Professional Documents
Culture Documents
by
Sonia K. Kang
A thesis submitted in conformity with the requirements for the degree o f M aster o f Arts
University o f Toronto
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Library and Bibliotheque et
Archives Canada Archives Canada
NOTICE: AVIS:
The author has granted a non L'auteur a accorde une licence non exclusive
exclusive license allowing Library permettant a la Bibliotheque et Archives
and Archives Canada to reproduce, Canada de reproduire, publier, archiver,
publish, archive, preserve, conserve, sauvegarder, conserver, transmettre au public
communicate to the public by par telecommunication ou par I'lnternet, preter,
telecommunication or on the Internet, distribuer et vendre des theses partout dans
loan, distribute and sell theses le monde, a des fins commerciales ou autres,
worldwide, for commercial or non sur support microforme, papier, electronique
commercial purposes, in microform, et/ou autres formats.
paper, electronic and/or any other
formats.
Canada
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Stigma Sensitivity and Stereotype Threat in Older Adults
Sonia K. Kang
University o f Toronto
Abstract
Although research has shown that older adults are negatively affected by aging stereotypes,
no method is currently available to identify which older adults may be especially susceptible
to these effects. The current research takes the first step toward identifying older adults most
susceptible to the effects o f stereotype threat and investigates the consequence o f stereotype
threat on the well-being o f older adults. We hypothesized that stereotype threat effects
stereotype threat such that stereotype threat effects would be most potent for those highly
identified with the older adult age group, high in stigma consciousness, or high in both state
and trait perceived stereotype threat. We also expected to see a decrease in positive affect for
those in the threat condition, with no change in positive affect for those in the no threat
condition. The results o f our study showed that stereotype threat disrupted performance on
the free recall task, but no moderation o f this effect was found. However, stereotype threat
effects on cued and recognition recall tasks were moderated by stigma consciousness, and
state and trait perceived stereotype threat. Results and implications for negative and positive
ii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1
“In all human relations - familial, ethnic, international - the engendering power o f
expectancy is enormous. If we foresee evil in our fellow man, we tend to provoke it;
Referred to by some as the “last bastion o f bigotry”, ageism is thought to represent the
most socially condoned and institutionalized form o f prejudice in the world today (Nelson,
2002). However, despite the prevalence o f age discrimination, relatively little research has
examined age-based prejudice compared to prejudice based on race or gender (Chasteen, 2000;
Nelson, 2002). This lack o f research is surprising given that unlike relatively fixed membership
in racial and gender categories, membership in age categories is transient, with individuals
transitioning into and out o f age groups with each passing year. For most people, being
classified as an older adult is simply a matter o f time and, because o f this, research on age-based
prejudice and its consequences is o f utmost importance. One major problem currently facing
older adults is stereotype threat, a mechanism through which anxiety about fulfilling a negative
stereotype about one’s group leads to decreased performance and, therefore, stereotype
fulfillment (Steele & Aronson, 1995). Akin to Allport’s notion o f provoking or eliciting
expected behaviour from a stereotyped other, stereotype threat works cyclically to create and
maintain stereotyped beliefs. Although research has shown that older adults are often negatively
affected by aging stereotypes (Chasteen, Bhattacharyya, Horhota, Tam, & Hasher, 2005; Hess,
Auman, Colcombe, & Rahhal, 2003), no method is currently available to identify which older
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2
adults may be especially susceptible to these effects. Research has also yet to systematically
examine the effect o f stereotype threat on the well-being o f older adults. The present research
project attempts to address these gaps in the literature by specifically examining the effect o f
stereotype threat on emotions o f older adults, and takes the first step toward identifying older
adults most susceptible to its negative effects. By assessing the specific emotional consequences
o f stereotype threat and identifying those most vulnerable to its effects, we hope to gain the
information necessary to develop a powerful and targeted intervention method to inoculate older
Age Stereotypes
Research on age stereotypes is an area unique from research on gender and racial
stereotypes in a number o f ways. One o f these is the relatively normative and institutionalized
existence o f negative views o f aging and o f the older adult community as a whole (for reviews
see Pasupathi & Lockenhoff, 2002; Wilkinson & Ferraro, 2002). Within a cultural context,
ageism can be seen embedded in our language, humour, music, art, literature, television, and
advertising.
According to Butler (1975), words used to refer to older adults and aging in general are
good indicators o f cultural views o f aging. Covey’s (1988) analysis o f English terms for older
people uncovered a grim view o f our society’s perception o f aging as the terms we use to refer to
older adults (e.g. grumpy, crotchety, slowpoke) reflect a decline in the status o f elders and an
increased focus on the debilitative effects o f aging. The frequent use o f “old” as an adjective to
mean useless or undesirable also points to our negative view o f older adults. More recent
research shows that the way we speak to older adults reflects our negative views regarding their
competence and worth (Coupland & Coupland, 1993; Ryan, Hamilton, & Kwong See, 1994).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3
Most work on age-related humour has focused on the content o f jokes, cartoons, and
birthday cards. This work has consistently found evidence for the propagation o f negative views
o f aging through these media (Dillon & Jones, 1981; Palmore, 1971; Smith, 1979). Similarly,
analyses o f 300 pieces o f music with age-related content revealed a negative view o f aging and
old age in the majority o f pieces (Cohen & Kruschwitz, 1990). Television and advertising are
also culpable for their infrequent and relatively negative portrayal o f elders and the aging process
(Robinson & Skill, 1995). Because o f the common and pervasive negative view o f aging and
older adults in our society, ageism is quite different from racism and sexism, both o f which are
relatively unaccepted and censured. The normalcy o f ageism makes its study extremely
important, as the eradication o f such a widely accepted prejudice will be particularly difficult.
Even though research on old age stereotypes suggests an overall negative attitude toward
older adults (Crockett & Hummert, 1987; Kite, Stockdale, Whitley, & Johnson, 2005), there is
still no consistent and reliable pattern o f findings in this area. A generally accepted framework
presents three characteristics o f old age stereotypes: they contain both positive and negative
elements, they are generally found to be pervasive across cultural and ethnic groups, and they are
According to the Stereotype Content Model (SCM; Fiske, Xu, Cuddy, & Glick, 1999),
stereotypes can be described along two critical dimensions: warmth and competence. A
stereotype may describe a group as either warm or cold, and competent or incompetent.
Research on the perception o f older adults has found that aging stereotypes are mixed, describing
older adults as warm but incompetent, reflecting the simultaneous existence o f both positive and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4
negative perceptions (Cuddy, Norton, & Fiske, 2005; Fiske et al., 1999; Hummert, Garstka,
Shaner, & Strahm, 1994). The SCM predicts that the degree to which groups are rated as warm
or competent has direct implications for the type o f emotional prejudices held about the group.
