refrains from
correcting a spouse’s inaccurate anecdote in the presence of guests) both
the decision and the subsequent inhibition may require considerable
cognitive work.
G. Experiment 1 sought to test the hypothesis that self regulating perceivers

may be able to categorise the behaviour they see and characterise the target
in terms of that behaviour, but they may be unable to use situational
constraint information to correct their characterisations.
II. EXPERIMENT 1
A. Method
1. Overview 
Subjects watched 7 silent clips from videotape of an interview, in 5 of 7
clips, the interviewee (target) appeared quite depressed and unhappy.
Half the subjects learned that in all seven clips the target had been
answering happiness inducing questions. Half of the subjects in each of
these conditions were required to perform a self regulation task (i.e. to
avoid looking at words that appeared at the bottom of the video screen)
while viewing the tape, and the remaining subjects were not.  
After viewing the tape, subjects rated the target’s dispositional sadness,
attempted to recall the questions that the target had answered, and
attempted to recall and recognise the words that had appeared at the
bottom of the screen.
2. Subjects 
50 female students at University of Texas
3. Instructions 
Subjects greeted by male experimenter who gave them a brief oral
introduction to the experiment, provided them with complete written
instructions, and then escorted each subject to a separate cubicle
(equipped with video monitor and video camera) where the subject
remained for the duration of the experiment. 
The written instructions explained the experiment as nding out the role
of non verbal behaviour in the interviewing process.  
Subjects would watch 7 short clips from a videotape of an interview that
had ostensibly taken place earlier in the year.
4. Situational Constraint Information 

Subjects were told that to protect privacy of target, the videotape would
be shown without any sound. However, subjects were also told that they
123
ff
fi
would be able to tell which of the 7 questions the target was answering in
any given clip because the question that the target was answering would
appear on the screen for 10s prior to the clip. 
Half subjects randomly assigned to the sad questions condition. 
Remaining subjects were assigned to the happy questions condition.
5. Parafoveal Optiscope 
Recorded eye movements as you watch the lm.
6. To Be Ignored Words 
Other subjects were told to learn a list of words while they watch the lm
that you are about to see.
7. Self Regulation Manipulation 

All subjects told that their primary task was to form an impression of the
target with particular emphasis on how dispositionally happy or sad she
was.  
Half of the subjects were randomly assigned to the unregulated condition.
They were told that as the lm proceeded they should not concern
themselves with the TBI words because they are not relevant to the
condition of the experiment that he was in. (COULD IGNORE TBI
WORDS) 
Remaining subjects assigned to self regulated condition. These subjects
given the above instructions with regard to TBI words, but in addition
were told that the optiscope would not work if the subjects moved her
eyes too much. (SHOULD IGNORE TBI WORDS)
8. Dependent Measures
a) Perceived dispositional sadness 

Subjects to familiarise themselves with dispositional sadness
measures. Measures required subjects to rate the target on four 13
point bipolar scales that were anchored at the endpoints with the
phrases a) is generally happy (unhappy) sort of person, b) is probably
pretty cheerful (somewhat depressed) much of the time, c) is generally
a light hearted (troubled) sort of person, and d) probably has an
optimistic (a pessimistic) outlook on life.
b) Recall of interview questions 

Subjects had to complete dispositional measures just described.
c) Recall of TBI words 

Subjects asked to recall as many of TBI words as they possible could.
Assumed that self-regulated subjects might feel reluctant to admit that
they had seen the TBI words because they had originally been
124
fi
fi
fi
forbidden to do so. But subjects were told that due to parafoveal
processing, they can still recall the words.
d) Recognition of TBI words 

Subjects asked to circle TBI words that appeared in the lm and
encouraged to guess if they were unsure.
B. RESULTS AND DISCUSSION
1. Perceived Dispositional Sadness
a) Unregulated subjects considered target more dispositionally sad

when she appeared sad while answering happiness inducing
rather than sadness inducing questions.
b) Unregulated subjects used situational constraint information to

correct (i.e. to discount or to augment) their characterisations of
the target.
c) Self Regulated subjects considered the target to be equally sad,

regardless of the type of questions she had ostensibly answered.
Self regulated subjects showed no evidence of inferential
correction (i.e. of either discounting or augmenting).
2. Recall of Questions
a) Self regulation can usurp resources and thereby impair the ability
to use situational constraint information.  
Possible that self regulated subjects did not use the situational
constraint information simply because they did not have the
information in memory.
b) Subjects in the sad questions condition remembered more

questions than did subjects in the happy questions condition.
c) Although self regulated subjects were less likely than unregulated

subjects to use the situational constraint information, they were
equally likely to have that information in memory.
3. Recall and Recognition of TBI Words
a) Self regulated and unregulated subjects were equally accurate in

their recognition and recall of the TBI words.
b) Both correct recall and recognition of the TBI words were quite
low.
125
fi
c) Both self regulated and unregulated subjects successfully
ignored the TBI words, and it is somewhat ironic in this regard
that self regulated subjects evidently made a conscious and
costly e ort to avoid performing a behaviour that, had they been
unregulated, they would easily have avoided anyway.
III. EXPERIMENT 2
A. Background
1. Experiment 1 suggests that active self regulation of behaviour can impair

some ongoing cognitive operations (correction) without impairing others
(characterisation). Subjects who were merely asked to regulate their gaze
drew more dispositional inferences about a target whose behaviour was
situationally induced than did subjects who were allowed to gaze where
they wished.
2. Experiment 2 attempted to model this common social predicament of

how human relations shows that a person’s behaviour does not always
faithfully express his or her true feelings, beliefs, and desires. 
Asked female subjects to use non verbal behaviour to ingratiate a male
confederate whose political opinions the subject was eliciting in an
interview.  
Confederate was always constrained to give conservative responses to
the interviewer’s questions. Predicted that subjects who ingratiated a
dislikable rather than a likeable confederate would devote a great deal of
conscious attention to masking their true feelings, and thus would
particularly unlikely to use situational constraint information when drawing
inferences about the confederate’s true political opinions. 
In contrast, predicted that observers who were not required to ingratiate
the confederate would be just as likely to use situational constraint
information when drawing inferences about the opinions of a dislikable as
of a likeable confederate.
B. METHOD
1. Overview 
Pairs of female subjects met a male confederate who behaved in either a
likeable or dislikable manner. 
One of the subjects (interviewer) interviewed the confederate (responder)
about his opinions on several political issues while the other subject
(observer) watched the interview.  
Both subjects were told that regardless of his true opinions, the
126
ff
responder would read experimenter generated conservative responses to
the political interview questions. 
In addition to asking questions, the interviewer was instructed to
ingratiate the responder during the interview.  
Finally, all subjects attempted to estimate the responder’s true political
attitudes.
2. Subjects 
46 female students at University of Texas.
3. Likeability Manipulation 
Subjects invited to participate in experiment on “how people encode and
decode communications”. 
2 female subjects arrived at the laboratory and were escorted by a male
experimenter to a waiting area. 
Experimenter explained that a third subject (actually a male confederate)
was due to arrive momentarily. When confederate arrived, experimenter
left the waiting area to check on some equipment, and confederate either
engaged the 2 subjects in friendly conversation (likeable confederate
condition) or ignored them thoroughly (dislikable confederate condition). 
Experimenter returned after 1 minute and explained that before the
experiment could begin he would have to retrieve some materials from a
nearby building. In the dislikable condition the confederate was visibly
annoyed with the delay (e.g. You must be kidding, you mean we have to
wait?), whereas in the likeable condition the confederate assured the
experimenter the delay was no sweat. 
After experimenter left waiting area for second time, dislikable
confederate stormed out of the waiting area after telling the subjects, “If
i’m not back on time, tell that jerk to start without me”. Conversely, the
likeable confederate remained in the waiting area, behaved pleasantly,
and o ered both subjects a stick of gum during the experimenter’s
absence.
4. Interview Task 
After returning, the experimenter escorted subjects and the confederate
to individual cubicles where each remained for the rest of the experiment. 
One subject randomly assigned role of interviewer. From her cubicle the
interviewer could see and ostensibly be seen by the confederate via
CCTV. Interviewer was told that her task would be to read 7 questions to
the confederate who had been assigned the role of responder. 
Was explained that the responder would answer each of these questions
by reading an experimenter generated response, that the interviewer’s
127
ff
primary task was to diagnose the responder’s true attitude on each of the
issues. Experimenter claimed that this would help him learn “how people
decode the communication of others”.
5. Ingratiation Task 
Interviewer was told that part of the experiment concerned “how people
encode, or send communications via non verbal behaviour”. Interviewer
was asked to use non verbal means to ingratiate the responder during the
interview.  
Second subject assigned the role of observer. Observer was give same
information as interviewer. Observer was instructed to watch the ensuing
interview with the goal of diagnosing the responder’s true attitude on
each of the 7 political issues. 
Observer was of course, informed that the interviewer would be trying to
ingratiate the responder, and the interviewer was informed that the
observer would be watching the interview.
6. Dependent Measures 
Both subjects given sheet of paper containing seven 13 point bipolar
scales anchored at the endpoints with the phrases responder is opposed
to and responder is in favour of, each of 7 political issues (military
spending, ghting communism, legalised abortion, nuclear weapons, gun
control, capital punishment, and school prayer). 
Completed scale as questions progressed.
1. Manipulation Checks
a) Subjects liked the likeable responder a great deal more than they
liked the dislikable responder.
b) Those who ingratiated a dislikable responder found the task more

di cult than did those who ingratiated a likeable responder.
c) Interviewers behaved in a somewhat more likeable way toward

the likeable than the dislikable responder.
d) Interviewers who ingratiated a dislikable confederate apparently

put forth exceptional e ort with unexceptional results.
2. Ratings of Responder’s Political Attitudes
a) FACTOR 1: POLITICAL CONSERVATISM
128
ffi
fi
ff
(1) Observers attributed similar degrees of conservatism to both
the likeable and dislikable responder, but interviewers
attributed more conservatism to the dislikable rather than to
the likeable responder.
(2) Being friendly to someone whom we actually dislike is a

di cult task that impairs the ability to use situational
constraint information and thereby facilitates dispositional
attributions about that person’s behaviour.
(3) Observers may have been more cognitively busy than those
subjects who interviewed a likeable confederate.
(4) No di erences emerged between observers’ ratings of the

likeable and of the dislikable responder.  
Suggests that disliking a person does not in and of itself lead
to dispositional attributions about that person’s behaviour. 
Also suggests that there is no simple tendency to consider
dislikable persons politically conservative. Rather, expression
of false feelings and inhibition of true feelings appear to be
critical ingredients in the recipe for uncorrected
characterisation.
b) FACTOR 2: TRADITIONAL MORALITY
(1) De ned by conservative attitudes toward school prayer and

legalised abortion. Conservative beliefs on these issues re ect
adherence to traditional moral positions.
(2) Likeable responder was seen as holding more traditional

moral values than was the dislikable responder. 
Before they heard responder speak, subjects already had
some information about him, namely that he was nice, or that
he was nasty. Thus, subjects may have predicted the
responder’s attitudes on moral issues from their knowledge of
the responder’s likability alone, attributing high morality to the
likeable responder and how low morality to the dislikable
responder regardless of what the responder said.
(3) Those interviewers who ingratiated a dislikable responder

were also those who performed the greatest amount of
cognitive work.
129
ffi
fi
ff
fl
(4) Interviewers’ ratings of the responder’s likability were
signi cantly correlated with their ratings of the responder’s
traditional morality, but not with their ratings of the
responder’s political conservatism.
(5) However, interviewers’ ratings of how much cognitive work

they performed (i.e. their answers to the question “How
di cult was the ingratiation task?”) were signi cantly
correlated with their ratings of the responder’s political
conservatism but not with their ratings of the responder’s
traditional morality.
(6) These partial correlations reveal the independent e ects of

cognitive work and perceived likability on interviewers’
attributions.
(7) Cognitive work increased the interviewers’ tendency to infer

conservatism but not morality from the responder’s behaviour,
whereas perceived likability increased the interviewers’
tendency to infer morality but not conservatism from the
responder’s behaviour.
(8) Observers attributed the same degree of conservatism to the

likeable and dislikable responders.
(9) Observers’ judgements of morality were just as strongly

a ected by the responder’s likeable or dislikable behaviour as
were the interviewers’ judgements of morality.
c) FACTOR 3: GUN CONTROL
(1) Uncorrelated.
IV. GENERAL DISCUSSION
A. GENERAL
1. For most people, social life is an extended exercise in the self

regulation of action.
2. Self regulation seemed to impair subjects’ abilities to use situational

constraint information when interpreting another’s actions.
3. Self regulators draw dispositional inferences about behaviours that

could easily have been explained with reference to situational forces.  
130
ff
ffi
fi
fi
ff
E ect was obtained when the to be regulated behaviour involved
glancing at an irrelevant word on a video monitor or when it involved
expressing one’s true disdain for another.
4. Self regulatory e orts can a ect the inferences perceivers draw

about those with whom they interact.
5. Not all self-regulatory attempts will necessarily impair social

inference process in this way. Some social rules (e.g. 3 of the 10
commandments) cross virtually all situational boundaries, and it
seems likely that ubiquitous restrictions are themselves over learned
and that their enforcement therefore requires little cognitive work.
Other restrictions however, do demand a high sense of occasion
(e.g. revelation of sexual proclivities or personal failures may be
appropriate towards intimates but not among strangers, etc.).
6. Self regulations are more easily imposed than others, and social
inference is most likely to be a ected when perceivers attempt to
oblige unfamiliar, temporary, or local norms.
B. SELF REGULATION IN EVERYDAY LIFE (JUDGING POWERFUL PERSONS)
1. People tend to draw dispositional inferences about those with

superior status (Thibaut & Riecken, 1955), and about those on whom
their outcomes depend (Berscheid, Graziano, Monson, & Dermer, 1979;
Miller, Norman, & Wright, 1978; cf. Erber & Fiske, 1983).
2. Powerful people may simply impair one’s ability to process

information.
3. Perceivers may embrace dispositional explanations of powerful

persons’ behaviours simply because the perceiver’s own self
regulatory actions have impaired his or her ability to draw accurate
social inferences.
C. JUDGING THE OUT GROUP
1. When individuals interact with members of other races, genders, and

nationalities, they often draw more dispositional inferences from
those behaviours than from the identical behaviours of their cohort.
Generally explained with reference to lack of cognitive complexity.
131
ff
ff
ff
ff
2. People think of out-groups in simple ways and are thus overly
in uenced by the unrepresentative actions of out group members
(Linville & Jones, 1980; Quattrone & Jones, 1980).
3. Unfamiliarity with the norms that govern an out group member’s

behaviour may leave one unable to estimate the consensus that such
actions enjoy. (E.g. one may assume a Japanese acquaintance is
particularly reserved without realising that such reserve is demanded
by Japanese custom and as such, tells one little about the
acquaintance as a unique individual).
4. One is often unsure of the norms that should govern one’s own
behaviour in the company of out group members, just as one often
knows little about the norms that govern an out group member’s
behaviour.
5. Interaction with out group members may cause individuals to be

overly cognizant of their own actions, which may in turn impair their
ability to draw accurate inferences about the out group members
with whom they are interacting.
D. JUDGING THE SELF
1. People often come to believe that which they once merely professed;
that is, people often do not use information about the situational
constraints on their own behaviour (Bem, 1972, and Wicklund & Brehm,
1976).
2. Everyone is from time to time, induced to mask some sentiments

and to express others. This is demanding, and may cause one to fail
to consider the very forces that led one to lie in the rst place. The
more elaborate and di cult the lie, the more likely one should be to
believe that one believes it. A person who is subtly induced to tell
another that a mundane peg-turning task is in fact quite thrilling may
at one level recognise that this lie was coerced.
3. Because telling a lie with apparent conviction requires a great deal of

e ort and attention, the person may not be able to use the coercion
information when drawing inferences about their own beliefs.
4. Responding to social pressures may leave us too busy to think about

our beliefs.
132
ff
fl
ffi
fi
E. LIARS AND OTHER SELF REGULATORS
1. Lying can be a strenuous labor that leaves liars especially vulnerable

to others’ lies.
2. The more attention one’s own actions require, the less attention one
can spend drawing inferences about the actions of others.
3. A successful lie must be told again and again, and this eases
cognitive burden over time.
4. A liar may not trust themselves, fail to correct their characterisation

of others, but they also place little faith in these characterisations
simply because they have learned from experience how easily one
individual’s view of another can be manipulated.
5. Habitual liars may to some extent disregard the inferences they draw
during interaction.
6. When people tell lies they often leak their true feelings through non
verbal behaviour. (Gilbert and Krull, 1988). When verbal and non verbal
behaviours are at odds, cognitive busyness can uniquely impair the
perceiver’s ability to draw dispositional inferences from verbal
behaviour. 
Thus, cognitively busy perceivers (of which liars are one sort) often
rely on them to be more accurate lie detectors. On balance, then, all
of these points suggest that liars and other self-regulators may well
achieve a limited immunity to the lies of others.
133
Week 11 Readings (Motivation and A ect)

MOTIVATION & AFFECT (1)
Je Greenberg, Tom Pyszczynski, Sheldon Solomon, Abram Rosenblatt, Mitchell