Groups that are seen as warm but incompetent tend to elicit pity. Although pity may seem
innocuous at first, it can actually have detrimental effects through the creation o f a self-fulfilling
prophecy (Allport, 1954; Cuddy, Norton, & Fiske, 2005; Harris, 2005). In the medical health
field in particular, researchers have found that doctors and other geriatric specialists often convey
the message that older adults are helpless through their linguistic expressions o f sympathy. This
message o f helplessness is internalized by some older adults and has actually been shown to
decrease their independence (Cohen, 1990). The feeling o f pity towards older adults is often in
conflict with a feeling o f admiration towards this group (Fiske et al., 1999). The contrasting
attitudes o f pity and admiration highlight the mixed character o f the old age stereotype and point
attitudes toward younger and older adults (Kite, Stockdale, Whitley, & Johnson, 2005) updating
their first meta-analytic review conducted in 1988 (Kite & Johnson, 1988). In the 2005 review,
232 effect sizes were analyzed across five categories: age stereotype (e.g., old-fashioned, talks a
lot about the past), attractiveness (e.g., pretty, wrinkled), competence (e.g., intelligent, good
memory), behaviour (e.g., willingness to interact with, make a phone call to), and evaluation
(e.g., generous, friendly). Across these five categories, it was found that older adults were
consistently judged more negatively than younger adults. The most negative bias towards older
adults was seen in the domain o f age stereotype, indicating that aging stereotypes about older
adults are much more negative than aging stereotypes about younger adults. Some negative
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5
aging stereotypes include the ideas that older adults are forgetful, complaining, slow, or rude.
Older adults were also consistently judged to be less attractive, less competent, less desirable
social interaction partners, and were evaluated less positively overall, than younger adults.
Although a negative bias towards older adults was seen across all five domains, the negative bias
in the behavioural and evaluation domains were relatively less severe than in the other
categories. This result echoes the assertion o f Cuddy, Norton, and Fiske (2005) that older adults
are viewed as warm but incompetent. In the meta-analytic review, Kite et al., (2005) found that
even though negative stereotypes exist about older adults, and they are seen as less attractive and
One study that specifically examined stereotypes o f older adults was conducted by
Hummert and her colleagues (Hummert, Garstka, Shaner, & Strahm, 1994). In this study, cluster
analyses were used to organize traits used to describe older adults by young, middle-aged, and
older adult participants. The findings suggest that older adults have more complex mental
representations o f older people as a group than do young or middle-aged adults as the older
adults grouped traits into significantly more categories than either o f the younger age groups. In
addition, consistent with other research that has found stereotypes o f the elderly to be mixed,
there were seven trait categories that were shared across all age groups, with some categories
being positive (e.g., perfect grandparent (loving, wise)) and others negative (e.g., severely
Another suggested feature o f aging stereotypes is that they are pervasive across cultures.
However, research in this are has found inconsistent results with respect to cross-cultural
differences. Levy and Langer (1994) explored cross-cultural views o f aging and their effects on
memory performance among young and old Americans, D eaf Americans, and the Chinese. This
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
study supports the notion that individuals from Eastern cultures view older adults more
positively and with more respect than individuals from Western cultures. The results o f Levy
and Langer’s research have been both supported (Harwood, Giles, McCann, Cai, Somera,
Gallois, & Noels, 2001) and challenged (Gattuso & Shadbolt, 2002; Sharps, Price-Sharps, &
Hanson, 1998; Yoon, Hasher, Feinberg, Rahhal, & Winocur, 2000) by subsequent research
efforts. It remains unclear whether or not aging stereotypes and their effects are pervasive across
cultures or if a distinction between Eastern and Western cultures in the domain o f aging
Following from work on age stereotypes, Erber has examined how existing aging
stereotypes affect the way that older adults are perceived by society. In one experiment (Erber,
Szuchman, & Rothberg, 1990), young and old participants read vignettes about younger and
older individuals who experienced identical memory failures. Participants were asked to indicate
how often that type o f memory failure would have to occur before they would suggest that a
medical or psychological intervention be sought out. The results o f this study showed that when
an older adult was depicted as experiencing a memory failure, the failure was most often
attributed to an internal cause, such as mental difficulty or lack o f ability. However, when
attributing the cause o f the younger individual’s memory failure, participants were more likely to
indicate lack o f effort or attention as possible causes. Also, younger adults judged memory
failures much more harshly than older adults, reporting that they would recommend a medical or
psychological intervention after only a few experiences o f memory failure. The results o f this
research add to the evidence suggesting that old age stereotypes, in this case, forgetfulness, lead
older adults to be treated differently and, often, more negatively than younger adults.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
7
The study o f aging stereotypes is important because stereotypes are known to affect both
the self-concept and actions o f the stereotyped individual (Guo, Erber, & Szuchman, 1999;
Steele, 1997; Wheeler & Petty, 2001). For example, holding negative age self-stereotypes
correlates with negative consequences such as decreases in cognitive performance, will to live,
positive affect, and physical health (Levy, 2003). Simple exposure to negative aging stereotypes
leads to behavioural changes such as a decrease in walking speed (Hausdorff, Levy, & Wei,
In another study, Levy (1996) looked to see if positive and negative aging stereotypes
could be activated via priming and whether or not this activation would have an effect on
memory performance. In this study, both young and old participants were primed with positive
or negative age-related primes. The results o f this study showed that older adults exposed to
negative age-related primes (e.g. decline, senile) showed a decrease in memory performance. In
contrast, older adults exposed to positive age-related primes (e.g. wise, guidance) actually
showed an improvement in memory performance. The results o f this study have important
implications for memory interventions with older adults. For example, simply making older
adults aware o f positive aging stereotypes may have cognitive benefits. In an extension o f this
work, Levy and her colleagues (Levy, Hausdorff, Hencke, & Wei, 2000) found that exposure to
negative aging stereotypes increased cardiovascular stress in older adults compared to those
exposed to positive age stereotypes. These results highlight the potential for negative aging
stereotypes to have detrimental effects not only on cognitive performance o f older adults, but
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
8
fulfilling prophecies for the targets o f these stereotypes. Being aware o f a stereotype about one’s
group may lead group members to question the applicability o f the stereotype to themselves.
Finding confirmation o f the stereotype in one’s characteristics or behaviours lends support to the
stereotype and, when the stereotype is negative, this can have disruptive effects on the self
Stereotype Threat
stereotyping is through a mechanism known as stereotype threat (Steele & Aronson, 1995).
Stereotype threat occurs when concerns about fulfilling a negative stereotype about one’s group
disrupt performance on tasks related to the stereotype. Stereotype threat has been shown to
affect members o f a variety o f stereotyped groups, including older adults. In one o f the earliest
examinations o f stereotype threat, Steele and Aronson (1995) looked at stereotype threat effects
condition, an experimental task was presented as diagnostic o f intellectual ability, activating the
stereotype that Black people perform poorly on intellectual tasks. This led African American
participants in the stereotype threat condition to perform more poorly than Black participants in
the non-stereotype threat condition or White participants in either condition. After controlling
for SAT performance, Black participants in the non-threat condition performed equally as well as
White participants. These results also showed that stereotype threat can be activated through
relatively minor experimental manipulations. For example, simply asking Black participants to
report their race on a piece o f paper was enough to impair their intellectual performance. As
well, Inzlicht and Ben-Zeev (2000) showed that the mere salience o f situational cues that activate
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
9
social identity, such as the race or gender o f other people in the room, was enough to activate
stereotype threat effects. Stereotype threat is especially worrisome because even those who do
not believe a stereotype or do not believe that it is true o f themselves can be affected by
Stereotype threat research conducted with older adult participants has shown that older
adults are susceptible to the effects o f stereotype threat (Chasteen, Bhattacharyya, Horhota, Tam,
& Hasher, 2005; Hess et al., 2003; Rahhal, Hasher, & Colcombe, 2001). For example, Hess and
his colleagues (Hess et al., 2003) found that older adults exposed to stereotype threat performed
more poorly on a recall task than both younger adults and older adults not exposed to threat.