Veeder, Shari Kirkland, Deborah Lyon. Evidence for Terror Management Theory II:
The E ects of Mortality Salience on Reactions to Those Who Threaten or
Bolster the Cultural Worldview
I. ABSTRACT
A. 3 experiments conducted to test the hypothesis, derived from terror

management theory, that reminding people of their morality increases
attraction to those who consensually validate their beliefs and decreases
attraction to those who threaten their beliefs.
B. Study 1: Subjects with a Christian religious background were asked to form

impressions of Christian and Jewish target persons. Before doing so,
mortality was made salient to half of the subjects. 
Supporting hypothesis, mortality salience led to more positive evaluations of
the in group member (Christian) and more negative evaluations of the out
group member (Jew).
C. Study 2: Mortality salience led to especially negative evaluations of an

attitudinally dissimilar other, but only among subjects high in
authoritarianism.
D. Study 3: Mortality salience led to especially positive reactions to someone

who directly praised subjects’ cultural worldviews and especially negative
reactions to someone who criticised them.
E. Study allowed us to understand in group favouritism, prejudice, and

intolerance of deviance.
F. Terror Management Theory was based from Ernest Becker (1962, 1973,
1975). This attempts to synthesise the various social science disciplines into
a general theory of human social behaviour. According to Becker,
sophisticated human intellectual abilities lead to the awareness of people’s
vulnerability and ultimate mortality. This awareness creates the potential for
paralysing terror concerning the vast array of aversive experiences that are
the eventual inevitability of death. As these abilities evolved, cultural
conceptions of reality began to emerge, and the potential for terror put a
press on evolving cultures such that any culture that was to survive needed
to provide means of managing this terror.
134
ff
ff
ff
G. Cultural Anxiety Bu er 
The means of conceptualising reality that allows for the possibility of
equanimity in the face of human vulnerability and mortality. People’s beliefs
about reality provide a bu er against the anxiety that results from living in a
largely uncontrollable, perilous universe, where the only certainty is death.
Consists of 2 components.
1. Faith in the validity of a cultural conception of reality that provides

meaning, standards of value, and the promise of immortality.
2. Belief that one is meeting or exceeding those cultural standards of value.
II. STUDY 1 
Hypothesis that mortality salience would increase liking for a member of one’s
own religious group (Christians) and decrease liking for a member of a religious
out group (Jews).
A. METHOD
1. Subjects 
26 female and 20 male Christian introductory psychology students
participated. 
6 additional subjects who identi ed themselves as having Jewish parents,
but data were not used.
2. Procedure 
4-5 subjects participated in each session. Subjects were told that the
study concerned personality and attitude variables that a ect the
impressions people form of each other, and that they would accordingly
ll out some personality questionnaires and then be given personality
information supplied by 2 other subjects that they would use to evaluate
those subjects. 
Explained to ensure privacy and control over factors that a ect
impressions. 
Subjects lled out a preliminary questionnaire packet at their cubicles,
containing the mortality salience manipulation for half of the subjects. 
Asked for their age, sex, no. of siblings, asked for their parents’ religious
a liations (select subjects for inclusion in the study). 
Also asked to ll out the Janis Field Self Esteem Inventory (Eagly, 1967),
and also a social issues survey that asked them to indicate on 10 point
scales their level of agreement with 10 social issue statements. 
Mortality salience manipulation placed at the end, write about what will
happen to them as they physically die and the emotions that the thought
of their own death aroused in them. 
135
fi
ffi
fi
fi
ff
ff
fi
ff
ff
When subjects completed questionnaires, experimenter collected those
materials, left and then returned with the background questionnaires. 
Subjects received 2 sets of initial impressions assessments and an
envelope, 2 versions of the set questionnaires were available, one lled
out by a Jew and the other by a Christian. 
Religious a liation manipulation had 2 parts. On background
questionnaire, parents’ religious a liations were Jewish for both parents
for the Jewish target and Methodist and Lutheran for the Christian target. 
On the Who Am I? questionnaire, the third response was either Jewish or
Christian.  
Di erent attitudes were presented to increase realism of targets and their
responses on background questionnaire and social issues questionnaire. 
Subjects were instructed to evaluate the subjects in a speci c order and
ll out the rst impression assessment as soon as they nished reading
about the rst subject. 
Each of the 2 initial impression assessments (one for each target) began
by asking subjects to indicate their current mood. Subjects indicated on 9
point scales the extent to which they felt happy, calm, irritated, secure,
angry, disturbed, hostile, and frustrated. Followed by Interpersonal
Judgement Scale (Byrne, 1971), which asked subjects to rate the target’s
intelligence, knowledge of current events, mortality, adjustment, and the
extent to which they would like and enjoy working with the target.
(Indicating how applicable each of 20 characteristics was to the target). 
Following traits of stingy, manipulative, arrogant, snobbish, and
obnoxious was included because they t the negative stereotype of Jews
portrayed in anti-Semitic literature. Other 15 traits were honest, cheerful,
reliable, trustworthy, argumentative, intelligent, warm, patient, kind,
ambitious, stable, sleazy, introverted, spineless, and impulsive.
3. RESULTS 
2 counterbalanced variables, namely (a) which version of the
questionnaire responses appeared to be Jewish, and (b) order of
presentation (whether the Jew or Christian was evaluated rst). 
Because data from non-Christians and the suspicious subject had to be
discarded, neither of these variables was perfectly counterbalanced,
therefore deemed important to include the counterbalanced factors in
preliminary analyses of ANOVAs as between group factors. 
Primary analyses to be reported were 2 (Mortality Salient vs. Non Salient)
x 2 (Christian First vs. Jew First) between subjects x 2 (Christian vs.
Jewish Target) within subjects, unweighted means, least squares
ANOVAs.
136
fi
ff
fi
fi
ffi
ffi
fi
fi
fi
fi
fi
a) Interpersonal Judgement Scale IJS 
Mortality salience increased attraction to the Christian, and
decreased attraction to the Jew. 
The Christian was liked more than the Jew only in the mortality
salient condition.
b) Trait Applicability Ratings 

Although mortality salient subjects, compared with control
subjects, rated the Christian more positively and the Jew more
negatively, the di erence was signi cant only for ratings of the
Christian. 
As with IJS scores, the Christian was rated signi cantly more
positively than the Jew only in the mortality salient condition.
c) Analysis of Overall Trait Rating Index 

3 way interaction revealed. 
Christians was rated signi cantly more positively than the Jew
only in the mortality salient Christian rst condition. 
Within the Christian rst condition, mortality salient subjects
rated the Jew signi cantly more negatively than did control
subjects.
d) Analysis of Negative Stereotype Index 

Expected assignment of greater negative stereotypic traits to the
Jew in the mortality salient condition occurred, but only when the
Christian was evaluated rst. 
In the mortality salient Christian rst condition, the Jew was
ascribed more negative stereotypic traits than the Christian, and
more negative traits than the Jew in the control Christian rst
condition. 
For nonstereotypic traits, mortality salience enhanced ratings of
the Christian but did not reduce ratings of the Jew. In the
mortality salient condition, subjects rated the Christian more
positively than the Jew, and rated the Christian more positively
than did subjects in the mortality non salient condition.
e) Mood Measures 
Mortality salient subjects reported being more disturbed than did
control subjects. 
Mortality salient subjects were more frustrated than control
subjects at both the rst and second assessment. They were
137
ff
fi
fi
fi
fi
fi
fi
fi
fi
fi
fi
more frustrated at the second assessment than at the rst
assessment.
4. DISCUSSION
a) Inducing Christian subjects to think about their mortality led them

to give more positive IJS ratings of fellow Christians and more
negative IJS ratings of Jews. This was the case for the overall
trait applicability ratings, and stereotype ratings.
b) On the negative Jewish stereotype ratings, mortality salience

again led to more negative trait ascriptions to the Jewish target
only when the Christian was rated rst. 
However, on items not relevant to the negative Jewish stereotype,
mortality salience increased the positivity of trait ascriptions to
the Christian target regardless of who was rated rst.
c) Across all measures, the Christian was rated more positively than
the Jew only in the mortality salient condition.
d) Findings show that positive reactions towards in group members

and negative reactions to out group members are mediated by
the implications that such individuals have for the individual’s
cultural anxiety bu er. 
According to Terror Management Theory, beliefs about the nature
of reality serve to bu er the anxiety that results from awareness
of human vulnerability and mortality. Reminding subjects of their
mortality increases their need for the anxiety bu er provided by
their beliefs and consequently increases the intensity of their
reactions to those who bolster and threaten those beliefs.
e) Mortality Salience decreased the favourability of trait ascriptions

to the out group member only when an in group member had
been previously evaluated. Perhaps the initial rating of the in
group member served to remind subjects of the particular
dimension of their beliefs that was threatened by the out group
member. 
Alternatively, rating the in group member may have created an
anchor against which ratings of the out group member could be
contrasted.
138
ff
ff
fi
ff
fi
fi
f) Struch and Schwartz (1989) found that aggression toward
ultraorthodox Israeli Jews by other Israeli Jews was correlated
with perceived con ict of interests and di erences in basic values
between groups, in addition, these relationships were mediated
by the degree to which subjects identi ed with their own group (in
group identi cation). 
This indicates that only individuals who strongly subscribe to a
particular worldview will react negatively to those perceived to
have an anti ethical alternative worldview.
(1) Current experiment’s ndings found instead, that evaluating an

in group member rst sometimes, but not always, encourages
negative evaluations of out group members. More research
needed on this.
g) Clear that religious a liation is a type of group membership that

seems especially relevant to coping with one’s fear of death.
Almost all religions provide explanations for existence and beliefs
about how death can be transcended.
III. STUDY 2 
E ects of mortality salience on reactions to those with similar and dissimilar
attitudes were observed. Addressed potential individual di erences in
interpersonal responses to mortality salience. 
Predicted that high authoritarians would be especially likely to respond to
existential threat by derogating dissimilar others and reacting positively to
similar others. Fear of death and vulnerability plays a central role in promoting
authoritarianism, predicted that di erences between high and low authoritarians
would be greater when mortality was made salient.
A. METHOD
1. Subjects 
60 male and 107 female students in lower level psychology courses. 
Subjects randomly assigned to conditions in a 2 x 2 x 2 (High
Authoritarianism vs. Low Authoritarianism x Similar vs. Dissimilar x
Mortality Salient vs. Control) factorial design.
2. Procedure 
Subjects participated in groups ranging in size from 3-6.  
Subjects led to individual cubicles and were told that the experiment
concerned problem solving behaviour. Subjects were also told that there
were 2 parts to the experiment. First part involve completing
139
ff
fi
fi
fl
ffi
fi
ff
fi
ff
ff
questionnaires, and second part involved them completing a problem
solving interaction with a randomly chosen partner. 
Subjects were given consent forms to sign and the experimental sessions
began (although actually only the rst part of the experiment took place).
They received a packet containing a survey of attitudes and the F-Scale
(Adorno et al., 1950). Survey of attitudes was used for attitude similarity
manipulation. Consisted of 12 items concerning a variety of issues (e.g.
discipline of children, sports, role of women in society, and university
grading system). 
Subjects simply asked to indicate whether they agreed or disagreed with
each statement. Every subject’s survey used to construct a bogus survey
supposedly lled out by their partner that appeared to be either quite
similar (75% item agreement) or dissimilar (25% item agreement) to the
subject’s own. 
After completing attitudes survey, subjects lled out the F scale. 
Questionnaires used to manipulate mortality salience were distributed,
and subjects were randomly assigned to ll out questionnaires
concerning either death or their favourite foods. 
Bogus attitude surveys were prepared by assistants in the meantime. 
Subjects were randomly assigned to either similar or dissimilar condition
for partner’s attitude surveys. 
Subjects then lled out a nal questionnaire containing measures of
attraction to the target, perceived similarity, and current mood state. 
Finally, mood state was assessed on a set of 10 mood adjectives.
B. RESULTS
1. MANIPULATION CHECKS 
Main e ect of the similarity manipulation. 
Subjects in the similar condition rated the target as more similar to
themselves than did subjects in the dissimilar condition.
2. EVALUATION OF PARTNER 
Main e ect for similarity was found. 
Subjects liked similar partners more than dissimilar partners. 
Predicted Threat x Similarity x Authoritarianism interaction was also
found. 
High authoritarians liked dissimilar others less under mortality
salient conditions than under control conditions. 
Low authoritarian subjects’ liking for their partners was not
signi cantly a ected by the mortality salience manipulation in either
condition. The only condition under which high authoritarians
140
fi
ff
ff
fi
fi
ff
fi
fi
fi
fi
di erent from low authoritarians was the mortality salient dissimilar
other condition. 
High authoritarians were more negative than low authoritarians in
their evaluations of dissimilar others only in the mortality salient
condition.
3. MOOD RATINGS 
No evidence that mood mediated the e ect of mortality salience.
C. DISCUSSION
1. Results of study supports hypothesis that mortality salience

encourages high authoritarians to derogate dissimilar others. In fact,
high authoritarians di ered from lows in their reactions to dissimilar
others only when mortality had been made salient.
2. Mortality salience enhanced the rejection of dissimilar others in

Study 2 only among high authoritarian subjects. 
Low authoritarians showed no hint of this tendency toward
increased derogation of dissimilar others when mortality was salient.
3. Raises possibility that low authoritarians may not be concerned with

mortality or do not engage in terror management. However, this is
unlikely. 
There is evidence suggesting that anxiety about death is widespread,
even among individuals who deny it on self-report measures
(Alexander, Colley, & Adlerstein, 1957; Templer, 1971). 
Second, authors have found powerful e ects of mortality salience in
a variety of other studies (reported here and in Rosenblatt et al.) in
which authoritarianism was not considered. 
Third, low authoritarian individuals’ world views are usually open
minded and tolerant of di ering opinions.
IV. STUDY 3 
Primary purpose of Study 3 was to test the possibility that if, as Terror
Management Theory posits, implicit validation and threat motivate these
reactions, then we would expect mortality salience to have even stronger e ects
on reactions to those who directly validate or threaten aspects of subjects’
cultural worldviews.
A. METHOD
141
ff
ff
ff
ff
ff
ff
1. Subjects 
70 male and 81 female American introductory psychology students
2. Procedure 
Subjects arrived in the lab in groups of 4-5. 
Seated so that they could not see each other’s forms. 
Subjects told the study concerned the relationship between personality
and political attitudes. 
Anonymity guaranteed. 
Given a packet to complete. Mortality salience manipulation, an
interviewee credentials page, the interview, and the primary dependent
measures assessing reactions to the interview.
a) Mortality Salience Manipulation 

Same questionnaire used in Studies 1 and 2.
b) Favourability of Interview to the Culture 

Subjects read a one page interview in which the interviewee evaluated
the U.S. political system. 
Pro US version (admitted US mistakes but was positive), Mixed
version (acknowledged value of US system but focused on more
negative aspects), and Anti US version (very hostile to the US
government).
c) Interviewee Credentials 
Half the subjects were told the interviewee was a Harvard political
science professor, and other half were told he was a president of the
American Communist Party.
d) Dependent Measures 
Subjects asked to indicate their emotional state at the current
moment. Based on Izard’s 1977 Di erential Emotions Scale. 
Subjects asked a series of 8 questions to assess feelings about
interview and interviewee. (How much they liked and wanted to meet
him, how much they would like to have him as an instructor, and how
knowledgeable they felt he was).
B. RESULTS
1. PRIMARY DEPENDENT MEASURES 

Consisted of 6 questions concerning liking for the interviewee and
agreement with his remarks.  
The more favourable the interviewee was to the culture the more
subjects generally liked him. However, the interaction occurred
142
ff
reliably only in the mortality salient condition. 
In the mortality salient condition, the positive interviewee was rated
signi cantly higher than both the mixed and negative interviewee. In
addition, the negative interviewee was liked signi cantly less than
the mixed interviewee. They also liked the positive interviewee more
and the negative interviewee less than did mortality non salient
subjects. 
Unexpectedly, there was no evidence that the credentials of the
communicator in uenced the extent to which subjects liked him.
2. ANALYSIS OF AGREEMENT MEASURE 

The more favourable the interviewee’s remarks, the more subjects
generally agreed with him.  
Within the mortality non salient condition, the positive interviewee
was agreed with more than the negative interviewee. 
Within the mortality salient condition, the positive interviewee was
agreed with more than the mixed interviewee, who was agreed with
more than the negative interviewee. 
Mortality salient subjects agreed more with the positive interviewee
and less with the negative interviewee than did mortality non salient
subjects. 
Credentials of the communicator had no e ect on subjects’
agreement with the communication.
3. AFFECT MEASURES 
Subjects tend to be surprised by the positive remarks when
attributed to the communist and by the negative and mixed remarks
when attributed to the Harvard professor. 
Main e ect of interview in the Mortality Salience x Interview
Interaction on ratings of how disturbing the remarks were. 
Subjects were generally less disturbed by the positive interview than
by the other two interviews. 
Mortality salient subjects tended to rate the negative interview as
more disturbing than did mortality non salient subjects. In contrast,
mortality salient subjects tended to rate the positive interview as less
disturbing than did mortality non salient subjects. 
Disturbed feelings did not substantially mediate interactions
according to ANOVA.
143
fi
ff
fl
ff
fi
C. DISCUSSION
1. Study 3 provided strong support for hypothesis that mortality

salience encourages positive reactions to someone who praises the
culture and negative reactions to someone who sharply criticises the
culture.
2. Contrary to expectations, credentials of interviewee did not alter the

primary e ects of mortality salience. Credentials had no e ects on
liking and agreement variables.
3. Fact that the “Harvard professor” was not even liked more or agreed
with more than the communist in general suggests that those
particular credentials do not di erentially in uence liking or
perceived credibility. Another possibility is that the vivid, compelling
nature of the interviewee’s remarks overwhelmed any potential e ect
of credentials. It may also be that subjects went out of the way to be
fair and judge the interviewee on the merits of his remarks over his
background.
V. GENERAL DISCUSSION
A. STUDY 1
1. Mortality salience led Christian subjects to make more positive

evaluations of a fellow Christian and more negative evaluations of a
Jew. When the Christian was rated rst, mortality salience also led
these subjects to ascribe more negative stereotypic traits to the
Jewish target person.
2. Mortality salience a ected the way Christian subjects evaluated

Christian and Jewish target persons.
3. Suggests that Terror Management plays a signi cant role in in group

favouritism and prejudice. Terror Management Theory may help
account for the historic role of Jews as victims of prejudice and
discrimination. 
Jews may have been a target of prejudice wherever and whenever
they have been for their longstanding faith has typically been viewed
as a serious threat to the cultural anxiety bu ers of the majority
groups around them.
144
ff
ff
ff
fi
fl
ff
fi
ff
ff
4. Religious background is one aspect of the individual’s cultural
worldview that is very explicitly tied to beliefs about death and the
possibility of an afterlife. They are important components of the
anxiety bu ering world view, but are not the only component. All
important beliefs and values are tied to overall conception of reality
that makes it possible for the individual to maintain equanimity.
5. Theory predicts that mortality salience should have similar e ects on

evaluations of others who di er with respect to other aspects of the
value system.
B. STUDY 2
1. Mortality salience led authoritarian subjects to more negative

evaluations of a potential interaction partner whose attitudes on a
broad range of topics were dissimilar to their own. Mortality salience
had a catalytic e ect on subjects’ authoritarian personality
tendencies in that high authoritarians were more rejecting of
dissimilar others than were low authoritarians only when their
mortality had previously been made salient
2. Increased rejection of others who di er merely on attitudinal

dimensions may be especially prominent among individuals high in
authoritarianism. From Terror Management Perspective, personality
characteristics are conceptualised as habitual styles of responding
to existential problems. The nding that high and low authoritarians’
evaluations of their partners di erent from the others only when
mortality was made salient is consistent with this interpretation.
3. Along with Adorn et al.’s (1990) contention that the authoritarian

personality is basically a defensive way of coping with threats, this
nding also suggests that death and vulnerability may be issues with
which authoritarians are especially troubled with.
C. STUDY 3
1. Mortality salience encouraged positive reactions to one who praised

the culture and negative reactions to one who criticised it.
2. Mortality salience intensi ed reactions of American subjects to one

who directly attacked or supported faith in the cultural world view.
145
fi
ff
ff
fi
ff
fi
ff
ff
ff
3. Reactions to the target depended exclusively on his view of the
culture and not at all on his group status. Whether he was a
legitimate in group authority gure or a member of a deviant group
had no impact on evaluations.  
In contrast, his remarks about the in group had a large e ect (US),
especially in the mortality salient condition.
4. Similar to ndings concerning role of perceived values in prejudice

(Byrne & Wong, 1962; Rokeach & Mezei, 1966), the implications of the
other’s behaviour for one’s cultural anxiety bu er, rather than group
membership per se, determined the individual’s reaction to that
person.
5. Terror Management may play a role in both the promotion of

nationalism, and this may explain the censorship and persecution of
those courageous or foolhardy enough to challenge central aspects
of a popular world view.
D. Findings support Terror Management Theory. Liking for others depends

largely on the impact such persons have on one’s cultural anxiety bu er.
Others with similar worldviews are liked because they provide
consensual validation for the individual’s own world view. Others with
dissimilar worldviews are disliked because they make salient the lack of
consensus for the individual’s beliefs and thus threaten faith in those
beliefs.
E. People should be especially motivated to obtain consensual validation

of beliefs when faced with reminders of their mortality. Consequently,
liking for similar others is increased and liking for dissimilar others is
decreased when mortality is made salient.
146
fi
fi
ff
ff
ff
MOTIVATION & AFFECT (2)
Vanessa Sawicki, Duane T. Wegener, Jason K. Clark, Leandre R. Fabrigar, Steven M.

Smith, and Steven T. Bengal. Seeking Con rmation in Times of Doubt: Selective
Exposure and the Motivational Strength of Weak Attitudes. 2011.
I. ABSTRACT
A. Strong attitudes exert greater in uence on social perceptions, judgements,

and behaviours. Some research indicates that strong attitudes are
associated with exposure to attitude con rming information. However,
uncertain attitudes might produce strong selective exposure to attitude
consistent information, especially when available information is unfamiliar.
B. 3 Experiments, participants reported attitude favourability, reported attitude

con dence (Study 1A and 2) or completed a doubt priming manipulation
(Study 1B), and selected information supporting or opposing an issue. 
When chosen information was relatively unfamiliar in all three studies,
uncertainty led to more attitude consistent exposure than certainty did.
However, when chosen information was more familiar (Study 2), the pattern
of e ects was signi cantly reversed. Certainty led to more attitude consistent
exposure than did uncertainty. Findings suggests that under certain
conditions, uncertainty can motivate people to seek attitude con rming
information, thereby creating a motivational basis for weak attributes to have
strong in uences on information seeking.
C. Past research indicates that people often seek attitude consistent

information (Smith, Fabrigar, & Norris, 2008).  
When people cannot consider all relevant information (e.g. time restriction,
information choices are costly), attitudes can serve as a shortcut in choosing
what to consider. When people hold a particular attitude, it indicates to them
what the “correct” position is (Festinger, 1950), which can guide their choices
of additional information. Yet if the person does not have a strong opinion
one way or the other, information seeking might be more balanced.
D. ATTITUDE STRENGTH AND INFORMATION SEEKING 

Conditions most likely to produce selective exposure e ects have been
studied in some detail (Frey, 1986; Hart et al., 2009; Smith et al., 2008), but
relatively little research addressed the types of attitudes that could enhance
preference for attitude consistent information. 
Consistent with the idea that attitudes can serve as a heuristic or shortcut to
identify which information to consider, some research suggests that strong
attitudes can serve as a stronger cue to selective exposure. (Brannon, Tagler,
and Eagly, 2007). 
147
fi
ff
fl
fi
fl
fi
fi
ff
fi
Hart et al. (2009) provided additional support for the attitude strength
hypothesis, as involving attitudes higher in commitment or relation to
particpants’ values produced stronger tendencies for attitude consistent
exposure. Analysis also included some unpublished research that measured
attitude certainty and showed greater attitude consistent exposure when
con dence was high rather than low (Brechan, 2002). These studies suggest
that attitudes associated with greater certainty result in greater attitude
consistent selective exposure, which makes sense if one considers attitudes
as a cue in guiding information choices.
E. MOTIVATION TO BOLSTER UNCERTAIN ATTITUDES 

Con dence and other strength related attitudinal properties might have
di erent implications when one considers the motives they produce. 
From motivational perspective, doubt, rather than certainty might often lead
to strong in uences of attitudes on selective exposure. 
Possibility of uncertainty motivating selective exposure also makes sense
from the perspective of the su ciency principle from the heuristic systematic
model of persuasion (Chaiken, Liberman & Eagly, 1989). According to this
principle, people generally strive to hold their attitudes with certainty, and
when actual con dence falls short of desired con dence, the person will
process information with the goal of gaining conviction.
II. STUDIES 1A AND 1B 

Test hypothesis that relatively uncertain rather than certain attitudes would
produce greater attitude consistent information seeking in a traditional selective
exposure paradigm.  
After reporting attitude favourability and attitude con dence toward nuclear
power (Study 1A) or completing a doubt priming manipulation (Study 1B),
participants were allowed to choose from a list of 8 articles (Study 1A) or 10
articles (Study 1B) with titles making it clear that the article argued for or against
building additional nuclear power plants. 
Study 1B was designed to not only incorporate an experimental manipulation of
attitude con dence (to address potential confounds with con dence measure)
but also address the generalisability of uncertainty e ects to conditions of
attitude expressions (which generally enhance selective exposure to attitude
consistent information; Smith, Fabrigar, Powell, & Estrada, 2007).
A. Method

Undergraduates, 100 for 1A, and 42 for 1B.  
Favourability of pre-exposure attitudes and attitude con dence were
measured (Study 1A), or con dence was manipulated (Study 1B).
148
ff
fi
fi
fl
fi
fi
fi
ffi
fi
ff
fi
fi
fi
2. Procedure 
1-7 individuals participated in each session. Participants completed the
study materials at a computer running MediaLab 2006, while visually
isolated from one another. 
Study presented as investigation of general impressions toward various
social issues. 
Participants to select information to read about the issue of potential
building of new nuclear power plants. 
Before information exposure, participants reported their attitudes on a
variety of issues (e.g. senior comprehensive exams, legalisation of
marijuana), including the topic of interest (nuclear power). 
Study 1A, participants reported strength related properties toward a
number of the issues. 
Study 1B, con dence was manipulated through a priming procedure, and
attitude expression instructions were presented immediately before the
exposure task but after a ller task.
3. Predictor Variables
a) Pre Exposure Attitude 

Participants reported their attitudes before exposure, toward the
building of additional nuclear power plant using 2 x 9 point scales. 
Scores were highly correlated and averaged to form an index of
preexposure attitude.
b) Attitude Con dence 

Study 1A, after reporting attitudes toward nuclear power, participants
completed the attitude certainty measure consisting of 2 items. 
Study 1B, participants were randomly assigned to recall previous
experiences in which they felt either doubt or con dence (Petty,
Brinol, & Tormala, 2002). During a 5 minute period, participants were
provided with 5 boxes in which to describe these doubtful or
con dent experiences.
4. Dependent Measure: Information Selection 

Participants given 2 minutes to select paragraphs to read from a list of
paragraphs concerning the building of additional nuclear power plants. 
List included brief descriptions of the information that would be
presented so that participants would know if they were choosing a
paragraph that supported or opposed nuclear power. 
Half paragraphs supported nuclear power, and half opposed it. 
Participants had the option to refrain from reading any information.
149
fi
fi
fi
fi
fi
B. RESULTS
1. Study 1A showed main e ect of pre-exposure attitude on

favourability of information chosen.  
Participants generally selected information that supported their
position on nuclear power. There was also main e ect of attitude
con dence. Lower con dence was associated with choosing pro-
nuclear power information.
2. Predicted Pre-Exposure Attitude x Con dence interaction was

signi cant in both studies. When participants were relatively
uncertain about their attitudes (attitude con dence one standard
deviation below the mean) or had recalled past experiences of doubt,
people with more favourable pre-exposure attitudes toward nuclear
power selected more pro-nuclear information to read.  
Participants who felt doubtful preferred information that was
consistent rather than inconsistent with their attitudes. However,
when participants were relatively con dent (attitude con dence one
standard deviation above the mean) or had recalled personal
experiences of con dence, the relation between pre-exposure
attitude and favourability of information chosen was non-signi cant.
C. DISCUSSION
1. Studies 1A and 1B reliably demonstrated that having doubt in one’s

attitude can enhance biases toward exposure to attitudinally
consistent information.
2. Feeling shaky about an attitude can lead to greater use of that

attitude in selecting new information. This nding is directly opposite
to the idea that strong attitudes would be a stronger cue in guiding
information choices (cf., Brannon et al., 2009).
3. Study 1B provided the rst experimental evidence of certainty’s/

uncertainty’s e ect on selective exposure.  
The stronger impact of uncertain attitudes on information choices
occurred when research participants did not expect to express their
attitudes after information exposure (Study 1A), and when they did
expect to express their attitudes (Study 1B). 
Con dent attitudes did not signi cantly predict information choices,
150
fi
fi
fi
ff
fi
fi
fi
ff
fi
fi
fi
fi
fi
ff
fi
fi
thus suggesting that con dence may lead to relatively even handed
examination of information (cf., Albarracin & Mitchell, 2004).
4. Uncertain attitudes can create stronger attitude consistent selective

exposure than can certain attitudes, but studies also beg the
question of when the current e ects occur instead of strong
attitudes having stronger e ects on selective exposure (Brannon et
al., 2007; Knobloch-Westerwick & Meng, 2009).  
One could have greater con dence in the current e ects if replicated
using a topic other than nuclear power.
III. STUDY 2 
Replicate and extend ndings of Studies 1A and 1B using new attitude object of
ca eine consumption. 
Seemed plausible that issue familiarity might be one dimension with important
consequences for the link between certainty and information choices. Familiar
and unfamiliar information may not be equally capable of meeting goals present
at information selection time. 
Motives are likely to be di erent when attitude con dence is high, and unfamiliar
information would require additional thought with less assurance that even
agreeable information would support the high level of con dence one currently
holds. 
Study 2 was designed to test moderating role of information familiarity in the
relationship between attitude con dence and selective exposure.
A. METHOD 
Similar to study 1A except that ca eine consumption served as attitude
object.
1. Predictor Variables 
Pre-exposure attitude and con dence measures. 
Information familiarity.
B. RESULTS
1. People who reported greater con dence in their attitudes toward

ca eine consumption rated the information they chose as being
more familiar than did people who reported less con dence in their
attitudes.
2. Consistent with Studies 1A and 1B, relatively uncertain attitudes

predicted information choices, but relatively certain attitudes did not.
In other words, when uncertain in their ca eine attitudes,
151
ff
ff
fi
ff
fi
ff
fi
fi
ff
fi
ff
fi
ff
fi
ff
fi
fi
participants selected attitude consistent unfamiliar information.
When con dent about ca eine stances, participants’ attitudes did
not predict exposure to unfamiliar information.
3. However, when selected information was more familiar, the opposite

pattern emerged. 
Similar to past research (Brannon et al., 2007; Knobloch-Westerwick &
Meng, 2009), relatively certain attitudes tended to predict information
choices, but relatively uncertain attitudes did not.
4. When con dent about their ca eine attitudes, participants tended to

select attitude consistent familiar information. When uncertain about
their ca eine stances, participants attitudes did not predict exposure
to familiar information.
C. DISCUSSION
1. Results of Study 2 replicated the ndings from the prior ndings but
with di erent attitude object. A potential moderator was also
identi ed to account for when attitude uncertainty or certainty
creates attitude consistent selective exposure.
2. Doubt rather than con dence fostered a preference for information

that agreed with the person’s attitude when that information was
unfamiliar rather than familiar, perhaps because new information that
agrees with a doubtful position is the clearest way to increase
con dence.
3. When information was familiar however, uncertainty did not lead to

con rmation bias. Makes sense because even if familiar information
is consistent with the person’s attitude, the information was
apparently not su cient to create high levels of attitude certainty.
4. When attitudes are held with strong conviction, a di erent pattern

emerged. Con dence led to a bias toward attitude consistent
information when information was familiar rather than unfamiliar.
When one possesses con dence and awareness of the arguments
surrounding a topic, con dence can take the path of least
resistance. Known information that disagrees with a con dent
attitude may be unappealing; this information may have been
previously argued down and deemed invalid.
152
fi
fi
fi
ff
ff
fi
fi
fi
ffi
fi
fi
ff
fi
ff
fi
ff
fi
fi
5. Uncertainty has the ability to guide exposure to information about an
unfamiliar topic (nuclear power) and past research demonstrating
con dence as a predictor of preference for pro-attitudinal
information regarding a well worn topic (e.g. abortion).
6. Information familiarity may be a key determinant of information’s

ability to serve attitude bolstering goals.
7. However, these data do not address whether people who are

uncertain in their attitudes tend to view all information on the topic
as being less familiar. Alternatively, it could be that uncertainty
creates a preference for unfamiliar rather than familiar agreeable
information that people seek out when they have the opportunity to
do so. 
Future work could examine these issues more directly by measuring
perceived familiarity of each paragraph selected and not selected to
compare the level of familiarity of chosen and unchosen information.
A. When people were relatively uncertain in their attitudes, they selected

information that supported rather than opposed their existing attitudes,
especially when the available information was relatively unfamiliar. 
Yet when people relatively con dent, their attitudes predicted
information selection only when information was familiar. 
Current studies provide evidence of a robust e ect of uncertain
attitudes in guiding information choices, and suggest that this e ect is
most likely to occur when information is capable of bolstering the
doubtful attitude by virtue of being novel.
B. Uncertainty might create motives to bolster the attitude that cause the
attitude to guide information seeking.  
Attitudes that have been traditionally viewed as being structurally weak
or associated with beliefs that undermine the strength of the attitude
can nonetheless be associated with strong motives that direct attitude
related information seeking.
C. Strength related features of attitudes may exert motivational in uences

on cognition and behaviour that are conceptually distinct from their
structural e ects. Sometimes these structural and motivational forces
work together. 
153
fi
ff
fi
ff
fl
ff
(E.g. in addition to creating and supporting an integrated cognitive
structure, putting e ort into elaborating about an attitude object may
make people motivated to hold on to that attitude.).
D. Motivation to bolster the attitude could result in strong use of the

attitude in some settings (e.g. driving information choices).
E. Uncertainty may not be the only attitudinal property with the potential to
exert motivational in uences that counter the more traditional structural
e ects. 
Attitudinal ambivalence is another possibility.
CROSS CULTURAL COGNITION (1)
Incheol Choi, Minkyung Koo, Jong An Choi. Individual Di erences in Analytic