Chasteen et al. (2005) found that the more older adults perceived themselves to be victims o f
stereotype threat, the worse their memory performance was. Although research has shown that
older adults are often affected by aging stereotypes, no method is currently available to identify
Moderation o f the stereotype threat effect has been studied previously by other
researchers. For example, researchers have shown that group members can be differentially
affected by stereotype threat based on the degree to which they value the domain in question
(Aronson, Lustina, Good, Keough, Steele, & Brown, 1999; Hess et al., 2003; Stone, Lynch,
Sjomeling, & Darley, 1999). Intuitively, it seems reasonable that negative stereotypes about a
certain domain would only affect an individual if that individual cares about the domain, or, at
least cares about being viewed as lacking ability in that domain. Aronson and his colleagues
showed that White men who highly valued achievement in the math domain were more affected
by a stereotype threat procedure that pitted performance o f White males on a math test with that
o f Asian males, than were White males who placed less value on math achievement. Likewise,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
10
Hess and his colleagues showed that stereotype threat in the domain o f memory was more
disruptive for older adults who reported to value their memory ability more highly than those
who did not seem to value memory achievement as much. Finally, Stone and his colleagues
showed that White males who viewed athletic ability as a measure o f self-worth, were more
affected by a stereotype threat manipulation comparing the athletic ability o f White and Black
men.
Other research has shown that stereotype threat effects can be moderated by identification
with the stereotyped group. For example, Schmader (2002) showed that women who were more
highly identified with their gender group were more affected by negatives stereotypes about
women and math ability. Following from this research, Pronin and her colleagues (Pronin,
Steele, & Ross, 2004) found that women were actually able to buffer the effects o f stereotype
reducing their identification with their gender. Similar tests with age group identification are
The goal o f the current research project is to identify those older adults who are most
susceptible to stereotype threat and to test our identification method by determining whether
a number o f potential moderators o f the stereotype threat effect: stigma consciousness, age group
identification, and both state (situational) and trait (dispositional) perceived memory-related
stereotype threat.
o f the likelihood that s/he will be stereotyped in a given situation. Stigma consciousness has also
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
11
been referred to as stigma sensitivity, the tendency to expect, perceive, and be influenced by
negative stereotypes about one’s group (Aronson & Inzlicht, 2004). It has been suggested that
stereotype threat (Aronson & Inzlicht, 2004). Stigma consciousness has been shown to moderate
stereotype threat effects in women in the domain o f math performance (Brown & Pinel, 2002).
In this study, women high in stigma consciousness performed worse on a math test while under
threat than women low in stigma consciousness also under threat. As expected, there was no
difference in math performance for women high or low in stigma consciousness in the no threat
condition. In the present study, we will measure age-based stigma consciousness using a
modified version o f the Stigma Consciousness Questionnaire (SCQ; Pinel, 1999), to determine
whether stereotype threat effects in older adults are conditional on seniors’ levels o f stigma
consciousness.
The second potential moderator in our study, age group identification, refers to the extent
to which individuals identify with their age group (Garstka, Branscombe, & Hummert, 1997).
As previously mentioned, group identification has been shown to moderate racial and gender
stereotype threat effects (e.g. Mendoza-Denton et al., 2002; Nosek, Banaji, & Greenwald, 2002;
Pronin, Steele, & Ross, 2004). Recently, age group identification has also been shown to
moderate prejudice toward older adults (Chasteen, 2005; Packer & Chasteen, 2006). These
researchers showed that young adults who were strongly identified with their age group
responded more negatively toward older adults after imagining themselves as an older adult than
did those younger adults who were weakly identified with their age group. However, it has not
yet been shown whether or not age group identification moderates susceptibility to stereotype
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
12
Our third potential moderator is a construct we call perceived stereotype threat (Chasteen
et al., 2005). Perceived stereotype threat measures the extent to which individuals expect and
perceive stereotyping against them. Thus far, it has been found that perceived stereotype threat
is inversely related to older adults’ memory performance, such that memory performance
decreases as perceived stereotype threat increases (Chasteen et al., 2005). We have modified the
original measure o f perceived stereotype threat in order to make a distinction between state
(situational) and trait (dispositional) perceived stereotype threat. State perceived stereotype
threat refers to the extent to which an individual expects and perceives stereotyping related to
Other measures o f interest in this study include stereotype activation, subjective age
identification, and memory identification. Stereotype activation will allow us to assess how
influenced by the manipulation should show activation o f negative age-related and general
o f age, irrespective o f their actual chronological age. For instance, a person might be aged 60
but feel like a 40 year-old. It may be the case that individuals who identify with a younger age
group may not find stereotypes about older adults to be personally relevant and would therefore
not be susceptible to stereotype threat effects. Memory identification refers to how important the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
13
Predictions
Our first set o f hypotheses refers to the main effect o f our stereotype threat manipulation.
We predicted that participants in the threat condition would perform more poorly on all tasks o f
Environmental Support Hypothesis (Craik, 1986), age-related differences are most pronounced
on tasks that provide low environmental support (e.g. context, cues) and, as such, we expected
the threat/no threat difference to be most pronounced on the free recall task, as this task provides
suggests that as we age, emotion-related goals become increasingly important. Following from
this, Carstensen and Turk-Charles (1994) showed that older adults recalled emotional material
better than neutral material when presented with a free recall task. Because o f this heightened
attention for and recall o f emotional materials, we expected memory for emotional information
to be less affected by stereotype threat than memory for neutral information. If this emotionality
bias is maintained under stereotype threat, it will provide us with another option for an
emotionally relevant may increase memory performance for previously neutral information.
In terms o f change in affect, we expected to see a decrease in positive affect for those in
the threat condition, with no change in positive affect for those in the no threat condition. We
did not predict a change in negative affect for either condition, as older adults report low levels
o f negative affect in general (Charles, Reynolds, & Gatz, 2001; Mroczek, 2001), and we
expected this to create a restricted range o f responses where we would be unlikely to see a
significant change.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
14
consciousness, age identification, and perceived stereotype threat. Stereotype threat effects
should be most potent for those highly identified with the older adult age group, high in stigma
consciousness, or high in both state and trait perceived stereotype threat, compared with those
Our hypotheses for the remaining variables were as follows. We expected stereotype
activation to be higher for those in the threat condition than for those in the no threat condition.
As well, we thought that memory identification and subjective age identification might act as
moderators o f stereotype threat. Those highly identified with the domain o f memory
performance will likely experience a greater decrease in memory performance due to stereotype
threat than those low in memory identification, as shown in previous research (Aronson, Lustina,
Good, Keough, Steele, & Brown, 1999; Hess et al., 2003). Lastly, stereotype threat might be
moderated by subjective age identification such that only those subjectively identified with their
Method
Forty-two female community dwelling older adults (Mage = 71.14 years, range = 6 2 - 8 4
years) volunteered to participate in this study. In this experiment, we included only females in
our examination because older females are known to be targeted much more often and severely
by age-based prejudice than their male peers (Kite & Wagner, 2002). By only testing female
participants were initially recruited through the Adult Volunteer Pool (AVP) in the Department
o f Psychology at the University o f Toronto, St. George campus. All participants completed at
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
15
least 11 years o f education (Mediation = 15.24 years, SD = 2.31 years) and the majority were o f
European descent (36 European, 3 African, and 3 other ethnic origin). O f the 11 participants
who did not consider English to be their first language, all learned English relatively early in
their lives (Mage = 10.40 years, SD = 7.21 years). The majority o f participants rated their
physical health as either good or excellent, and only one participant rated her health as poor.
Participants were assigned to conditions such that the threat (n = 20) and no threat (n =
22) condition were balanced for age (p > .69) and SCQ score {p > .15). Participants were paid
Materials
In the first part o f this study, we mailed a set o f questionnaires to members o f the Adult
Volunteer pool. Participants completed these questionnaires in their own homes. The mail-out
contained a modified version o f the SCQ (Pinel, 1999) and an age group identification
Age stigma consciousness. We modified the SCQ from its original form in order to
measure sensitivity to age stereotypes. The modified scale consisted o f ten statements that
address old age and aging stereotypes and the perceptions and experiences participants m ay have
had with them (e.g., “I never worry that my behaviours will be viewed as stereotypic o f older
adults”, “M y being an older adult does not influence how younger people act with me”).
Participants responded to these items on a scale from 0 (strongly disagree) to 6 (strongly agree).
Prior to analyses, the first item o f the questionnaire was removed due to poor internal
consistency. The remaining nine item scale had acceptable internal consistency: Cronbach’s a =
0.70.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
16
Age group identification. The age group identification scale (Garstka et al., 1997)
consisted o f 13 items assessing an individual’s identification with their age group (e.g. “I value
responded to the items on a scale from 1 (strongly disagree) to 7 (strongly agree). The 13 item
age group identification scale had excellent internal consistency: Cronbach’s a = 0.97.
In the second part o f the study, participants completed the experimental tasks in a
laboratory on campus. These measures appear next in the order in which they were completed.
Current affect. Participants completed a 20 item emotion scale to assess their current
affect both before and after the experimental manipulation. The scale contains 10 positive items
upset”). Responses to the items were made on a scale from 1 (not at all) to 4 (very much). Both
administrations o f the emotion scale had good internal consistency: pre Cronbach’s a = 0.81, and
post Cronbach’s a = 0.80. In order to examine participants’ affective reaction to the stereotype
threat manipulation, change in negative affect and change in positive affect were computed by
Prose passage. Participants were presented with a prose passage o f approximately 650
words taken from Agatha Christie’s (1972) novel Elephants Can Remember. This passage was
used by Carstensen and Turk-Charles in their examination o f older adults’ memory bias for
emotional information (Carstensen & Turk-Charles, 1994). In this passage, a godmother and her
adult goddaughter discuss past and current social events and an interaction the godmother has
with a mutual acquaintance. For analysis, this passage is divided into 56 information “chunks”,
which can be further divided into 25 neutral and 31 emotional items. In the threat condition, the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
17
prose passage appeared in a stylized booklet designed to look like a formal test booklet. In the
no threat condition, the prose passage was presented in simple word-processor format.
Filler tasks. Participants completed a series o f filler items consisting o f the Shipley
vocabulary test (Shipley, 1940), a digit comparison task, and a verbal fluency task.
Recall tasks. Memory for the prose passage was assessed using tests o f free, cued, and
recognition recall. The free recall task asked participants to report everything they could about
the content o f the passage they had read a few minutes before. Free recall responses were scored
for the number o f information “chunks” recalled out o f a possible total o f 56 chunks. The
number o f emotional and neutral items recalled was also recorded, and analyzed as the
percentage o f emotional or neutral items recalled out o f the total number o f emotional or neutral
The cued recall task consisted o f 15 short answer questions that asked participants to
recall specific information about the characters and plot o f the passage (e.g. “What was the
colour o f the dress described in the passage?”, “What is Mrs. Oliver’s first name?”).
The recognition recall task consisted o f 15 quotations taken from the passage.
Participants were asked to identify the speaker o f each quotation from a list o f five options.
The appearance o f memory tasks differed between the threat and no threat conditions. In
the threat condition, the tasks were framed as tests diagnostic o f memory performance, and were
made to appear as formal test booklets. Participants were asked to list their age and initials at the
top o f each page o f each task. In the no threat condition, the tasks were framed as reading
comprehension tasks and were presented in simple word processor format. Participants in the no
threat condition were only asked to list their initials at the top o f each page o f each task.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
18
Stereotype activation. In order to assess activation o f old age stereotypes and bias or
questionnaire consisted o f 30 word completion items, each o f which can be completed with
multiple words. O f the 30 items, 10 target old age stereotypes (e.g. lonely, senile, weak), 10
target words related to bias or prejudice (e.g. bigot, hate, degrade), and 10 were filler items (e.g.
tone, dice, gold). Participants were instructed to “fill in the blanks to complete the fragment to
Perceived stereotype threat. Participants completed a nine item measure to asses their
perception o f stereotype threat during the experiment. The perceived stereotype threat (PST)
measure was initially adapted by Chasteen and her colleagues (2005) from Steele and Aronson
(1995). We further modified the perceived stereotype threat measure in the present study in
order to assess both state (situational) and trait (dispositional) perceptions o f stereotype threat.
Responses were made on a scale from 1 (strongly disagree) to 5 (strongly agree). Analyses were
conducted with a four item measure o f state PST (e.g. “Today I felt the experimenter expected
me to do poorly because o f my age”) and a three item measure o f trait PST (e.g. “In general,
people often underestimate my memory ability because o f my age”). Both o f these scales had
good internal consistency: state PST Cronbach’s a = 0.86, and trait PST Cronbach’s a = 0.81.
participants’ thoughts and feelings before, during, and after the experimental procedure.
Subjective age identification. A three item questionnaire was used to determine the age
circle an age group in 10 year increments from 20s to 90s. The three items on this questionnaire
were taken from Goldsmith and Heiens’ (1992) examination o f subjective age identification (e.g.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
19
“I feel as though I am in my . . “In most ways, I am like a person in his/her.. “My interests
and activities are mostly those o f a person in his/her.. A composite score o f subjective age
identification was created by averaging participants’ responses to the three items: Cronbach’s a =
0.90.
from the math identification questionnaire (Brown & Josephs, 2000). Responses to this
composite score o f memory identification was created by averaging participants’ responses to the
three items (e.g. “M y memory ability is important to me”, “How I perform on memory tasks is
Procedure
brought into the laboratory for the second part o f the study. Participants were assigned to either
the threat or no threat condition. Those in the threat condition participated along with a female
young adult confederate; participants in the no threat condition participated in same-age pairs. In
the threat condition, the confederate was instructed to act as though she was completing all o f the
Threat condition. To create our stereotype threat manipulation, we used methods devised
in prior research (Inzlicht & Ben-Zeev, 2000; Schmader & Johns, 2003), including having
participants list their age on the memory tasks, completing the task in the same room as a young
adult confederate, and telling them that the tasks are diagnostic o f memory ability. When
welcomed to the laboratory, participants were told that they would be participating in a study
involving a reading comprehension task that would be diagnostic o f their memory ability.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
20
Participants were specifically told that we were interested in their ability to remember the
information contained in the passage they would read. Participants then read and completed the
consent form.
After hearing some general experimental instructions, participants completed the emotion
scale. Next, participants were presented with the prose passage and were told that their memory
for the passage would be tested later. Participants were allowed 6 minutes to read the passage;
for this and all other tasks, participants in the threat condition were timed conspicuously with a
stopwatch.