Versus Holistic Thinking. 2007.
I. ABSTRACT
A. Authors constructed the Analysis Holism Scale AHS to measure analytic vs.
holistic thinking tendency.
B. Study 1, used exploratory and con rmatory factor analysis, and a 24 item
scale was developed.
C. Study 2, convergent and discriminant validities were tested.
D. Studies 3 and 4, the known group di erence validity was examined by

comparing scores on the AHS of Americans and Koreans (Study 3) and of
Korean students of Oriental medicine and Korean students of non-Oriental
medicine majors (Study 4). 
Results here show that Koreans and Korean students of Oriental medicine
scored higher on the AHS than did Americans and Korean students of non-
Oriental medicine majors, respectively.
E. Studies 5 and 6 tested predictive validity by examining associations of the

AHS with performances on 2 cognitive tasks (categorisation and causal
reasoning). 
Those with high scores on the AHS displayed holistic pattern of
performances on each task more than did those with low scores.
F. ANALYTIC HOLISTIC THINKING
1. Widely accepted that East Asians hold a holistic assumption that every
element in the world is somehow interconnected, whereas Westerners
154
ff
ff
fl
fi
ff
ff
tend to view the universe as composed of independent objects (e.g.
Munro, 1985; Nakamura, 1964/1985; Needham, 1962).
2. Studies on various social and cognitive domains such as attention (Chua,

Boland, & Nisbett, 2005; Hedden et al., 2000; Ji, Peng, & Nisbett, 2000;
Masuda & Nisbett, 2001), attribution (Choi & Nisbett, 1998; Morris &
Peng, 1994), categorisation (Choi, Nisbett, & Smith, 1997; Ji & Nisbett,
2001; Norenzayan, Smith, Kim, & Nisbett, 2002), memory (Masuda &
Nisbett, 2001), logical reasoning (Norenzayan et al., 2002), and tolerance
of contradiction (Peng & Nisbett, 1999) to compared East Asians and
westerners.
G. ATTENTION: FIELD VS. PARTS
1. East Asians are holistic, and attention tends to be oriented toward the
relationship between objects and the eld to which those objects belong.
2. Westerners tend to focus attention more on an object itself rather than on

the eld to which it belongs (Hedden et al., 2000; Ji et al., 2000; Masuda
& Nisbett, 2001).
3. This apparent di erence in allocation of attention allows East Asians to

see the whole picture with more ease than they would see individual
parts, whereas the reverse is the case for Westerners.
4. Consequently, East Asians are more eld dependent than are Westerns
(Witkin, Dyk, Faterson, Goodenough, & Karp, 1974), in that East Asians
nd it more di cult to separate an object from the eld in which it is
embedded than do Westerners.
5. By the same logic, East Asians are generally better than Westerners in
detecting the relations among objects in a background eld (Ji et al.,
2000).  
Such a di erence even occurs in eye movement (Chua et al., 2005).
H. CAUSALITY: INTERACTIONISM VERSUS DISPOSITIONISM
1. In explaining causal relationships, East Asians assume the presence of

complex causalities and focus more on the relationships and interactions
between an actor and his or her surrounding situations than do their
Western counterparts, who primarily consider the internal dispositions of
an actor.
2. East Asians consider a greater amount of information than Westerners do

before making a nal attribution (Choi, Dalal, Kim-Preito, & Park, 2003). 
Also less likely than Westerners to make the fundamental attribution error
155
fi
fi
ff
ffi
ff
fi
fi
fi
fi
fi
(Choi & Nisbett, 1998; Choi, Nisbett, 7 Norenzayan, 1999; Lee, Hallahan,
& Herzog, 1996; Miller, 1984; Morris, Nisbett, & Peng, 1995; Morris &
Peng, 1994).
I. PERCEPTION OF CHANGE: CYCLIC VERSUS LINEAR
1. East Asians tend to view a phenomenon as non static and expect that a
state of constant change exists because of the complex pattern of
interactions among the elements.
2. Westerners perceive most objects as independent, thus the essence of

an object does not dramatically change over time, nor is it a ected by
other factors.
3. When predicting future events, East Asians tend to possess a cyclical

view that assumes constant uctuations, whereas Westerners maintain a
linear perspective that expects similar patterns of change or stability has
have been displayed in the past (Ji, Nisbett, & Su, 2001; Peng & Nisbett,
1999).
J. CONTRADICTION: NAIVE DIALETICISM VERSUS FORMAL LOGIC
1. When 2 contradictory opposites exist, East Asians tend to pursue a

compromised middle ground. Naive dialecticism. (Peng & Nisbett, 1999).
2. Westerners resolve contradictions by choosing one of the two opposite

propositions. (Peng, 1997; Peng & Nisbett, 1999).
II. STUDY 1: AHS DEVELOPMENT (ANALYSIS HOLISM SCALE)
A. METHOD 
4 domains as the essential constructs of analytic holistic thinking dimension 
-Locus of attention (parts vs. whole) 
-Causal theory (dispositional vs. interactional) 
-Perception of change (linear vs. cyclic) 
-Attitude toward contradictions (formal logic vs. naive dialecticism) 
Generated a list of 80 items that represented each of the 4 domains but
selected 40 as preliminary items after considering face validity, overlaps in
meaning and relevance to the literature. 
303 students.
1. Exploratory Factor Analysis 

Conducted as part of item selection process. 
Reliability was found to be 0.74, considered high in general (Stangor,
1998).
156
fl
ff
2. Con rmatory Factor Analysis 
2 factors of “locus of attention” and “causality” were integrated into
one, and thus a three factor model might be more suitable than a
four factor model. 
In the end, discovered that 4 factor model ts better with the original
model by Nisbett et al. 2001 than does a 3 factor model.
III. STUDY 2: CONVERGENT AND DISCRIMINANT VALIDITY OF AHS 

Testing convergent and discriminant validity of the AHS by examining its
relations with other scales.  
Examined correlations with the following three scales: the Attributional
Complexity Scale (Fletcher, Danilvics, Fernandez, Peterson, & Reeder, 1986), the
Sternberg Wagner Self Assessment Inventory on the Global Style (SWSAI;
Sternberg & Wager, 1991), and the Rahim Organisational Con ict Inventory II
(ROCI-II; Rahim, 1983). Each one is supposed to measure one’s causal
complexity, relative emphasis on the whole vs. the part, the attitude toward
con ict, respectively, all of which are key characteristics of holistic thinking.
1. ACS 
Scale consists of 28 items. Construct of attributional complexity includes
“preference for complex rather than simple explanations”, “awareness of
the extent to which people’s behaviour is a function of interaction with
others”, and “tendency to infer abstract or causally complex internal
attributions” (Fletcher et al., 1986).
2. SWSAI 
Global thinking style indicates the extent to which people focus on and
prefer dealing with the larger issues rather than the details. 8 items.
3. ROCI-II 
5 sub-dimensions, each representing a unique style of dealing with
con ict: integrating, avoiding, dominating, obliging, and compromising
styles.
4. INDCOL 
Indicates the tendency to place more value on personal goals over in
group ones, whereas reverse case is the case for collectivism (Triandis,
1995).
5. SCS 
Markus and Kitayama 1995, on the independent self versus
interdependent self, constructed to measure the degree to which ones
de nes oneself in relation with others or independent from social
contexts.
157
fi
fl
fl
fi
fi
fl
B. METHOD
1. Participants 
328 students, 131 male, 194 female, 3 unidenti ed gender
2. Procedure  
Each participant administered a packet of questionnaires including all the
above packets.
1. Convergent Validity
a) Signi cant positive correlations between the AHS and the 3

scales that were supposed to measure thinking styles (ACS,
SWSAI, ROCI-II). 
AHS (Perception of change, attitude toward contradiction, locus
of attention, and causality). 
SWSAI (Locus of attention, attitude toward contradiction, and
causality). 
ROCI-II (Attitude toward contradiction, and causality).
2. Discriminant Validity 
No signi cant correlations were found between the overall score of
the AHS and 2 scales measuring cultural di erences in value and self
construal.
IV. STUDY 3: THE AHS AND A BETWEEN CULTURE COMPARISON 

Selected Koreans and Americans as a representative of East Asian culture and
North American culture, respectively. 
Expected that Koreans would score higher than Americans on the AHS. 
Another purpose of Study 3 was to examine whether the 4 factor model based
on theoretical assumptions could be supported in other cultures beyond the
Korean culture.
A. METHOD 
104 Korean students at Seoul National University (50 male, 54 female) and
87 American students at University of Michigan. 
Ran through the packets.
B. RESULTS
1. Mean composite of AHS was signi cantly higher for Korean than
American students. 
Korean students preferred to resolve con icts by making a
158
fi
fi
fi
fl
ff
fi
compromise, expected more changes in future events, gave more
attention to the whole rather than to small details, and maintained
more complex causal beliefs than their American counterparts.
2. AHS is valid in the sense that the members of a holistic culture

(Koreans) scored higher than did the members of an analytic culture
(Americans) in the scale.
V. STUDY 4: THE AHS AND A WITHIN CULTURE COMPARISON 

If AHS is valid, the former group of students should score higher than the latter
on the AHS. (Oriental medicine in Korea vs. other students in Korea).
A. METHOD
1. Participants 
129 students.
2. Procedure 
Did packets.
B. RESULTS
1. Oriental medicine students exhibited signi cantly higher score on

the AHS than did students in others majors.
2. Oriental medicine students in comparison with psychology class

students reported that they paid more attention to the whole (vs.
parts) and possessed more complex causal beliefs
3. But 2 groups did not di er in attitude toward contradiction, nor

perceptions of change.
4. AHS has the ability to successfully discriminate a more holistic group

from a more analytic group in both between culture and within
culture contexts.
VI. STUDY 5: THE AHS AND SIMILARITY JUDGEMENT 

Conducted to address issue of if the AHS is valid, then we should expect that
those with higher scores on AHS to utilise family resemblance more and rules
less than those with lower scores on the AHS.
A. METHOD
1. Participants 
92 students from University.
159
ff
fi
2. Procedure 
Provided with total of 9 series of stimuli sets that were randomly selected
from the original sets that were randomly selected from the original set
that were randomly selected from the original sets used in Norenzayan et
al. 2002 (Study 2).
B. RESULTS
1. AHS predicted number of family resemblance based judgement such

that the more holistic people made the greater number of
judgements based on the family resemblance strategy (hence the
fewer rule based judgements).
2. European Americans clearly preferred the rule based strategy over

the family resemblance strategy, whereas Koreans displayed the
opposite pattern in Norenzayan et al. (2002).
3. Analytic Koreans in Study 5 are still more holistic than European

Americans in Norenzayan et al. (2002).
4. AHS is able to predict individual di erences in the habitual mode of

categorisation, such that those with higher scores on the AHS
preferred a family resemblance strategy over a rule based strategy in
categorisation more than did those with lower scores.
VII. STUDY 6 THE AHS AND CAUSAL COMPLEXITY 

East Asians are expected to consider a greater amount of information to explain
a certain event. 
Choi et al. (2003) conducted a series of studies to test this prediction. 
Provided American, Asian American, Korean participants with a short scenario
of murder incident, adapted from an actual newspaper headlines, in which a
graduate student killed his advisor. 
Participants were also provided with a list of 97 items of information that might
or might not be relevant to the explanation of the incident. 
-Graduate student’s history of mental disorders 
-Whether or not graduate student had a history of violence 
-The way the professor dressed 
-Whether the professor ever unfairly gave the graduate student a bad evaluation 
-Whether the graduate student and the professor had o ces on di erent oors 
Participants asked to eliminate the irrelevant information from the list. 
Measured participants’ level of individualism collectivism and their AHS scores
and tried to show that one’s causal complexity was better explained by the AHS
than by individualism collectivism.
160
ff
ffi
ff
fl
A. METHOD
1. Participants 
119 students in introductory psychology class.
2. Procedure 
Same procedure as Choi et al. (2003) to measure one’s causal
complexity, also administered AHS and INDCOL.
B. RESULTS
1. The more holistic one was, the fewer items of information on judged
to be irrelevant. 
Out of 97 items holistic participants excluded about 39 items,
whereas analytic participants excluded about 45 items. 
Whether one was individualistic or collectivistic did not predict how
many items of information were judged to be irrelevant.
2. Scores of AHS were associated with performance styles of similarity

judgement (Study 5) and causal complexity (Study 6).
3. Replicated the well established contrast between East Asians and

North Americans in given cognitive tasks.
VIII. GENERAL DISCUSSION
A. AHS possessed an adequate level of psychometric reliability and validity

(Studies 1 and 2).
B. AHS was able to di erentiate 2 ethnic groups (Americans vs. Koreans in

Study 3) and 2 subgroups within a culture (students of Oriental Medicine
vs. students of other majors in Korea in Study 4).
C. Those with high scores, compared to low scores, of the AHS were
shown to judge similarity based more on overall similarity than on rules
(Study 5) and to consider a greater amount of information in causal
reasoning (Study 6). 
AHS can be used as a measure of the analytic holistic thinking style.
D. Although analytic and holistic thinking could be conceptualised as 2

independent dimensions, the AHS treated them as if they were the ends
of a single dimension for the ease of development.
E. Reverse Coded ITEMS
161
ff
1. No good a priori reason to believe that those with a strong
acquiescence tendency would rely on overall similarity rather than
rule in categorisation (Study 5) and consider more information in
causal attribution than their counterparts (Study 6). 
In Study 5, the task asked participants to choose one of the 2
options (similarity based strategy vs. rule based strategy), which has
little to do with the acquiescence bias. 
Yet AHS was still able to predict participants’ choices in the task.
2. Second, if lack of enough reverse items created an artefact, then we

should expect that the factor of change (which has 5 reverse items
out of 6 total items) and the other 3 factors (causality, attention,
contradiction) would be grouped into 2 di erent factors. 
However, tness was found to be poor.
3. Third, comparison between Americans and Koreans in Study 3

showed the expected di erence even in the change. If the response
bias severe problem, Koreans would have endorsed the reversed
items (i.e. analytic items of the change factor more than Americans
resulting in circularity of change less than Americans. However that
was not the case.
4. Finally we shortened the AHS. (possible).
F. BENEFITS OF AHS
1. Although an individual di erence approach is not the best way to

study cultural in uences on behaviour, it can certainly provide useful
information about the interplay between culture and behaviour.
(Practical).
2. O ers insights into antecedents and consequences of analytic

holistic thinking.
3. AHS addresses some questions that priming research cannot. Which

individual characteristics are associated with the analytic or holistic
thinking cannot be handled in priming research.
4. AHS can provide a hint to the question by administering the AHS

after priming and examining which component of the analytic holistic
dimension became particularly salient.
162
ff
fi
fl
ff
ff
ff
CROSS CULTURAL COGNITION (2)
Etsuko Hoshino Browne, Adam S. Zanna, Steven J. Spencer and Mark P. Zanna,
Shinobu Kitayama, Sandra Lackenbauer. On the Cultural Guises of Cognitive
Dissonance: The Case of Easterners and Westerners. 2005.
I. ABSTRACT
A. Cognitive dissonance and e ects of self a rmation on dissonance arousal

were examined cross culturally. In Studies 1 and 2, European Canadians
justi ed their choices more when they made them for themselves, whereas
Asian Canadians (Study 1) or Japanese (Study 2) justi ed their choices more
when they made them for a friend.  
In Study 3, an interdependent self-a rmation reduced dissonance for Asian
Canadians but not for European Canadians. 
In Study 4, when Asian Canadians made choices for a friend, an independent
self-a rmation reduced dissonance for bicultural Asian Canadians but not
for monocultural Asian Canadians.  
Studies demonstrate that both Easterners and Westerners can experience
dissonance, but culture shapes the situations in which dissonance is aroused
and reduced.
B. Research in this article demonstrates that cross cultural research helps the
eld understand relatively basic social psychological processes better, in this
case, both cognitive dissonance and self-image maintenance processes.
C. CULTURALLY IDEAL SELF CONCEPTS, COGNITIVE DISSONANCE, AND

SELF AFFIRMATION
1. The possibility of having made a irrational choice or having made a

decision in uenced by others, could induce Westerners to feel that their
culturally ideal self-concepts are threatened. These threatened feelings
could lead them to sense that their self integrity is damaged and
consequently, to justify their choices as a means of reducing cognitive
dissonance. (E.g. Heine et al., 1999; Markus & Kitayama, 1991; Triandis,
1989, 1996).
2. East Asians on the other hand, hold interdependent self views, tending to
attach greater importance to smooth and harmonious interpersonal
163
fi
fi
ffi
fl
ff
ffi
ffi
fi
relationships with their in group members (E.g. Heine et al., 1999; Markus
& Kitayama, 1991; Triandis, 1989, 1996).
3. Interdependent East Asians are expected to justify their decisions when

they fear they might have made interpersonally inconsiderate decisions,
which can have various implications for their relationships with in group
members, and not when they have made suboptimal decisions for
themselves, which a ect only the decision makers alone.
4. Cross culturally variable self concepts have implications for self image
maintenance. (Steele, 1988; Spencer et al., 1993). 
Self A rmation Theory suggests that people are motivated to maintain an
image of self integrity. When a negative event threatens their beliefs that
they are morally adequate and adaptive, individuals try to restore their
positive self images by a rming some positive, valuable aspects of their
self concepts.
5. When people’s culturally valued sense of self is threatened and people

experience cognitive dissonance, an e ective self image maintenance
method (other than justifying the decision they have made) is to a rm the
self in a culturally adaptive and appropriate way (i.e. a rming the
independent self for Westerners and the interdependent self for East
Asians).
D. PAST RESEARCH ON COGNITIVE DISSONANCE FROM CROSS CULTURAL

PERSPECTIVES
1. Heine and Lehman (1997), demonstrated in cross cultural di erences in

dissonance reduction between North Americans and East Asians. 
Across Canadian and Japanese participants, Canadians showed the
usual justi cation of their choices of CDs as a means of reducing
dissonance. They did not justify their decisions when provided with an
opportunity to a rm themselves through positive feedback on a
personality test. 
Japanese counterparts by contrast, did not show a tendency to justify
their choices of CDs. They did not rationalise their decisions because
East Asians do not experience cognitive dissonance in the conventional
free choice paradigm. Suggested that cognitive dissonance was a
culturally constructed phenomenon speci c to North American culture.
2. Sakai (1981)’s earlier research with an induced compliance paradigm

demonstrated contrasting results with East Asian Samples. 
Asked high school students to make a counter attitudinal speech on
abolition of coeducation either publicly (i.e. their names, a liated classes,
164
ffi
fi
ffi
ff
ffi
ff
fi
ffi
ffi
ff
ffi
and grade were included in an audiotaped speech) or anonymously. He
found that those who made the speech publicly showed signi cant
attitude change (i.e. higher endorsement of abolition) than those who
made the speech anonymously. Sakai attributed this attitude change in
the public speech condition to dissonance reduction.
E. UNCONVENTIONAL FREE CHOICE PARADIGM
1. Although free choice paradigm is known to present threats to Westerners

who hold independent self concepts, authors needed to consider realistic
situation in which East Asians who hold interdependent self concepts
were likely to experience cognitive dissonance. 
For East Asians, knowing preferences and anticipating the desires of
close in group members are important aspects of interpersonal
relationships. Failing to meet these cultural standards might not only hurt
the others’ feelings and thereby harm close interpersonal relationships,
but also make the decision maker feel incompetent. This will threaten self
integrity and lead to dissonance arousal.
2. Chose food as object of testing. Experimenters created free choice

paradigm where participants ranked and rated a list of Chinese food
entrees based on either their own preferences (self condition) or their
close friends’ preferences (friend condition). Participants were asked to
choose free lunch gift certi cate either for themselves (the self condition)
or for their close friends (the friend condition).
II. STUDY 1: POSTDECISIONAL JUSTIFICATION AMONG EUROPEAN

CANADIANS AND ASIAN CANADIANS 
Examined how culturally ideal self concepts (i.e. independent self concepts for
European Canadians vs. interdependent self concepts for Asian Canadians)
interact with situations in which both cultural groups experience dissonance
arousal and rationalise their decisions as a means of dissonance reduction.
A. Method
1. Participants 
126 undergraduate students.
2. Procedure
a) Strength of Identi cation with Asian and Canadian Cultures. 

Asian Canadians were asked which ethnic group they most identify
with and how much they identify with that ethnic group on an 11 point
scale. The longer Asian Canadians stayed in Canada, the less they
identi ed with Asian culture.
165
fi
fi
fi
fi
b) Also asked potential participants to rate the extent to which they
identify with Canadian culture on an 11 point scale.
c) Materials and free choice paradigm 

Participants reported to lab individually and greeted by experimenter.  
Randomly assigned to self or friend condition. 
Provided a cover story stating that group of researchers were
investigating the psychology of decision making in real life situations. 
Friend Condition, participants were told that “past research has
shown that survey responses are more meaningful and more accurate
when respondents picture themselves making decisions for another
person.”. 
Self Condition, were not given those additional explanations and
questions. 
All participants given a list of Chinese entrees that contained 25
dishes and asked to choose 10 items on the basis of either their own
preferences (self condition) or close friends preferences (the friend
condition). 
Asked to rank the top 10 most preferred entrees. 
Passed a free coupon for the 5th and 6th choice respectively, and
asked to choose one for themselves or for their friends when the
restaurant opened the following month. 
More detail of the entrees was revealed (e.g. price). 
Asked to rerate the 10 items. 
Study had 2 x 2 (cultural group: European Canadian vs. Asian
Canadian) and (target of coupon choice: self vs friend) between
subjects factorial design.
B. RESULTS
1. European Canadians in the self condition tended to show a greater

spread of alternatives than Asian Canadians in the self condition.
2. Among Asian Canadians, those in the friend condition tended to

show a greater spread of alternatives than those in the self
condition.
3. European Canadians in the friend condition showed less spread of

alternatives compared with European Canadians in the self condition
and Asian Canadians in the friend condition, although not signi cant.
166
fi
4. The stronger the Asian cultural identi cation the more
interdependent tendency there would be in post decisional
justi cation.
5. In friend condition, the more strongly Asian Canadians identi ed with

Asian culture, the more they justi ed their choices for their friends;
the simple slope was signi cant. 
In contrast, in self condition, the more strongly Asian Canadians
identi ed with Asian culture, the less they showed post decisional
justi cation for the choices made for themselves, the slope was also
signi cant.
C. DISCUSSION
1. European Canadians showed signi cant post decisional justi cation

when they made choices for themselves.
2. European Canadians’ post decisional justi cation of their choices

made for their friends was less than that of choices made for
themselves and less than Asian Canadians’ justi cation of the
choices made for their friends.
3. Asian Canadians did not engage in post decisional justi cation when
they made choices for themselves.
4. Asian Canadians tended to engage in post decisional justi cation

when they made choices for their close friends.
5. Weakly identi ed Asian Canadians did not justify their choices for
their friends counteracts such concerns about the use of Chinese
food in this research. This dispels any worries about potential
priming e ect.
6. Possible di erences in types of friends or felt closeness toward

friends can be safely eliminated as alternative explanations for mean
di erences of the spread of alternatives in the friend condition
between the 2 cultural groups.
III. STUDY 2: REPLICATION OF STUDY 1 AMONG EUROPEAN CANADIANS AND

JAPANESE 
Conceptual Replication of Study 1. But using Japanese.
A. Method
167
ff
fi
fi
fi
fi
ff
ff
fi
fi
fi
fi
fi
fi
fi
fi
fi
fi
fi
1. Participants 
197 students
2. Procedure 
Similar, just changed to some Japanese materials.
B. RESULTS
1. European Canadians in the self condition tended to show greater

spread of alternatives than Japanese in the self condition.
2. Japanese in the friend condition tended to show a greater spread of

alternatives than Japanese in the self condition.
3. Pattern was similar to Study 1.
4. Means of the spread of alternatives of European Canadians in the

self condition and of Japanese in the friend condition were
signi cantly di erent from zero.
5. Means of spread of alternatives of European Canadians in friend

condition and of Japanese in the self condition were not signi cantly
di erent from zero.
C. META ANALYSES ACROSS STUDIES 1 AND 2 

Conducted as simple contrasts of the means were not statistically signi cant
through the 2 studies. Tested for interaction across 2 studies.
1. Both European Canadians and East Asians showed the most

justi cation of their choices when their culturally important self
concepts were threatened by making choices. European Canadians
justi ed their choices made for themselves, whereas East Asians
justi ed their choices made for their friends.
D. DISCUSSION
1. Extent to which European Canadians engaged in justi cations for

their choices they made for their friends was not di erent from zero.
2. But di cult to claim that European Canadians justi ed the choices

they made for their friends or that they did not justify the choices
they made for their friends.
3. Evidence exists that show North Americans do experience

dissonance interpersonally or vicariously when they observe their in
group members engage in attitude inconsistent behaviour (Norton et
168
ff
fi
fi
fi
fi
ffi
ff
fi
ff
fi
fi
fi
al., 2003). 
If European Canadians felt personal responsibility for choices they
made for their friends, they would likely to justify their choices for
their friends. 
If they held strong in group identi cation with their individualistic
Canadian culture, they would espouse more individualistic
orientations and thus would be unlikely to justify choices for their
friends.
IV. STUDY 3: POSTDECISIONAL JUSTIFICATION AND SELF AFFIRMATION

AMONG ASIAN CANADIANS 
Investigated e ect of self a rmation as a means of countering cognitive
dissonance in the friend condition among Asian born Asian Canadians who
strongly identi ed with Asian culture. Examined this group because they
engaged in post decisional justi cation because of a threat to their
interdependent cultural ideals in Study 1, and are therefore, a group that should
show clear e ects of an interdependent self a rmation.
A. Method
1. Newly Devised Self A rmation Manipulations 

Tailored to a rm culturally ideal interdependent self concepts. 
Participants to select the most personally important value from a list of
various values (business economics, social life relationships, religion
spirituality, etc) and explain why in written form, it is important to them.
Explain why these values uniquely describe who they are.
2. Participants 
Potential Asian Canadian participants surveyed during mass testing
session at University. 61 undergraduates.
3. Procedure 
Randomly assigned to 1 of 3 self a rmation conditions. No self
a rmation, independent self a rmation, or interdependent self
a rmation. 
Used only the friend condition. 
Coupons and entrees again. 
Presented the values for self a rmation manipulation as another
experimenter’s request, unrelated to current experiment. 
Participants in independent self a rmation condition were asked to
choose 1 value that was the most personally important to them from a list
of 6 di erent values and to write a paragraph about how the selected
value describes who they are. 
169
ffi
ffi
ff
ffi
ff
ff
fi
ffi
ffi
ffi
ffi
fi
ffi
ffi
fi
ffi
Participants in interdependent self a rmation condition were asked to
select the most important value for both themselves and their family
members and to explain in a paragraph why they share those particular
values with family members. 
Participants in the no self a rmation condition did not receive these
manipulations.
B. RESULTS
1. Mean of spread of alternatives of the no self a rmation and

independent self a rmation conditions was signi cantly di erent
from the mean of the spread of alternatives of the interdependent
self a rmation condition.
2. Di erence between the no self a rmation condition and

independent self a rmation condition was not signi cant.
3. Di erence between no self a rmation condition and interdependent

self a rmation condition was signi cant.
4. Di erence between independent self a rmation condition and

interdependent self a rmation condition was not signi cant.
5. Means of spread of alternatives in no self a rmation condition was

signi cantly di erent from zero, but those in the independent self
a rmation condition and in the interdependent self a rmation
condition were not.
C. DISCUSSION
1. Asian Canadians who did not have a chance to a rm themselves

engaged in post decisional justi cation when they made choices for
their close friends, replicating Study 1 and Study 2. (Asian Canadians
and Japanese respectively).
2. Interdependent self a rmation was an e ective means for Asian

Canadians to bu er or reduce their threatened feelings. 
Through an interdependent self a rmation opportunity, Asian
Canadian participants seemed to be able to maintain their cultural
ideals, even after making choices for their close friends, and to
signi cantly reduce both their levels of dissonance arousal and their
concomitant need to justify their choices.
170
ffi
ff
ff
ff
fi
fi
ffi
ffi
ff
ff
ffi
ffi
ffi
ffi
ffi
ffi
fi
ffi
ffi
ffi
fi
ffi
ff
ffi
ffi
ffi
fi
fi
ffi
fi
ff
3. By contrast, although Asian Canadians in the independent self
a rmation condition had an opportunity to a rm themselves,
a rming an independent self concept was not an e ective means of
neutralising the threat from the possibility of making non optimal
decisions for valued in group members. They tend to engage in post
decisional justi cation of the choices that they made for their close
friends.
4. However, the degree to which those in the independent self

a rmation condition engaged in the justi cation fell between mean
spread of alternatives obtained in the no self a rmation and
interdependent self a rmation conditions. 
If independent self a rmation opportunity was not helpful at all to
Asian Canadians, the degree to which they engaged in post
decisional justi cation should not have been equivalent to the no self
a rmation condition.  
One possible explanation for this more pronounced e ect of the
interdependent self a rmation on post decisional justi cation could
be that interdependent self a rmation is somehow stronger or has
better e cacy than the independent self a rmation, apart from the
fact that it a rms Asian Canadians’ culturally important
interdependent self concepts people possess, the interdependent
self a rmation should always have a stronger e ect in reducing
psychological discomfort or bu ering people from self threats.
5. Small study conducted, 37 European Canadian men to make choices

for themselves after receiving either an independent or an
interdependent self a rmation. 
Results demonstrated that an independent self a rmation
eliminated spread of alternatives whereas interdependent self
a rmation did not.
V. STUDY 4: POST DECISIONAL JUSTIFICATION AND INDEPENDENT SELF

AFFIRMATION AMONG BICULTURAL AND MONOCULTURAL ASIAN
CANADIANS 
Excluded interdependent self a rmation condition and examined the e ects of
the independent self a rmation only among Asian Canadians who either
identi ed only with their Asian culture (monoculturals) or identi ed with both
Canadian and Asian cultures (biculturals).
A. Method
171
ffi
ffi
ffi
ffi
ffi
fi
ffi
ffi
ffi
fi
fi
ffi
ffi
ffi
ffi
ffi
ffi
ffi
ff
fi
ffi
ffi
ffi
ff
ffi
ff
ff
fi
fi
ff
1. Participants 
120 undergraduate students. 
*The longer Asian Canadians stayed in Canada, the more strongly they
identi ed with Canadian culture. This is how the experimenters
determined who is mono and who is bicultural.
2. Procedure 
Randomly assigned to either no self a rmation or independent self
a rmation condition. 
Similar procedure with the food coupons, rank entrees, and asked to
choose one value that was most personally important to them from 6
di erent values and write short paragraph on why.  
Participants in no self a rmation condition did not receive the value
manipulation.
B. RESULTS
1. Signi cantly marginal e ect for interaction between self a rmation

condition and identi cation with Canadian culture.
2. The more participants identi ed with Canadian culture in the

independent self a rmation condition, the less post decisional
justi cation they engaged in.
3. Identi cation with Canadian culture was unrelated to post decisional

justi cation in the no self a rmation condition.
4. Among bicultural participants, those in independent self a rmation

conditions were predicted to show less post decisional justi cation
than those in the no self a rmation condition.
C. DISCUSSION
1. Asian Canadians who did not have an opportunity to a rm

themselves engaged in post decisional justi cation when they made
choices for their close friends. The strength of their identi cation
with Canadian culture did not matter in this situation. However, the
strength of identi cation with Canadian culture did matter when
Asian Canadians could a rm their independent styles.
2. Bicultural Asian Canadians who strongly identi ed with both Asian

and Canadian cultures were able to use the independent self
a rmation opportunity e ectively to bu er or reduce threatened
feelings and maintain their cultural ideals even after making choices
172
ffi
ffi
ff
fi
fi
fi
fi
fi
fi
ffi
fi
ffi
ff
ffi
ff
ffi
ffi
fi
ffi
ff
fi
fi
ffi
fi
ffi
ffi
fi
for their close friends. They did not need to justify their choices to
reduce dissonance arousal.
3. Independent self a rmation opportunity was not helpful to

Monocultural Asian Canadians, who did not strongly identify with
individualistic Canadian culture and thus did not embrace
independent self concepts. They continued to engage in post
decisional justi cation of the choices that they made for their close
friends.
4. Versatility that Bicultural Asian Canadians demonstrated in the use

of self a rmation suggests that they can switch or alternate the
psychological frames of reference that originated in 2 di erent
cultures to act appropriately according to the social environment in
which they are situated.  
They seem to be able to cross with ease the boundary between an
interdependent self, which is a more indigenous Asian self concept,
and an independent self, which is a more predominant North
American self concept.
VI. GENERAL DISCUSSION
A. People regardless of cultural backgrounds experienced cognitive

dissonance after making choices that were important to them and
subsequently engaged in e orts to justify their decisions.  
However, the situation in which they experienced cognitive dissonance
and justi ed their decisions were di erent across di erent cultural
groups.
B. Like Westerners, East Asians do engage in rationalisation of their

decisions. East Asians consistently justi ed their choices across the 4
studies conducted, when those choices pertained to their culturally
important self concepts.
C. East Asians engaged in dissonance reduction through rationalisation of

their decisions when their culturally important self concepts were
threatened by the need to make choices for their in group members.
D. Asian Canadians and Japanese justify their choices for their close
friends. Interpersonal concerns lead East Asians to experience
cognitive dissonance and subsequently engage in dissonance
reduction.
173
ffi
fi
fi
ffi
ff
ff
fi
ff
ff
E. Asian Canadians and Japanese who chose gift certi cate for
themselves did not justify their choices because they did not have any
interpersonal concerns. (Kitayama et al. 2004 study).
F. European Canadians seemed to experience some dissonance arousal

when they made their choices for their close friends, but the dissonance
they experienced was not to the same degree as when they made their
choices for themselves, nor as much as that experienced by East Asians
who made choices for their friends.
VII. ALTERNATIVE EXPLANATIONS FOR EAST ASIANS’ JUSTIFICATION IN FRIEND

CONDITION
A. Participants used their own food preferences pretending they were their
friends. 
But possibility was pre-empted through experimental design.
B. Friend condition is due to vicarious dissonance. 

Possibility dispelled through experimental design and not witnessing
their friends making their decisions.
C. Unintended social projection. 