Next, participants completed the filler tasks. They were given 4 minutes to complete the
Shipley vocabulary test, 6 minutes to complete the digit comparison task, and 5 minutes to
After the filler tasks, participants completed the memory tasks, which included measures
o f free recall, cued recall, and recognition recall. Before each section, participants were told that
they were completing a memory test that would measure how much they were able to remember
about the passage they had read earlier. Participants were instructed to write their initials and
age at the top o f each page o f each booklet, and were given 5 minutes to complete each recall
task.
Immediately following the recall tasks, participants completed the emotion scale for the
second time. Following this, participants completed the rest o f the experimental materials in the
suspiciousness check and a demographics sheet. Participants were then debriefed verbally and
given a debriefing sheet to read. At this time, the confederate left the room so that the older
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
21
adult would have the opportunity to privately ask questions if she so desired. The experimenter
answered the participants’ questions if asked, and the participant was then compensated,
No Threat Condition. A similar procedure was used in the no threat condition, with the
following exceptions. When welcomed to the laboratory, participants were told that they would
Participants were specifically told that we were interested in their thoughts and opinions about
the passage they would read. Participants then read and completed the consent form.
After hearing some general experimental instructions, participants completed the emotion
scale. Next, participants were presented with the prose passage and were told that we would later
ask them about their thoughts and opinions about the contents o f the passage. Participants were
allowed 6 minutes to read the passage; for this and all other tasks, participants in the no threat
For the memory tasks, participants were told that they were completing reading
comprehension tasks that would help us to learn more about their thoughts and opinions about
the contents o f the passage they had read a few minutes before. Participants were instructed to
write their initials at the top o f each page o f each booklet, and were given 5 minutes to complete
Results
The following analyses explore the dependent variables o f interest in this study, as well
as the potential moderators o f our stereotype threat manipulation. Our primary dependent
variable o f interest, recall performance, was assessed in three different ways using tasks o f free,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
22
cued, and recognition recall. In addition to our analyses on recall performance in general, we
also conducted exploratory analyses to examine potential differences in recall for emotional
versus neutral information contained within the prose passage. Another dependent variable o f
interest was participants’ affect from pre to post experimental manipulation. Our main
independent variables o f interest were the stereotype threat manipulation, stigma consciousness,
age group identification, and perceived stereotype threat. We were also interested in how the
stereotype threat manipulation might be moderated by the other independent variables, as well as
variables, stereotype threat, or the condition manipulation, was treated categorically, while all
other independent variables were treated continuously and examined using regression analyses.
first wanted to establish that our manipulation o f stereotype threat was effective. To do this, a
series o f t-tests were conducted on the recall and affect data, with all tests being two-tailed. The
stereotype threat manipulation had a significant effect on participants’ free recall performance
(maximum recall = 56 chunks), t(40) = 2.20, p < .05, d = 0.68, such that participants in the threat
condition (M = 6.30, SD = 3.47) recalled less than those in the no threat condition (M = 8.59, SD
= 3.29).
We next examined free recall for emotional versus neutral elements. Means are reported
as the percentage o f emotional or neutral items recalled out o f the total number o f emotional (31)
or neutral (25) items in the passage. A significant difference in the percentage o f emotional
items recalled by participants in each condition was found, t(40) = 3.13, p < .01, d = 1.00, with
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
23
those in the no threat condition (M = 9.82, SD = 5.93) recalling a higher percentage o f emotional
items than those in the threat condition (M = 5.00, SD = 3.70). There was no effect o f the
stereotype threat manipulation on recall for neutral items, f(40) = 1.04,/? =.30), d = 0.16.
No significant effects o f the stereotype threat manipulation were found for the other two
measures o f recall, cued, t(40) = 0.53,/? = 0.60, d = 0.16, and recognition, t(40) = 0.61, p = .54,
Participants’ affect data were analyzed for changes in positive and negative emotions
from before to after the experimental manipulation. As hypothesized, participants in the threat
and no threat condition did not differ in the extent to which they experienced a change in
negative emotions, t(40) = 0.62,p = .54, d = 0.19. In terms o f change in positive affect, there
was a trend in the hypothesized direction, suggesting that those in the threat condition (M = -
0.31, SD = 0.59) may have experienced more o f a decrease in positive emotions than those in the
Hierarchical linear regression analyses were used to test the predictive value o f the
independent variables on each dependent variable o f interest. In each case, condition (threat or
no threat) and the independent variable in question were entered in step 1 to assess main effects.
Condition was entered using a dummy-coded variable (0 = no threat, 1 = threat). The interaction
between condition and each independent variable was added in step 2. In all cases, z-score
standardized values o f the independent variables were used (Jaccard, Wan, & Turrisi, 1990).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
24
Stigma Consciousness
Stigma consciousness scores ranged from 1.33 to 4.44, with a mean o f 2.63 {SD = 0.74).
Stigma consciousness scores o f participants assigned to the threat (n = 20, M = 2.80, SD = 0.62)
and no threat (n = 22,M = 2.47, SD = 0.81) conditions did not differ, *(40) = 1.47,/? = .15, d =
0.46.
interaction for both cued {fi = 0.482, p < .05; Step 2: AR2 = 0.15, p < .05) and recognition (J3 =
0.54, /?< .01; Step 2: AR2 = 0.18,/? < .01) recall performance. For both types o f recall, a similar
relationship was found using post-hoc tests. In the threat condition, both cued and recognition
recall performance increased as stigma consciousness increased (cued: /? = 0.48,/? < .05;
recognition: /? = 0.61,/? < .01). In the no threat condition, recall performance remained stable as
stigma consciousness increased (cued: /? = -0.19,/? > .38; recognition: /? = -0.250,/? > .26).
Stigma consciousness did not significantly predict participant results on any o f the other
Because one participant in the no threat condition chose not to complete the age group
identification scale, analyses are based on 21 participants in the no threat condition and 20
participants in the threat condition. Scores on the age group identification scale ranged from 1 to
7 with a mean o f 4.75 {SD = 1.66). Age group identification scores o f participants assigned to
the threat (M = 4.86, SD =1.51) and no threat (M = 4.64, SD = 1.82) conditions did not differ,
A significant effect o f age group identification on free recall ( f = -0.26, p = .09; Step 1:
R = 0.18,/? < .05) emerged such that performance on the free recall task decreased as
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
25
identification with the older adult age group increased. No interaction with condition was found,
/? = -0.08,/? = 0.68; Step 2: AR2 = 0.01 ,/? = .68; for this analysis, only step 1 was significant.
A significant effect for the condition by age group identification interaction also emerged
for the dependent variable change in negative affect, /? = -0.47, p < .05; Step 2: AR = 0.13, p <
.05. In the threat condition, there was no relationship between age group identification and
negative affect, ft = -0.33, p = .16; in the no threat condition, participants experienced a marginal
No other significant effects for age group identification were found, all Fs < 2.11.
Mean state PST scores ranged from 1 to 5, with a mean o f 1.89 (SD = 1.03). Although
we expected state PST to be higher in the threat condition, there was no main effect o f condition
on state PST, (1(40) = 0.74, p - .46, d - 0.23, linear regression analysis showed a trend for the
effect o f state PST on free recall o f neutral items such that free recall o f neutral items decreased
Analyses also revealed a significant effect o f state PST on cued recall when collapsed
across condition, /? = -0.41 ,P < .01; Step 1: R2 - 0.17, p < .05. Performance on the cued recall
Also for cued recall, a trend for the condition by state PST interaction emerged, ft - -0.37,
p = .09; Step 2: AR2 = 0.06,/? = .09. As shown in Figure 1, participants in the threat condition
experienced a decline in cued recall as state PST increased (fi = -0.57,/? < .01) while cued recall
performance for those in the no threat condition remained stable as state PST increased, [i = -
0.16,/? = 0.48.