Dismissed through procedures and precautions.
VIII. EFFICACY OF SELF AFFIRMATION AMONGST ASIAN CANADIANS (3 and 4)
A. Self A rmation as an alternative to post decision justi cation, can be

an e ective means for Asian Canadians to counter threats to their self
concepts.
B. Providing Asian Canadians with opportunities to a rm themselves

reduced their need to engage in dissonance reduction as a strategy to
restore threatened self integrity.
C. This Self A rmation nding depends substantially on both the locus

and strength of Asian Canadians’ cultural identi cation. Asian
Canadians who strongly identi ed with Asian culture with an
opportunity to a rm their interdependent self concepts signi cantly
reduced their need to counter threatened self integrity by justifying the
choices they made for close friends.
D. Bicultural Asian Canadians who identi ed strongly with both Asian and
Canadian cultures with an independent self a rmation opportunity
reduced their need to counter threatened self integrity by engaging in
174
ff
ffi
ffi
ffi
fi
fi
fi
ffi
fi
ffi
fi
fi
fi
post decisional justi cation. 
In contrast, providing Monocultural Asian Canadians, who only
identi ed strongly with Asian culture, with the same opportunity did not
reduce their tendency to justify their choices. 
The independent as well as the interdependent self a rmation
manipulation served as an e ective means for countering threatened
self integrity.
E. East Asians justify their choices for their friends not under the pretence
of being considerate to their friends but because they experience
interpersonal dissonance by making choices for their friends, and they
try to resolve the psychological discomfort they experience by
rationalising their choices.
IX. ASIAN CANADIANS’ BICULTURALISM AND FLUIDITY OF SELF AFFIRMATION
A. Strength of East Asians’ identi cation with Asian culture is of crucial

importance. Although Asian Canadian participants in the research were
all born in East Asian countries, the degree to which they identi ed with
Asian culture varied greatly. 
This variability in the strength of Asian cultural identi cation was
important in their experience of cognitive dissonance and their
subsequent justi cation of their choices.
B. East Asians who only weakly identi ed with culture are ethnically Asian,
but their cultural ideals seem to be similar to independent self concepts
of Westerners did.
C. East Asians who strongly identi ed with Asian culture held

interdependent self concepts as their cultural ideals. Experienced
cognitive dissonance when their interdependent self concepts were
threatened, and they engaged in dissonance reduction through either
justi cation of their choices for their close others or interdependent self
a rmation.
D. Bicultural East Asians who strongly identify with both Asian and
Canadian cultures seem to embrace both interdependent and
independent self concepts as their cultural ideals. 
They are capable of using independent self a rmation as a means of
reducing dissonance, when their interdependent self is threatened.
Smoothly cross the boundary of 2 cultures and readily switch between
the 2 cultural mind sets. (Hong et al., 2000).
175
ffi
fi
fi
fi
fi
ff
fi
fi
fi
ffi
fi
ffi
fi
E. Provide support for the uidity of self a rmation processes (Steele,
1988; Steele et al., 1993; Spencer et al., 1993). Individuals have a pool of
positive self concepts within a large self system, and they can a rm
themselves and maintain overall self integrity by using some of these
positive attributes that are not necessarily under threat.
1. Both Monocultural and Bicultural East Asians who experience threat

to a particular interdependent self concepts (e.g. friendships) can
use another interdependent self concept (i.e. family relationships) to
a rm themselves. When they experience a threat to their
relationships with their close friends by the possibility of making an
inconsiderate choice or the wrong choice of Chinese food, they can
a rm themselves by using an important value that they share with
their family.
2. Bicultural East Asians, whose interdependent self concepts are

threatened by interpersonal concerns that arise from the possibility
of making non optimal choices for their close others, can a rm
themselves using either their independent or their interdependent
self concepts. In fact, they seem to have more resources with which
to self a rm than Monocultural East Asians who can a rm
themselves only using interdependent self concepts.
FINAL TUTORIAL ARTICLES SUMMARY

STEREOTYPES UNDER COGNITIVE LOAD
Jer rey W. Sherman, Frederica R. Conrey, & Carla J. Groom. Northwestern

University. Encoding Flexibility Revisited: Evidence for Enhanced Encoding of
Stereotype Inconsistent Information Under Cognitive Load. 2004.
I. ABSTRACT
A. Experiment tested 2 key components of the Encoding Flexibility Model of

stereotyping. Results demonstrated that a cognitive load increased the
attention paid to stereotype inconsistent information, and decreased the
attention paid to stereotype consistent information. Cognitive load also
enhanced the perceptual encoding of inconsistent information while
diminishing the perceptual encoding of consistent information. Implications
of these results for the role of e ciency and the interaction of motivation and
ability in social cognition are discussed.
176
ffi
ffi
ff
ffi
fl
ffi
ffi
ffi
ffi
ffi
B. The idea that stereotypes perpetuate themselves through con rmatory
biases has attained the status of an axiom of social psychology. This
principle is a rmed in every major social psychology textbook, and is a
central organisational theme in prominent reviews of the stereotyping
literature (e.g. Fiske, 1998; Hamilton & Sherman, 1994; Hilton & von Hippel,
1996). 
Common corollary to this theme is the notion that these con rmatory biases
ought to be particularly evident when perceivers su er from diminished
processing capacity. 
 
One such bias concerns processing of stereotype consistent and
inconsistent information. Suggested that particularly when processing
resources are depleted, consistent information is more thoroughly attended
to and encoded than inconsistent information.
1. Firstly, this is because it ts with an existing expectancy, stereotype

consistent information is simply easier to comprehend than stereotype
inconsistent information and is therefore more likely to be successfully
encoded into memory, particularly when resources are low and encoding
is di cult. In this way, stereotypes act as passive lters that permit the
processing of consistent information and block the processing of
inconsistent information under low capacity conditions (Bodenhausen &
Lichenstein, 1987; Macrae, Hewstone, & Gri ths, 1993; Stangor & Duan,
1991; Stangor & McMillan, 1992).
2. Second, motivational concerns have been proposed to direct resources

away from inconsistent and toward consistent information when
resources are low. Principles of selective exposure suggest that people
prefer to not attend to information that challenges their beliefs if they do
not have resources to counter argue that information (Frey, 1986). 
In addition, the cognitive miser views of social cognition suggest that
people are further motivated by the desire to exert as little e ort as
necessary in forming social impressions (Sherman, Lee, Besseno , &
Frost, 1998; Sherman, Macrae, & Bodenhausen, 2000). 
These motives collectively have been suggested to lead perceivers to
actively avoid stereotype inconsistent information (which challenges
beliefs and is di cult to comprehend) when capacity is depleted, and to
shift attention to consistent information (which con rms beliefs and is
relatively easily understood). In this way, they act as active lters that
promote processing of consistent information and diminish processing of
inconsistent information under low capacity conditions. (Bodenhausen,
1988).
177
ffi
ffi
ffi
fi
ffi
fi
fi
ff
fi
ff
fi
fi
ff
C. THE ENCODING FLEXIBILITY MODEL
1. Many self perpetuating biases in processing of stereotype relevant

information are present. A number of these biases are more prevalent
when processing capacity is depleted (Sherman et al., 1998, 2000).
2. There may be important exceptions to these generalities. They cast

signi cant doubt on the motivated “active” lter model and suggest that
the comprehension based “passive” lter model fails to provide a
complete account of the processing of consistent and inconsistent
information.
a) Evidence that attention in fact, shifts toward inconsistent information

and away from consistent information when resources are depleted.
(Sherman, 2001; Sherman & Frost, 2000; Sherman et al., 1998)
b) Sherman proposed the Encoding Flexibility Model EFM of

stereotyping (Sherman et al., 1998). 
According to this model, when resources are low, concerns for
e ciency more so than concerns for defence or sloth drive the use of
stereotypes.  
E ciency is de ned as the ratio of product gained (social information
in this case) to energy expended. It argues that processing is not
wholly biased toward either consistent or inconsistent information
when capacity is low. Rather, di erent aspects of consistent and
inconsistent information are encoded in order to maximise the amount
of information gained for the e ort expended.
3. Consistent information ts with prior expectancy and is easily

understood, it enjoys an advantage in conceptual encoding. Perceivers
are more able to extract the abstract, gist meaning of consistent than
inconsistent information, particularly when resources are depleted.
a) Because consistent information is relatively easy to comprehend and

con rms prior knowledge, it receives relatively little attention, and the
details of the information are not carefully encoded (Johnston &
Hawley, 1994; von Hippel, Jonides, Hilton, & Narayan, 1993).
b) Instead, having extracted the basic gist, attention can be redirected

toward other items of information that are not so easily understood,
including inconsistent information. Thus, as processing capacity is
depleted, consistent information enjoys relatively greater conceptual
encoding than inconsistent information.
178
ffi
ffi
fi
fi
fi
fi
ff
ff
fi
fi
c) At the same time, attention and perceptual/contextual encoding
(encoding of physical details and contextual speci cs of the stimulus)
shift away from consistent and toward inconsistent information.
Together, these exible encoding processes maximise the joint
encoding of consistent and inconsistent information. 
Conceptual advantage for consistent acts ensures that their essential
abstract meaning is extracted for possible use, whereas the
attentional and perceptual/contextual shifts toward inconsistent acts
aid in the comprehension of these behaviours and help to ensure that,
even if their abstract meaning cannot be extracted, at the least they
will remain available for later inspection, consolidation, and potential
use.
4. Abundant evidence exists for conceptual encoding advantage for

consistent information when capacity is depleted. 
Target trait judgements are more stereotypical when target information is
learned under low capacity conditions (Bodenhausen, 1990;
Bodenhausen & Lichenstein, 1987; Gilbert & Hixon, 1991; Kruglanski &
Freund, 1983; Macrae et al., 1993; Pratto & Bargh, 1991). 
Other research has shown that free recall favours consistent over
inconsistent information when the content is learned with diminished
capacity (Bodenhausen & Lichenstein, 1987; Macrae et al., 1993;
Sherman & Frost, 2000; Stangor & Duan, 1991).
a) Stereotypic traits were more likely to be primed by relevant behaviours

than counter stereotypic traits under low capacity encoding
conditions. (Sherman et al., 1998).
5. Evidence of shifts in encoding away from consistent and toward

inconsistent information under low capacity conditions is far less
prevalent.
a) Participants devoted relatively greater attention to inconsistent than

consistent information when capacity was constrained compared to
when it was full (Sherman et al., 1998, Experiments 1 & 2).
6. Consistent and inconsistent behaviours can occur simultaneously.

(Sherman et al., 1998, Experiment 3). 
Strong evidence for attention hypothesis of the EFM remains elusive.
7. No evidence available to support the perceptual/contextual encoding

hypothesis of the EFM.
D. OVERVIEW & PREDICTIONS
179
fl
fi
1. Purpose of experiment is to seek direct support for both attentional and
perceptual encoding hypotheses of the EFM. 
Participants asked to form impression of a target that belonged to a
stereotyped group. 
Half were placed under a cognitive load as they formed their impressions. 
Information about the target were presented in pairs. 
Following presentation of each pair, an x appeared on either left or right
side of screen 
Participants’ task was to press either a key marked left or a key marked
right as quickly as possible upon appearance of the x to indicate on
which side of the screen the x had appeared. 
Among pairs of behaviours were four that included one stereotype
consistent and one stereotype inconsistent behaviour. 
For these pairs, the x appeared on the side of the screen corresponding
to the side on which either the consistent or inconsistent item had
appeared. Interest centred on participants’ latencies to respond to the x
as a function of processing capacity and whether the x appeared in the
same position as the consistent or inconsistent behaviour. 
These latencies measure the extent to which participants were attending
to the consistent and inconsistent items of a pair when the x appeared. 
The more attention being paid to a particular item, the less time it should
take to respond to an x probe that appears in the same position as that
item.
2. When processing capacity is depleted, participants’ ability to process

both items is signi cantly constrained.
3. Following initial task, some participants were asked to complete a

graphemic cued recognition measure. Participants were presented with a
list of words on a sheet of paper. Some of the words on the sheet were
similar graphemically to words that had appeared in the behavioural
descriptions during the impression formation task (e.g. shell) was
presented as a cue for the word (shelf), which had appeared in one of the
behavioural descriptions. 
Participants were asked to circle only the items that looked like words
that had been presented in the the impression formation task.
4. Measure of the perceptual rather than conceptual encoding because

participants rely on physical attributes of the cues rather than their
semantic content in order to perform that task (Blaxton, 1989; Mulligan,
1998; Roediger, 1990). 
Conceptual measures are sensitive, perceptual measures are not.
180
fi
5. Graphemic cued recognition measure is also an explicit measure of
memory because participants are directed to think back to the original
information and try to remember it. (Mulligan, 1998; Roediger, 1990).
II. EXPERIMENT
A. METHOD
1. Participants 
196 University students

Participants asked to engage in an experiment on impression formation.
They were told that they would be reading information that had been
drawn from a magazine article about a person named Bob Hamilton, and
that we were interested in how people form impressions based on
simultaneous exposure to multiple pieces of information. 
Bob was described as either a skinhead or a priest who lived in Chicago. 
Descriptions of Bob consisted of 24 behaviours, 8 of which were
pretested to be kind, 8 of which were pretested to be unkind, and 8 of
which were pretested to be irrelevant to the kind unkind dimension. 
Participants in the skinhead condition, the unkind behaviours were
stereotype consistent and the kind behaviour were stereotype
inconsistent. Reverse was true for participants in the priest condition. 
Thus, same behaviours (and graphemic cue words) served as both
stereotype consistent and inconsistent stimuli, depending on the target. 
Behavioural stimuli were presented on participants’ computer screens in
pairs, with one behaviour on left and one on right. 
OF 12 pairs of items, 4 contained consistent and inconsistent behaviours,
4 contained consistent and irrelevant behaviours, and 4 contained
inconsistent and irrelevant behaviours. 
Participants completing the x probe task were instructed to press keys
marked left or right as quickly as possible to indicate on which side of the
screen the x had appeared. Response times were recorded. 
Participants not performing the x probe task simply read the item pairs
and formed an impression of the target. 
As impressions were formed, half of the participants were placed in a low
processing capacity condition. Further informed that experiment was
concerned with people’s ability to perform multiple tasks at the same
time.  
Cognitive load manipulated by asking participants to hold an 8 digit
number in memory as they performed the impression formation and x
probe tasks. 
181
At the end, participants were given a 5 minute ller task in which they
were asked to solve a series of puzzles. 
Also given a sheet of paper with 50 words, and asked to circle word that
appeared during the impression formation task in the behaviours. (Only 4
existed).
B. RESULTS
1. Predicted interaction between processing capacity and item type.
2. In high processing capacity condition, participants required more

time to respond to x probes following inconsistent than consistent
items, suggesting that greater attention was paid to the consistent
items of the pairs.
3. In low processing capacity condition, participants required more

time to respond to x probes following consistent rather than
inconsistent items.
4. Responses to x probes following inconsistent items were marginally

faster in the low capacity than the high capacity condition.
5. Responses to x probes following consistent items were slower in the

low than the high capacity condition (not reliably so).
6. Compared to counterparts with full processing capacity, participants

with limited capacity devoted relatively greater attention to the
inconsistent than the consistent item in the pair.
C. GRAPHEMIC CUED RECOGNITION TASK 

To test the hypothesis that diminished processing capacity enhances the
perceptual encoding of inconsistent relative to consistent behaviours, the
proportions of correct probes circled corresponding to words in consistent
and inconsistent behaviours were analysed.
1. In high processing capacity condition, participants recognised a

greater proportion of consistent than inconsistent probe words
(di erence was not really reliable).
2. In low processing capacity condition, participants recognised a

greater proportion of inconsistent than consistent items. (di erence
was marginally reliable).
3. Inconsistent probes were recognised with signi cantly greater

accuracy in the low than the high processing capacity condition.
182
ff
fi
fi
ff
4. Consistent probes were more poorly recognised in the low than the
high processing capacity condition (not reliably so).
5. Participants with limited capacity encoded relatively greater

perceptual detail of the inconsistent than the consistent item in the
pairs.
III. DISCUSSION
A. Results provide strong support for the Encoding Flexibility Model.
B. When there is direct competition between encoding of stereotype

consistent and inconsistent behaviour, the imposition of a cognitive load
increased the relative attention devoted to the inconsistent behaviour.
C. Results from graphemic cued recognition measure provide the rst

support for the perceptual / contextual encoding hypothesis of the EFM.
D. Imposition of a cognitive load enhanced the encoding of the perceptual

details of inconsistent compared to consistent behaviours. 
Does not mean to take this data to imply that memory will typically
favour unexpected events that have been encoded under cognitive load. 
Contrary, many studies have shown that free recall favours consistent
information when the stimuli have been learned under low capacity
conditions. (Bodenhausen & Lichenstein, 1987; Macrae et al., 1993;
Sherman & Frost, 2000; Stangor & Duan, 1991). 
Distinction between ndings is that free recall is the prototypic measure
of conceptual processing (which is enhanced for expected information
under cognitive load), whereas graphemic cued recognition is a
measure of perceptual memory (which is enhanced for unexpected
information under cognitive load).
E. Support for Encoding Flexibility Model EFM does not require

participants attend more carefully to the inconsistent than consistent
item when resources are depleted.  
What is required is a reliable shift of attention away from consistent and
toward inconsistent information in these conditions compared to high
capacity conditions. 
Support for the Encoding Flexibility Model does not require greater
perceptual encoding of inconsistent than consistent items when
resources are depleted. 
What is required is a relative increase in perceptual encoding of
183
fi
fi
inconsistent relative to consistent information in these conditions. 
Model predicts all processing will shift away from inconsistent and
toward consistent information when capacity is low.
F. ON EFFICIENCY
1. Encoding Flexibility Model proposes that perceivers allocate their

resources more strategically when resources are low to maximise
the amount of information gained for the e ort expended. Whereas
consistent behaviours enjoy a conceptual encoding advantage,
attention and perceptual encoding shift toward inconsistent
behaviours.
2. Beyond maximising the overall extraction of information, these

encoding tradeo s simultaneously promote stability and plasticity in
stereotypic expectancies. 
Conceptual advantage for consistent information reinforces the
expectancy, whereas shifts in attention and perceptual processing
help to maintain vigilance that the expectancy may be in need of
revision. 
Dual pursuits of stability and plasticity in expectancies have been
argued to be central components of an adaptive cognitive system
(Johnston & Hawley, 1994; McClelland, McNaughton, & O’Reilly, 1995;
Sherry & Schacter, 1987; Tulving, Markowitsch, Kapur, Habib, & Houle,
1994).
3. Said models proposed that increases in stereotype use are

associated with decreases in individuation, particularly in the
encoding of stereotype inconsistent information (Brewer, 1988; Fiske
& Neuberg, 1990). 
However, current and previous results show that decreases in
processing capacity may simultaneously increase both stereotyping
(through conceptual encoding advantages for consistent
information) and individuation (through attentional and perceptual
encoding shifts toward inconsistent information). 
Data argues that stereotype use and individuation should be
conceived as 2 separate but related continua, rather than as mutually
exclusive processing models. (Nelson, Acker, & Manis, 1996). 
Stereotyping may be increased via conceptual processing while
individuation is simultaneously increased via attentional and
perceptual processing.
184
ff
ff
G. ON MOTIVATION
1. The selective exposure component of the active lter model shows

that processing goals are dependent upon su cient capacity.
Con rmatory processing is assumed to be the default processing
style when resources are depleted. Hence, the suggestion is that
perceivers may pursue the goal of encoding all relevant (including
inconsistent) information only if they have enough resources to
permit it (Bodenhausen, Macrae, & Sherm, 1999; Fiske & Neuberg, 1990;
Pendry & Macrae, 1994).
2. When resources are scarce, even perceivers who may have been
initially motivated to fully individuate a target are expected to
abandon that motive and revert to con rmatory processing because
it is easier and or because it allows them to leave desired
conclusions unchallenged. 
In contrast, Encoding Flexibility Model proposes that processing
ability and motivation are independent, although related.  
Although diminished capacity may in uence the manner in which a
particular goal is pursued, it does not change the basic motive of the
perceiver. Thus, an accuracy motivated perceiver will not abandon
the goal to process all relevant information when resources are low,
but may well have to pursue the goal in a more e cient manner.
3. Does not mean that attention and perceptual processing will always
shift toward inconsistent information when resources are depleted.  
Perceivers are not always motivated by accuracy.
4. Stereotypes are exible cognitive tools that may be adapted to the

current needs of the perceivers, whatever they may be. As e cient
tools, they are particularly likely to be recruited for these purposes
when capacity is diminished.
185
fi
fl
fl
fi
ffi
ffi
fi
ffi
IMPLICIT AND EXPLICIT ATTITUDE GENERALISATION
Kate A. Ranganath & Brian A. Nosek. University of Virginia. Implicit Attitude

Generalisation Occurs Immediately; Explicit Attitude Generalisation Takes
Time. 2008.
I. ABSTRACT
A. People are able to explicitly resist using knowledge about one person to
evaluate another person from the same group. After learning about positive
and negative behaviours performed by one individual from each of 2 di erent
groups, participants were introduced brie y to new individuals from the
groups. Implicit evaluations of the original individuals readily generalised to
the new individuals; explicitly, participants resisted such generalisation. Days
later, both implicit and explicit evaluations of the original individuals
generalised to the new individuals. Results suggest that associative links
(e.g. shared group membership) are su cient for implicit attitude
generalisation, but deliberative logic (e.g. individual group members are not
necessarily the same) can reduce explicit generalisation by association.
When knowledge distinguishing who did what is unavailable, such as after
forgetting, associative knowledge provides the basis of explicit evaluation.  
Simple association linking one individual to another can produce implicit
attitude generalisation immediately and explicit attitude generalisation
eventually.
B. Subtle In uence of Associative Relations
1. Gawronski and Bodenhausen (2006) suggested that most important

distinctions between associative and propositional reasoning is that
associative relations can be activated regardless of whether a person
believes those associations are accurate or inaccurate, whereas
propositional processes allow for the deliberation involved in conscious
declarations of truth or falsity.
2. Recent work showing the distinctiveness of implicit (associative) and

explicit (deliberative) evaluations provide evidence that people can
possess multiple evaluations of a single target, even evaluations they
would reject as false if given the opportunity to deliberate (Greenwald &
Banaji, 1995; Nosek, 2007).
C. Resources Preventing Generalisation Fade with Time
1. Clear memory for who did what is necessary for observers to keep their
evaluations straight and enables them to prevent the actions of one
individual from in uencing their judgements of another.
186
fl
fl
ffi
fl
ff
2. Verbatim memory is forgotten and becomes inaccessible quickly, leaving
people to rely on gist memory (Clark & Clark, 1977).
3. Quicker loss of verbatim memory relative to gist memory suggests that

over time, speci c information used to deliberately distinguish group
members will fade. If associations between group members still exist in
memory without the conscious individuating information, then passage of
time should increase the di culty of avoiding associative generalised
from one person to another.
II. EXPERIMENT  
Study took place in 2 sessions.  
First session, participants read about behaviours performed by 2 individuals.
Reemolap and Vabbenif (Gregg, 2000), each of whom was described as
belonging to a large and diverse social group (low in entitativity). One of the
individuals performed primarily positive behaviours, and the other performed
primarily negative behaviours. 
Participants then read brief introductions to 2 new individuals, Bosaalap and
Ibbonif, who belonged to the same social groups as original individuals. 
Information about new individuals were minimal, and pretesting showed it to be
relatively neutral.  
IAT and Semantic di erential ratings were used to measure participants’ implicit
and explicit attitudes, toward either the original individuals or the new
individuals. 
Three days after rst study session, participants were invited back to complete
the implicit and explicit attitude measures again. Hypothesised that days after
the initial learning episode, participants would show not only the implicit attitude
generalisation predicted for the rst session, but also full explicit attitude
generalisation. After the delay, we expected attitudes induced toward original
individuals to provide the basis for explicitly evaluating the new individuals.
A. METHOD
1. Participants 
Volunteers at Project Implicit Research. 
831 people, 82% completion rate.
2. Measures 
-IAT  
(Assesses associations among 2 concept categories e.g. Reemolap and
Vabbenif) and 2 evaluative attributes (e.g. good and bad, the attributes
used in this study) by requiring that participants categorise stimulus items
representing the 4 categories as quickly as possible using 2 keys of a
computer keyboard. Used e and i keys to identify. 
187
fi
fi
ff
ffi
fi
IAT consisted of 7 blocks of trials. 
-Self Report Measures 
(Explicit attitude measured). 
-Manipulation Check 
(Checked to see if participants could identify the 4 individuals to whom
they were introduced by matching names and pictures.).
3. Procedure
a) First Session 
Attitude induction procedure, designed to create clear preference for
one individual over the other. 
Participants saw a series of sentences describing behaviours
performed by Reemolap and Vabbenif. 
Each sentence accompanied by picture of the individual,
counterbalanced across participants, who performed that behaviour. 
Reemolap performed predominantly positive behaviours and Vabbenif
performed predominantly negative behaviours. 
Participants were next introduced to new people. Bossalap and
Ibbonif were described as belonging to the same groups as Reemolap
and Vabbenif, Laapians and Ni ans respectively.  
Both groups were described as diverse groups with many members
who perform all di erent types of behaviours. Information
distinguishing the new individuals consisted of a cartoon picture
showing both and these descriptions 
(Ibbonif is a sculptor, likes gardening, biking, and card games, kind
and thoughtful, but slightly greedy) 
(Bosaalap is a painter, enjoys cooking, hiking, music, warm and
considerate, but slightly dishonest). 
These descriptions were equivalent. 
Finally, participants randomly assigned to complete the implicit and
explicit attitude measures about either the original people (Reemolap
and Vabbenif), or the new people (Bossalap and Ibbonif).
b) Second Session 
Participants were randomly assigned to complete implicit and explicit
attitude measures toward either the original people or new people.
B. RESULTS
1. Attitudes were Induced.
188
ff
ffi
a) Individual who performed primarily positive behaviours was
explicitly preferred over the individual who performed primarily
negative behaviours.
b) Individual who performed primarily positive behaviours was

implicitly preferred over the individual who performed primarily
negative behaviours.’
2. Explicitly, generalisation was resisted; implicitly, evaluations

generalised to new individuals
a) Participants resisted using information about one individual in

their immediate explicit evaluations of another individual from the
same group.
b) Explicit attitude generalisation was not entirely avoided. 

Attitudes toward the new individuals did show a preference
consistent with what was learned about their fellow group
members; participants preferred the new individual who belonged
to the same group as the original person who performed primarily
positive behaviours.
c) Participants explicitly resisted generalisation.
d) Induced implicit attitudes toward original people generalised

completely to the new people. New individual who shared group
membership with the original person who performed primarily
positive behaviours was implicitly preferred over the new
individual who shared group membership with the original person
who performed primarily negative behaviours.
e) Participants were unable to prevent the information about original

individuals from in uencing their implicit responses toward the
new individuals. Automatically and immediately treated the
individual group members as interchangeable.
3. Induced Attitudes Persisted Over Time
a) Individual who performed primarily positive behaviours was

preferred over the individual who performed primarily negative
behaviours.
4. Implicit and Explicit Attitudes Generalised After a Delay
189
fl
a) At second session, self reported attitudes toward original people
generalised completely to the new people.
b) New person who performed primarily positive behaviours was

preferred over the new person who shared group membership
with the original person who performed primarily negative
behaviours.
c) No di erence between e ectiveness of induction on evaluations

of the original and new individuals.
d) Generalised implicit attitudes that were induced immediately

persisted over time. 
Second testing session, the implicit preference between the new
people showed an e ect consistent with the induction materials
describing the original people. 
No di erence between the e ectiveness of the induction on
evaluations of the original.
e) Implicit associations would have a stronger in uence on explicit

judgement as the details of the induction faded in memory.
III. DISCUSSION
A. When people deliberately resist using the behaviours of one individual

to make explicit judgements about another individual from the same
group, the simple association due to shared group membership is
su cient for implicit evaluations of one group member to generalise to
another.
B. Generalisation occurs explicitly even when evaluations are measured

days after the original induction.
C. Despite participants’ ability to consciously resist using group

information as a basis of evaluation immediately, associations between
group members still exist in memory and in uence evaluation as time
passes.
D. Balance Principles predict that if one person who is viewed negatively is

associated with a second person by group membership, then the
second person is likely to inherit the negative evaluation (Greenwald et
al., 2002; Heider, 1958). 
Resisting transferring attitudes from one person to another requires the
190
ffi
ff
ff
ff
ff
ff
fl
fl
ability to di erentiate distinct knowledge about group members and the
cognitive capacity to deliberately refrain from generalisation. 
As information individuating 2 group members fade, resisting
generalisation becomes more di cult. 
THOUGHTS AS MATERIAL OBJECTS TO AFFECT IMPACT ON EVALUATION
Pablo Brinol, Margarita Gasco, Richard E. Petty, and Javier Horcajo. Treating
Thoughts as Material Objects Can Increase or Decrease Their Impact on
Evaluation. 2013.
I. ABSTRACT
A. In Western dualistic culture, it is assumed that thoughts cannot be treated as

material objects; however, language is replete with metaphorical analogies
suggesting otherwise. In this research report, authors examined whether
objectifying thoughts can in uence whether the thoughts are used in
subsequent evaluations.
B. In Experiment 1, participants wrote about what they either liked or disliked

about their bodies. Then the paper on which they wrote their thoughts was
either ripped or tossed in the trash or kept and checked for errors. 
When participants physically discarded a representation of their thoughts,
they mentally discarded them as well, using them less in forming judgements
than did participants who retained a representation of their thoughts.
C. Experiment 2 replicated this nding on thought representation and also

showed that people relied on their thoughts more when they physically kept
them in a safe place, putting their thoughts in their pockets, than when they
discarded them.
D. Final study revealed these e ects were stronger when the action was
performed physically rather than merely imagined.
II. EXPERIMENT 1
A. Method
1. Participants and design 

83 students at public high school. 
Randomly assigned to cells of a 2 x 2 between subjects factorial design.
191
ff
fl
ff
fi
ffi
(thought direction: positive vs. negative) and (Treatment: thought disposal
vs. control).
2. Procedure 
Participants received written instructions asking them to complete several
tasks. 
Each participant randomly assigned to generate and write down either
positive or negative thoughts about his or her own body during a 3
minute period. 
Positive thought condition, participants were told to list as many positive
thoughts about their bodies as they could 
Negative thoughts condition, they were told to list as many negative
thoughts about their bodies as they could. 
After listing, each participant was randomly assigned to either the thought
disposal or control condition. 
All participants were told to look back at the thoughts they wrote. 
Thought disposal condition, participants were asked to contemplate their
thoughts and then throw them into the trash can located in the room. 
Control condition, participants were asked to contemplate their thoughts
and to check for any grammar or spelling errors they could nd. 
Dependent variable was participants’ attitudes toward their bodies. 
Assessed using 3 x 9 point semantic di erential scales.
B. RESULTS
1. Signi cant main e ect of thought direction, thought direction x

treatment interaction.
2. Di erence between positive thoughts condition and negative

thoughts condition was signi cant only for participants in the control
condition and not for participants in the thought disposal condition.
C. DISCUSSION
1. Thought disposal treatment can in uence judgements by invalidating

people's thoughts, results showed that attitudes of participants who
physically threw away their thoughts showed less impact of the
thought direction induction than did the attitudes of participants who
physically retained their thoughts.
III. EXPERIMENT 2 
5 objectives, to provide a conceptual replication of Experiment 1 and show a
decrease in thought use after physical thought disposal. Second, authors
attempted to specify the aspect of the thought disposal treatment that was
192
ff
fi
ff
fi
fl
ff
fi
responsible for the e ects on attitudes by more clearly isolating the behavioural
component of the induction. Third, authors changed the topic of the thought
task in order to generalise their ndings. Fourth, to provide an extension of the
previous ndings by showing that thought use could be increased when
participants physically kept their thoughts safe rather than discarding them.
Finally, previous experiment established that treating thoughts as real objects
can in uence attitudes, but remained an open question whether the impact of
treating thoughts as objects could extent to other outcomes that are
downstream consequences of attitudes, such as behavioural intentions.
A. Method

284 undergraduate students 
2 x 3 between subjects factorial design. (thought direction: positive vs
negative), (treatment: thought disposal, control or thought protection).
2. Procedure 
Participants received written instructions asking them to complete several
tasks. 
Randomly assigned on rst task to list either positive or negative
thoughts about the Mediterranean diet for 3 minutes. 
Reminded that this diet involves high consumption of fruits, vegetables,
legumes, and unre ned cereals, with olive oil as basic fat. 
After listing thoughts, each participant randomly assigned to complete
either thought disposal, thought protection, or control task. 
Participants in thought disposal removed their thoughts physically. 
Participants in thought protection told to put it in their pocket. 
Participants in control condition, told to remove the page and pass to
experimenters for inspection. 
Dependent variable were participants attitudes and behavioural intentions
toward the Mediterranean diet.
3. RESULTS
a) Participants who generated thoughts in favour of the diet held

more favourable evaluations than those who generated thoughts
against the diet.
b) To fully interpret interaction, authors conducted a series of further

contrasts. 
First, the comparison between the thought disposal and control
conditions revealed signi cant interaction between thought
193
fl
fi
fi
ff
fi
fi
fi
direction and condition. E ect of thought direction on evaluation
of diet was greater for control than for thought disposal
participants. (Replicated pattern in Experiment 1).
c) Those who generated positive thoughts in the control condition

(write thoughts pass to experimenters) reported more favourable
evaluations of the diet than did those who generated negative
thoughts.
d) Participants in the thought disposal condition who generated

negative thoughts reported more favourable evaluations of the
diet than those asked to generate positive thoughts.
e) When people treated thoughts as trash, their judgements were

the opposite of their thoughts. 
People might be especially likely to do the opposite of their
thoughts when they have very high doubt about what they have in
mind, when they overcorrect for the direction of their thoughts
(Martin, 1986), or when they are concerned about thoughts that
are framed or perceived to be represented in dichotomous
manner (Beck, 1993; Ellis, 1973).
f) New thought protection condition led to greater thought use than

the control condition. E ect of thought direction on evaluations of
the diet was greater for the thought protection condition than for
the control condition.
g) For thought protection participants, those who generated positive

thoughts reported more favourable evaluations of the diet than
did those who generated negative thoughts.
A. Experiment 1 show that relative to control subjects, participants who

literally threw their thoughts away used those thoughts less in forming
attitudes.
B. Experiment 2 conceptually replicated this research for a di erent topic.