There were no other significant results involving state PST, all Fs < 1.72.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
26
Participants’ mean levels o f trait, or chronic, PST ranged from 1 to 5, with a mean o f 2.67
(SD = 1.09). As hypothesized, our stereotype threat manipulation had no effect on trait PST,
With increased trait PST, performance on the free recall task decreased, ft = -0.41 ,P <
.01; Step 1: R2 = 0.27, p < .01. No interaction with condition was found, ft = -0.19, p = .30; Step
2: AR2 = 0.02, p = .30. A comparable relationship was found for the effect o f trait PST on free
recall o f neutral items. Just as with state PST, free recall o f neutral items decreased as trait PST
increased,/? = -0.39,p = .01; Step 1: R 2 = 0.18,/? < .05. Similarly, a trend emerged such that free
• • • • 7
recall o f emotional items also decreased as trait PST increased, ft = -0.27, p = .06; Step 1: R =
When examining performance on the cued recall task, a significant condition by trait PST
interaction was found, ft = -0.41,/? < .05; Step 2: AR2 = 0.10,/? < .05. In the threat condition,
cued recall performance marginally decreased as trait PST increased (ft = -0.43,/? = 0.06). In the
no threat condition, cued recall performance remained stable as trait PST increased (ft - 0.13, p >
A significant condition by trait PST interaction was also found for the dependent variable
change in positive affect, ft = -0.56,/? < .01; Step 2: AR2 = 0.18,/? < .01. As shown in Figure 3,
those in the threat condition experienced a decrease in positive affect from pre to post stereotype
threat manipulation as trait PST increased (ft = -0.58,/? < .01). In the no threat condition,
There were no other significant effects o f trait PST, all Fs < 2.25.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
27
Stereotype Activation
restricted range o f responses on this item. Out o f the 20 possible bias/prejudice or old age
stereotype word completion items, participants only completed between the range o f 0 and 4
items (M = 1.40, SD = 1.13). Counter to our hypothesis, our stereotype threat manipulation had
no effect on stereotype activation, t(40) = 0.25, p > .80, d = 0.08. However, given the extremely
restricted range o f responses, it is not surprising that we did not find a main effect o f condition.
activation interaction for the dependent variable change in negative affect, ft - 0.55, p - .01; Step
2: AR2 = 0.15,/? = .01. As illustrated in Figure 4, those in the threat condition experienced an
increase in negative affect from pre to post stereotype threat manipulation as stereotype
activation increased (J3 = 0.60,/? < .01). In the no threat condition, negative affect remained
No other significant results were found in the analyses on stereotype activation, all Fs <
2.35.
Memory Identification
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
28
Discussion
threat effect in older adults. Firstly, we demonstrated that older adults under stereotype threat
performed more poorly on a test o f free recall performance than those not under threat,
replicating the results o f other researchers in this area (e.g., Hess et al., 2003). However, this
effect did not transcend into tasks o f cued and recognition recall performance, perhaps because
these tasks did not represent an adequate level o f difficulty for our participants. Indeed, other
researchers have shown that stereotype threat disrupted women’s math performance on difficult,
but not easy, tests o f mathematic ability (Spencer, Steele, & Quinn, 1999).
In order to more specifically examine this stereotype threat effect on free recall
performance, we were interested in whether memory for emotional and neutral information
might be affected differently by stereotype threat. Although we initially predicted that memory
for emotional information might be maintained under stereotype threat, our results show
otherwise. Here, we found that those in the no threat condition recalled more emotional
information than those in the threat condition, but that these groups did not differ in the amount
o f neutral information recalled. However, free recall o f neutral items was affected by both state
and trait perceived stereotype threat, such that recall o f neutral items decreased as perception o f
stereotype threat increased. Ideas on how to more powerfully address this hypothesis are
discussed below.
One o f the main goals o f our study was to test potential moderators o f the stereotype
threat effect. Similar to Brown and Pinel (2003), we showed that stigma consciousness
moderated the stereotype threat effect for cued and recognition recall. However, this moderation
o f cued and recognition recall performance was actually in the direction opposite to our
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
29
difficult to determine why this occurred, but it may be the case that those who were more
conscious and expectant o f age-based stigma were also trying the hardest to overcome its effects.
This effort may only have been successful on the relatively easier tasks o f cued and recall
performance, where stereotype threat effects were not as potent. In addition, a more descriptive
and age-relevant measure o f stigma consciousness may be necessary in order to truly tap this
concept in older adults. Because we used a modified version o f the SCQ in our research, we may
not actually be capturing the construct we aim to assess. The development o f a more powerful
SCQ for specific use with older adults is an ongoing focus o f our research.
Another potential moderator o f interest in our study was age group identification, the
degree to which older adults identified with the older adult age group. Our results show that age
group identification was associated with recall performance, such that the more older adults
identified with their age group, the lower their recall performance. To our knowledge, this is the
first time that identification with one’s age group has been shown to moderate stereotype threat
effects.
In this experiment, we were also able to show that both state (situational) and trait
cued recall performance. For both types o f perceived stereotype threat, there was a steeper
decline in cued recall performance with increased perceived stereotype threat in the threat
condition. This effect may only have been seen with cued recall performance because o f the
relative easiness o f this task compared with the free recall task. Again, those who are more
aware o f stereotype threat, may have been putting more effort into overcoming its effects, but
their efforts were only fruitful on the task o f cued recall. Despite this discrepancy, these results
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
30
are still important in that they point to the fact that both situational and chronic perceptions o f
stereotype threat have a negative effect on older adults’ cognitive task performance. The idea o f
chronic perceived stereotype threat is particularly interesting in that it suggests the existence o f
especially stereotype threat prone individuals, the very individuals who we would try to target
Our last independent variables o f interest, subjective age identification and memory
identification, failed to reveal any significant results. This may be due to the fact that these
measures were presented after the threat manipulation in order to best structure the experimental
session, or perhaps because these scales were not the best measures o f these constructs. In future
research, we m ay adjust the timing o f presentation o f these measures and use different scales to
Another main goal o f this study was to determine the specific emotional effects o f
stereotype threat on older adults. Our results surprisingly showed that those in the threat
However, we also showed that those in the threat condition experienced a decrease in positive
affect as trait (dispositional) perceived stereotype threat increased and an increase in negative
affect as stereotype activation increased. These results, although conflicting, are the first to show
a change in well-being for older adults under stereotype threat. The results for trait perceived
stereotype threat and stereotype activation make intuitive sense. It seems reasonable that one
would feel less happy and more upset in a threat situation if they are dispositionally more aware
o f and affected by stereotype threat and have more negative stereotypes activated in the situation.
However, the results for age group identification are less straightforward in their interpretation.
If one is more identified with their age group, why would they feel less upset in a threat
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
31
situation? Perhaps this reflects a sort o f resignation or acceptance, such that those who are
highly identified as an older adult have accepted their situation, and are less bothered by negative
stereotypes or threatening situations. Another explanation might provide more support for the
socioemotional selectivity theory in that it may be the case that those older adults who are old
both in body and in mind are those most attuned to positively valenced information and are
therefore attending to only the positive aspects, even in a threatening situation. This would
explain why these older adults are less upset in a threatening situation than those less identified
with the older adult age group and therefore perhaps less biased toward positive information. In
the future, we might ask older adults to report what aspects o f the situation contributed to their
ratings o f positive and negative emotions. It may be the case that highly identified older adults
remember more positive aspects o f the testing situation itself, and are therefore less likely to
have experienced an increase in negative emotions. It also may be the case that highly identified
older adult have a positivity bias even in their own emotions, such that their attention is focused
solely on their positive emotions, and these are the emotions they are more likely to report.
The results o f this study support the general finding that stereotype threat disrupts
performance o f stereotyped individuals on stereotype relevant tasks. This finding runs counter to
a recent study by O ’Brien and Hummert (2006), which finds evidence for self-stereotyping
(Levy, 1996), but not for stereotype threat. Self-stereotyping occurs when certain cues activate
group stereotypes and then lead to behaviours consistent with that stereotype. In this particular
study, middle-aged adults (48-62 years old) were told that their memory performance would be
compared with older adults over 70 years old or young adults under 25 years old. Stereotype
threat research suggests that those in the young adult comparison group should perform more
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
32
poorly than those in the older adult comparison group, whereas research on self-stereotyping
suggests the opposite. The results o f this study followed the self-stereotyping prediction, such
that middle-aged adults in the older adult comparison group actually performed worse than those
in the young adult comparison group. However, our study differs from this study in that our
participants were much older than the late middle-aged adults who participated in the O ’Brien
and Hummert study, and were therefore more likely to find the effects o f stereotype threat more
potent. In our future work, we might include middle-aged adults in order to see how their
The idea o f “transitioning” into the older adult age group presents interesting potential for
future work in this area. For example, it is likely the case that only older adults who recognize
and accept their older adult status would be affected by such phenomena as priming, self-
stereotyping, and stereotype threat. Indeed, Levy (2003) posited that two prerequisites need to
be met in order for age stereotypes to have an effect on older adults. First, older adults must
objectively join the old age membership group. This usually entails reaching a certain age
arbitrarily determined by one’s society as constituting status as a senior. The second, and
theoretically more important step, involves actual identification with group members. This
requires individuals to actively think about themselves as an older adult. These two prerequisites
rarely occur at the same time; instead, the second prerequisite is usually only met years after the
first. Levy asserts that only those who identify with the old age group will be affected by age
stereotypes and was able to show that older, but not younger, adults were affected by negative
old age primes (Levy, 2006). However, other researchers (Bargh, Chen, & Burrows, 1996)
disagree with this assertion and have shown that young adults can also be primed to conform to
old age stereotypes. These findings raise interesting questions for work on old age stereotypes in
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
33
general. Is it really the case that “you are only as old as you feel”, or does this subjective
“feeling”, or identification, really play any role? Further research with middle-aged adults may
shed some light on this issue as well as on issues underlying age group identification. For
example, what factors lead some middle-aged adults to transition quickly and easily into the old
age group while others resist and continue identifying with a younger age group almost
indefinitely? Are these factors biological, psychological, or both? Longitudinal studies on age
group identification would also help to uncover the nature o f age group identification over the
lifespan. For example, is it the case that those who are strongly identified with their age group at
age 25 are also strongly identified at 75? Perhaps, the opposite is true. In either case, there
remains much to be learned about the influence o f age group identification on the processes o f
One o f our initial hypotheses was that memory for emotional information might be
maintained under stereotype threat, however, our results did not support this hypothesis. In order
to better test this idea in the future, we likely need to further specify our hypothesis. Much o f the
work that follows from the socioemotional selectivity theory shows not only a bias for emotion-
related information, but a specific bias for positively valenced emotional information (e.g.
Carstensen & Mikels, 2005; Fung & Carstensen, 2003; Mather & Carstensen, 2005). The theory
behind these findings is that as we age or near endpoints in any segment o f our lives, our
attention becomes focused on positive information. Further research will attempt to discern
between memory for emotional information with positive versus negative valence in order to see
whether only memory for positive information is maintained under stereotype threat. The
passage used in this experiment did not contain enough clearly positive or clearly negative
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
34
In future work, it would also be interesting to see if our age group identification results
are moderated by an individual’s beliefs about older adults. For example, someone who is
highly identified with the older adult age group and has a positive view o f older adults would
probably be less affected by stereotype threat than an individual who is highly identified with the
older adult age group but views the group negatively. This finding also highlights the potential
o f in-group affirmation to reduce the effects o f stereotype threat. Research with gender
stereotypes (McIntyre, Paulson, & Lord, 2003) has shown that making positive group
achievements salient helps to alleviate the negative effect o f stereotype threat on women’s math
performance. A similar intervention with older adults may also help to inoculate this group
The results o f this research have significant implications for the treatment o f older adults
within our society. Because negative aging stereotypes and, therefore, stereotype threat, are so
prevalent in our society (Kite et al., 2005), it is important to identify those who are most
vulnerable to the deleterious effects o f these phenomena. Once these individuals have been
identified, intervention techniques can be implemented to combat these effects. By ensuring the
health o f our growing older adult population, we protect the healthcare and social systems o f our
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
35
References
Allport, G.W. (1954). Traits Due to Victimization. In G.W. Allport, The Nature o f Prejudice (pp.
Aronson, J., & Inzlicht, M. (2004). The ups and downs o f attributional ambiguity: Stereotype
Aronson, J., Lustina, M.J., Good, C., Keough, K., Steele, C.M., & Brown, J. (1999). When
White men can’t do math: Necessary and sufficient factors in stereotype threat. Journal
Brown, R.P., & Josephs, R.A. (2000). The importance o f unimportance: The mathematics
Austin.
Brown, R.P., & Pinel, E.C. (2003). Stigma on my mind: Individual differences in the experience
Butler, R.N. (1975). Why Survive? Being Old in America. New York: Harper and Row.
Butler, R.N. (1980). Ageism: a foreword. Journal o f Social Issues, 36, 8-11.
Carstensen, L.L. (1987). Age-related changes in social activity. In L.L. Carstensen & B.A.
Edelstein (Eds.), Handbook o f clinical gerontology (pp. 227-237). New York: Pergamon
Press.
Carstensen, L.L. (1991). Selectivity theory: Social activity in life-span context. In K.W. Schaie
(Ed.), Annual review o f gerontology and geriatrics, 11, 195-217. New York: Springer.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
36
Carstensen, L.L. (1993). Motivation for social contact across the life span: A theory o f
Nebraska Press.
Carstensen, L.L., & Mikels, J.A. (2005). At the intersection o f emotion and cognition: Aging and
Carstensen, L.L., & Turk-Charles, S. (1994). The salience o f emotion across the adult life span.
Charles, S.T., Reynolds, C.A., & Gatz, M. (2001). Age-related differences and change in positive
and negative affect over 23 years, Journal o f Personality and Social Psychology, 80,
136-151.
Chasteen, A.L. (2000). The role o f age and age-related attitudes in perceptions o f elderly
Chasteen, A.L., Bhattacharyya, S., Horhota, M., Tam, R., & Hasher, L. (2005). How feelings o f
Cohen, E.S. (1990). The elderly mystique: Impediment to advocacy and empowerment.