Included new conditions in which participants kept their thoughts in
their possession (e.g. pocket). Compared with participants in other
conditions, those who kept their thoughts showed increased reliance on
their thoughts in forming their attitudes and behavioural intentions.
194
ff
ff
ff
C. Further research required on the possibility of many other interventions,
boundary conditions for e ects on actions relevant to attitude change
and resistance.
D. Important question is to what extent performing the physical action with

the thoughts adds anything beyond merely imagining the action. 
New study was conducted by the authors, asking participants to type
their negative thoughts about the Mediterranean diet and instructed to
create electronic le with those thoughts. 
Some were asked to use mouse to move electronic le into the recycle
bin (physical disposal condition), and others had to use the mouse to
move le to the storage disk (physical protection condition). In
remaining 2 conditions, participants asked to mentally imagine they
would perform that action, but without having to do it. (Mental
protection vs. mental disposal).
1. Participants in the physical disposal condition reduced reliance on

negative thoughts in evaluating the diet compared with participants
in the thought protection conditions.
2. Evaluations of participants in their imagination conditions did not

di er after visualising the action of moving thoughts to be recycled
or to be stored.
3. Di erence between physically moving the negative thoughts to

recycle bin and merely imaging that action was signi cant. 
Di erence between physically storing the le and simply imagining
doing so was signi cant.
4. Physical actions can be more powerful even when performed on a

computer than mere imagination.
5. Practical implications can be found in treatments, and detaching and

separating negative thoughts can decrease negative evaluations.  
The very same treatment (thought disposal) produced the opposite
e ect when thoughts were positive. 
Suggests that techniques in mindfulness treatments can back re, at
least for some people and for some situations.
6. People can develop a closer relationship with their positive thoughts

(e.g. physically carrying them).
195
ff
ff
ff
ff
fi
fi
fi
ff
fi
fi
fi
fi
AUTOMATIC ACTION AND INACTION
C. Neil Macrae, Lucy Johnston. Help, I Need Somebody: Automatic Action and
Inaction. 1998.
I. ABSTRACT
A. Goal of behavioural control is of central importance in everyday life. When

production of an unwanted action can have deleterious consequences for
perceivers, there is considerable virtue in the possession of a mental system
that edits its behavioural products to meet the demands of a challenging
world. Accordingly, in an attempt to extend existing work on this topic (e.g.,
Bargh, Chen, & Burrows, 1996), in the present research the authors have
investigated the extent to which automatic elicitation of action may be
moderated by features of the task environment and perceivers’ goal states.
Findings were equivocal. When inhibitory cues were present in the
environment (i.e. Experiment 1), or perceivers had a competing goal in mind
(i.e. Experiment 2), automatic behavioural priming e ects were eliminated.
We consider the implications of our ndings for recent treatments of
behavioural priming and action control.
B. Action Initiation
1. Automatic behavioural priming is a rather routine consequence of normal,

cognitive functioning. Were action to unfold in any other way, life as we
know it would be fraught with peril (Bargh, 1997; Norman & Shallice,
1986).
2. To be able to do just about anything at all (e.g. driving, dating, dancing),

action initiation needs to be decoupled from the ine cient (i.e. slow,
serial, resources consuming) workings of the conscious mind, otherwise
action inevitably would prevail (Baars, 1997; Mandler, 1997; Norman &
Shallice, 1986). That behaviour however, can be triggered implicitly, as
has been acknowledged for many years (Bargh, 1990, 1994 1997; James,
1890; Ko ka, 1935; Lewin, 1943; Prinz, 1990; Schank & Abelson, 1977).
196
ff
fi
ff
ffi
3. People’s behavioural outputs can be triggered by factors of which they
are completely unaware. Action can unfold, in other words, when the
“lights are o and nobody’s home”.
4. Automatic priming e ects, much like any other cognitive event, are
moderated by a variety of forces. They are controllable. (Bargh, 1997).
C. Action Inhibition
1. Behavioural priming demands a critical match between contents of the

activated action schema and the current task environment, otherwise
inaction will prevail.
2. In complex social settings, correspondence between primed constructs

and the behavioural arena is not the only factor that may moderate the
automatic elicitation of action. Inaction may frequently occur even in
situations where the criterion of correspondence has been satis ed.
3. At any particular point in time, a range of internal (e.g. goals, moral

guidelines) and external (e.g. situational cues, task context) forces will be
competing for the control of one’s behaviour. These forces will frequently
have quite antagonistic behavioural implications, some specifying action,
others inaction.
4. The process of selection whereby willed and nonwilled actions compete

for behavioural control is termed contention scheduling; con ict is
resolved through inhibition of competing action schemas (Norman &
Shallice, 1986; Shallice & Burgess, 1993).
D. Experiment 1 investigates the inhibitory e ect of environmental features,

while Experiment 2 investigates the inhibitory e ect of goal states on the
automatic elicitation of action.
II. EXPERIMENT 1
A. Method

64 female undergraduates 
2 x 2 between subjects design. (Prime: help or control) and (Pen type:
normal vs. leaking).
2. Procedure and Stimulus Materials 

Priming manipulation was a Scrambled Sentence Test under the guise of
a language task. 
For each of the 15 items, participants were requested to use 5 presented
words to produce a grammatical English sentence of 4 words in length. 
197
ff
ff
ff
ff
fl
fi
For each item, 5 words were presented in a scrambled order. 
2 versions of scrambled sentence test were constructed. 1 intended to
prime the construct of helpfulness, while the other was a control
condition in which no speci c construct was activated. 
In priming condition, 10 out of 15 sentences contained an adjective or a
verb semantically related to helping. 
10 critical priming stimuli were (helped, assistance, aided, supported,
provided, encouraging, facilitated, promoted, fostered, and furthered). 
In control condition, critical items were replaced by neutral words that
were not associated with helping in any way. 
Participants told that they would be participating in 2 short experiments
on language use. 
First experiment was investigation on the construction of grammatical
English sentences. 
Participants told to complete the test in their own time and inform the
experimenter, which was blind with regard to priming manipulation. 
After which the experimenter would say she was going to get the
experimenter who was conducting the second experiment. 
Experimenter accidentally drops some items she was carrying. 
Depending on experimental condition, dropped items were either regular
pens or pens that were leaking. 
Results con rmed that participants believed it would be helpful to pick up
both regular pens and leaking pens that were dropped by an individual. 
In a helpful situation we expected the following e ects to emerge where
participants would be unwilling to engage in picking up leaking pens
because of possible costs involved, getting covered in ink.  
Whereas leaking pens should eliminate the automatic elicitation of
helping behaviour, regular pens should not impede the occurrence of
assistance. 
If pen was not picked up in 10 seconds, experiment was terminated.
1. When normal pens were dropped on the laboratory oor,

participants’ helping rate was higher in the prime than control
condition.
2. When pens were leaking, there was a dramatic reduction in helping

rates and no e ect of the priming manipulation was observed.
3. Helping rates were higher for normal than leaking pens in both prime
and control condition.
198
fi
ff
fi
ff
fl
4. Inhibitory impact of environmental cues on the automatic elicitation
of action (Norman & Shallice, 1986; Powers, 1973) were demonstrated.
5. When normal pens were dropped on the oor, primed participants

were more likely to help the experimenter than their colleagues in the
control condition. This replicates previous research of this kind and
re ects automatic elicitation of action. (Bargh et al., 1996).
6. However, when pens were leaking, this automatic helping e ect was
eliminated.
III. EXPERIMENT 2 
Investigated that in social behaviour, with processing objectives inhibiting the
e ects of primed action tendencies when these forces are in con ict and specify
incompatible behavioural outputs. (E.g. imagine situation in which one has been
primed to be helpful, but one is 5 minutes late for an important appointment?)
(Darley & Batson, 1973). Would one still o er assistance to a needy stranger in
such a setting? Authors suspect not. 
Experiment 2 investigated this possibility with a paradigm based closely on this
illustrative example.
A. Method

64 female undergraduates.
2. Procedure and Stimulus Materials 

Replication of Experiment 1, but with couple of modi cations. 
After completion of Scrambled Sentence Test SST, and when
experimenter explained that the participant was now to leave the lab and
walk to the foyer of the Psychology Department where the next
experimenter would be waiting; 
Experimenter will deliver 1 of 2 possible messages. For half of the
participants, experimenter announced that experimental session was
running on time. For others, she intimated that it was running 5 minutes
behind schedule.  
Participant would now leave, but Experimenter would drop items, which
were regular pens. 
To see whether there would be help rendered, where automatic helping
e ect would emerge.
199
ff
ff
fl
ff
fl
fi
ff
fl
1. When experimental session was running on time, participants’
helping rate was higher in the prime condition than in control
condition.
2. In contrast, when session was 5 minutes behind schedule, there was

a dramatic reduction in helping rates and no e ect of priming
manipulation was observed.
3. Helping rates were higher when experimental session was on time,

than when it was running late in both the prime and control
condition.
4. Experiment 2 demonstrated the modulatory impact of perceivers’

goals on the automatic elicitation of action (Norman & Shallice, 1986;
Powers, 1973). 
When experimental session was on time, primed participants were
more likely to help the experimenter than their colleagues in the
control condition. 
However when the session was running late (i.e. participants were in
a hurry), this automatic helping e ect was eliminated.
5. Findings were at odds with Darley and Batson’s (1973) classic study
where helping behaviour was una ected by a priming manipulation
(i.e. reading the parable of the Good Samaritan). Compared to
participants in the control condition (no prime), those primed to be
helpful were indeed more helpful when the experiment was running
on time but not when it was running late.
6. When covert priming manipulation was employed conceptually

comparable e ects were observed.
A. Goal of Control
1. Once primed, behavioural schemas can prompt perceivers to

perform a variety of actions. Participants may interrupt an
experimenter with unusual rapidity, emit hostile expressions when
asked to repeat a laborious task, dawdle away from the lab,
articulate a word list at great speed, or perform with distinction on a
test of their general knowledge. (Bargh et al., 1996; Dijksterhuis & van
Knippenberg, 1998; Macrae et al., 1998). 
These e ects are theoretically noteworthy because they are
200
ff
ff
ff
ff
ff
triggered implicitly. Perceivers have no awareness of causal origins
of their behaviour.
2. By denying introspective access to most of its workings, the mind

can respond exibly and adaptively to the demands of a complex
stimulus world (Baars, 1997; Dennett, 1996).
3. Many behaviours we routinely ascribe to the workings of deliberative

intentional processing in reality require no conscious processing at
all (Bargh, 1990, 1997; Bargh et al., 1996; Bodenhausen & Macrae, 1998;
Carver & Scheier, 1981; Chartrand & Bargh, 1996; Macrae, Bodenhausen,
& Milne, 1998).  
In a sense, things (e.g. actions and decisions), just happen.
4. Probable that most control processes are cognitively impenetrable,

operating in the silent world of the unconscious mind. (Baars, 1988).
B. On Resisting Assisting: Doing the Right Thing
1. Predictions supported that elicitation of automatic action can be

moderated by features of task environment and perceivers’ goal
states. Situational cues (and associated behavioural tendencies) and
processing goals can inhibit the e ects of activated behavioural
schemas (Norman & Shallice, 1986).
2. When participants were primed to be helpful, critical features of task

environment and perceivers’ current goals moderated the elicitation
of automatic behaviour. 
When experimenter dropped regular pens on lab oor, primed
participants were more likely than control participants to o er
assistance. Quite di erent results emerged if the pens were faulty or
if the experimental session was running behind schedule. 
Automatic elicitation of action can be moderated by both exogenous
(situational cues) and endogenous (processing objectives) factors.
(Norman & Shallice, 1986).
3. An adaptive control system actively prevents the elicitation of

inappropriate or incompatible actions.
4. Whether or not a speci c action occurs is determined by the relative

strengths of activated schemas, with schemas routinely provoking
con icting behavioural tendencies. The goal of action control is to
prevent the elicitation of a contextually unwanted action.
201
fl
fl
ff
fi
ff
fl
ff
5. Behavioural control is a competition between activated schemas,
with environmental cues and internal psychological states either
facilitating or inhibiting the elicitation of action (Norman & Shallice,
1986). 
Quite how much, when, and for whom primed action tendencies
shape behaviour, however, are questions that require empirical
clari cation.
C. CONCLUSIONS
1. Behavioural control is of central importance in social interaction.

When an errant response can land one in unseemly trouble, there is
considerable value in the possession of a mental system that
constantly modi es, updates, and edits its behavioural products. The
extent that this system is nely tuned to internal and external cues,
the better its prospects of guiding one’s actions in a purposive
manner. (Powers, 1973).
2. Through action tendencies associated with implicit registration of

external cues in the immediate task environment, behaviour can be
shaped, cajoled, and guided (promoted or inhibited) to meet the
demands of an ever changing stimulus world.
202
fi
fi
fi

PL3238 Summarised Study Guide

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

PL3238 Summarised Study Guide

Uploaded by

Copyright:

Available Formats

Tuesday, 10 August 2021

Week 1 Readings Summary

A. Elemental Origins of Social Cognition Research.

3. Proper methods of testing psychological science emerged. Empirical methods.

2. Gestalt Psychology drew on these initial holistic insights. In contrast to analysis

4. Focused on a person’s own subjective perceptions, not on objective analysis.

6. The Total Psychological Field, (and hence behaviour), is determined by 2 pairs

III. The Ebb and Flow of Cognition in Psychology.

B. Cognition in Experimental Psychology.

1. Wundt’s work at the dawn of empirical psychology as already described, relied

2. Behaviourists believe that specifying an observable stimulus (S) and response ®,

3. Linguists criticised the failure of the stimulus response framework’s attempts to

4. Information Processing arose out of work on how people acquire knowledge

C. Cognition in Social Psychology

a) Consistency Seekers are motivated by perceived discrepancies among

b) Naive Scientist consist of attribution theories that concern how people

d) Motivated Tactician is a fully engaged thinker who has multiple cognitive

A. Mentalism: A Commitment to Cognitive Elements

1. First assumption is that people have an unabashed commitment to mentalism

2. Social Schema. Schema is a cognitive structure that represents one’s general

B. Cognitive Processes in Social Settings

1. Formation of cognitive elements.

C. Cross Fertilisation, Studying Social Cognitive Processes

1. Social Cognition borrows relatively ne-grained cognitive theory and methods

D. Beyond Cognition, Real World Social Issues

V. People Are Not Things

H. People are unavoidable complex. Cognitions cannot be studied without making

Baumeister - Thinking About People

I. Social Psychologists and Thinking About People

A. The Consistency Seeker

B. The Self-Esteem Maximiser

C. The Terror Manager

D. The Information Seeker

E. The Information Processor

F. The Foolish Mistake Maker

G. The Non-di erent Individual / Situational Responder

H. The Impression Manager

I. The Naturally Selected Animal

K. The Group Member

L. The Benighted Layperson

Week 2 Readings Summary

• The world in which we make judgements is full of uncertainties, and we do not

• Occurs because people ignore prior probabilities (base-rates) in the population.

• Representativeness and Medicine

• Representativeness and the conjunction fallacy

Schwarz, N. et al. (1991). Ease of Retrieval as information. Availability

1. Tversky & Kahneman, 1973. Most frequently studied study of Experiment 8,

2. Tversky & Kahneman, 1973, Experiment 3. Subjects were found to overestimate

II. Experiment 1: If It Is So Di cult to Recall, It Cannot Be Typical.

III. Experiment 2: An Extended Replication

IV. Experiment 3: Misattributing the Ease of Recall.

A. Method. 78 female students of a West German University. 2 (examples of assertive

B. Results and Discussion.

V. Final Notes and General Discussion

B. Informative Functions of Subjective Experiences

I. PUTTING ADJUSTMENT BACK IN THE ANCHORING AND ADJUSTMENT

3. Results and Discussion.

3. Results and Discussion.

2. Results and Discussion

3. FINAL GENERAL DISCUSSION

Week 3 Readings Summary

II. Possible Moderation of Ease of Retrieval E ects

B. Evidence Favouring High Elaboration Conditions.

IV. Overview of Present Research

C. Cognition in Social Psychology 

1. Formation of cognitive elements. 

I. Social Psychologists and Thinking About People 

A. The Consistency Seeker 

B. The Self-Esteem Maximiser 

C. The Terror Manager 

D. The Information Seeker 

E. The Information Processor 

F. The Foolish Mistake Maker 

G. The Non-di erent Individual / Situational Responder 

H. The Impression Manager 

I. The Naturally Selected Animal 

K. The Group Member 

L. The Benighted Layperson 

III. Experiment 2: An Extended Replication 

IV. Experiment 3: Misattributing the Ease of Recall. 

B. Results and Discussion. 

B. Informative Functions of Subjective Experiences 

3. Results and Discussion. 

3. Results and Discussion. 

2. Results and Discussion 

3. FINAL GENERAL DISCUSSION 

II. Possible Moderation of Ease of Retrieval E ects 

B. Evidence Favouring High Elaboration Conditions. 

IV. Overview of Present Research 

I. Positive Test Strategy 

A. Biased Memory Search 

B. Biased Evidence Seeking 

C. Hypothesis Con rmation in Evaluating Oneself 

D. Choosing vs. Rejecting 

E. Analysing Reasons for Predictions 

III. Covariation Detection 

A. Assessing Covariation from 2 x 2 Tables 

C. Stereotype Formation Through Illusory Correlation 

D. Failure to Detect Actual Correlations 

E. Accuracy or Inaccuracy in Covariation Detection 

G. Story Models in Jury Decision Making (Explanation Based Judgement) 

IV. Counterfactual Thinking 

B. Ease of Replicating Events Mentally 

C. Functions of Counterfactual Thoughts 

II. Experiment Subjects 

I. State of Mind When Encoding Events 