Cohen, E.S. & Kruschwitz, A. (1990). Old age in America represented in 19th and 20th century
Coupland, M. & Coupland, J. (1993). Discourses o f ageism and anti-ageism. Journal o f Aging
Studies, 7, 279-301.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
37
Covey, H. (1988). Historical terminology used to represent older people, Gerontologist, 28, 291-
297.
Craik, F.I.M. (1986). A functional account o f age differences in memory. In F. Kliz & H.
Crockett, W.H. & Hummert, M.L. (1987). Perceptions o f aging and the elderly, Annual Review
Cuddy, A.J.C., Norton, M.L, & Fiske, S.T. (2005). This old stereotype: The pervasiveness and
Dillon, K. & Jones, R. (1981). Attitudes toward aging portrayed by birthday cards. International
Erber, J.T., Szuchman, L.T., & Rothberg, S.T. (1990). Age, gender, and individual differences in
Fiske, S.T., Xu, J., Cuddy, A.C., & Glick, P. (1999). (Dis)respecting versus (dis)liking: Status
Fung, H.H., & Carstensen, L.L. (2003). Sending memorable messages to the old: Age
Garstka, T.A., Branscombe, N.R., & Hummert, M.L. (1997, May). Age group identification
across the lifespan. Paper presented at the annual meeting o f the American Psychological
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
38
Gattuso, S., & Shadbolt, A. (2002). Attitudes toward aging among Pacific Islander health
Goldsmith, R.E., & Heiens, R.A. (1992). Subjective age: A test o f five hypotheses. The
Guo, X, Erber, J.T., & Szuchman, L.T. (1999). Age and forgetfulness: Can stereotypes be
Harris, D.K. (2005). Self-fulfilling prophecy. In E.B. Palmore, L. Branch, & D.K. Harris (Eds.),
Harwood, J., Giles, H., McCann, R.M., Cai, D., Somera, L.P., Ng, S.H., Gallois, C., & Noels,
K. (2001). Older adults’ trait ratings o f three-age groups around the Pacific rim. Journal
Hausdorff, J.M., Levy, B.R., & Wei, J.Y. (1999). The power o f ageism o f physical function o f
Hess, T.M., Auman, C., Colcombe, S.J., & Rahhal, T.A. (2003). The impact o f stereotype threat
Hummert, M.L., Garstka, T.A., Shaner, J.L., & Strahm, S. (1994). Stereotypes o f the elderly held
Inzlicht, M., & Ben-Zeev, T. (2000). A threatening intellectual environment: W hy females are
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
39
Jaccard, J., Wan, C.K., & Turrisi, R. (1990). The detection and interpretation o f interaction
Kite, M.E., & Johnson, B.T. (1988). Attitudes toward older and younger adults: A meta-analysis.
Kite, M.E., Stockdale, G.D., Whitley, B.E., & Johnson, B.T. (2005). Attitudes toward younger
and older adults: An updated meta-analytic review. Journal o f Social Issues, 61, 241-266.
Levy, B. (1996). Improving memory in old age through implicit self-stereotyping. Journal o f
Levy, B.R. (2003). Mind matters: Cognitive and physical effects o f aging self-stereotypes.
Levy, B.R., & Langer, E. (1994). Aging free from negative stereotypes: Successful memory in
China and among the American Deaf. Journal o f Personality and Social Psychology, 66,
935-943.
Levy, B.R., Hausdorff, J., Hencke, R., & Wei, J. (2000). Reducing cardiovascular stress with
1-9.
Mather, M., & Carstensen, L.L. (2005). Aging and motivated cognition: The positivity effect in
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
40
McIntyre, R. B., Paulson, R. M., & Lord, C. G. (2003). Alleviating women's mathematics
Mendoza-Denton, R., Downey, G., Purdie, V.J., Davis, A., & Pietrzak, J. (2002). Sensitivity to
Mroczek, D.K. (2001). Age and emotion in adulthood. Current Directions in Psychological
Nelson, T.D. (2002). Ageism: Stereotyping and prejudice against older persons. Cambridge,
Nosek, B.A., Banaji, M.R., & Greenwald, A.G. (2002). Math = male, me = female, therefore
O ’Brien, L.T., & Hummert, M.L. (2006). Memory performance o f late middle-aged adults:
Packer, D J . & Chasteen, A.L. (2006). Looking to the future: How possible aged selves influence
Palmore, E. (1971). Attitudes toward aging as shown by humor. The Gerontologist, 11, 181-186.
Pasupathi, M., & Lockenhoff, C. (2002). Ageist behavior. In T. D. Nelson (Ed.), Ageism:
Stereotyping and prejudice against older persons, 201-246. Cambridge, MA: MIT
Press.
Pinel, E.C. (1999). Stigma consciousness: The psychological legacy o f social stereotypes.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
41
Pronin, E., Steele, C.M., & Ross, L. (2004). Identity bifurcation in response to stereotype threat:
Rahhal, T. A., Hasher, L., & Colcombe, S. J. (2001). Instructional manipulations and age
differences in memory: Now you see them; now you don't. Psychology and Aging, 16,
697-706.
Robinson, J., & Skill, T. (1995). The invisible generation: Portrayals o f the elderly on prime-time
Ryan, E.B., Hamilton, J.M., & Kwong See, S. (1994). Patronizing the old: How do younger and
older adults respond to baby talk in the nursing home?, International Journal o f Aging
Schmader, T., & Johns, M. (2003). Converging evidence that stereotype threat reduces working
Sharps, Matthew J., Price-Sharps, J.L., & Hanson, J. (1998). Attitudes o f young adults toward
older adults: evidence from the United States and Thailand. Educational Gerontology,
24, 655-660.
Shipley, W.C. (1940). A self-administering scale for measuring intellectual impairment and
Smith, M. (1979). The portrayal o f elders in magazine cartoons. The Gerontologist, 19, 408-412.
Spencer, S.J., Steele, C.M., & Quinn, D.M. (1999). Stereotype threat and women’s math
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
42
Steele, C.M. (1997). A threat in the air: How stereotypes shape intellectual identity and
Steele, C.M, & Aronson, J. (1995). Stereotype threat and the intellectual test performance o f
Stone, J., Lynch, C.I., Sjomeling, M., & Darley, J.M. (1999). Stereotype threat effects on Black
and White athletic performance. Journal o f Personality and Social Psychology, 11, 1213-
1227.
Wheeler, S.C., & Petty, R.E. (2001). The effects o f stereotype activation on behavior: A review
Wilkinson, J.A. & Ferraro, K.F. (2002). Thirty years o f ageism research. In T.D. Nelson (Ed.),
Ageism: Stereotyping and prejudice against older persons, 338-358. Cambridge, MA:
MIT Press.
Yoon, C., Hasher, L., Feinberg, F., Rahhal, T. & Winocur, G. (2000). Cross-cultural differences
in memory performance between younger and older adults: The role o f culture-based
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
43
Figure Captions
Figure 3. Change in positive affect from pre to post experimental manipulation as a function of
trait (dispositional or chronic) perceived stereotype threat by condition. A zero value would
indicate no change in positive affect from pre to post experimental manipulation. Positive values
indicate an increase in positive affect; negative values indicate a decrease in positive affect.
Figure 4. Change in negative affect from pre to post experimental manipulation as a function o f
stereotype activation by condition. A zero value would indicate no change in negative affect
from pre to post experimental manipulation. Positive values indicate an increase in negative
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
44
Figure 1.
12 -
No Threat
10 - Threat
0)
0
d
as
E
1 8 “
a>
o_
as
o
ai 6 —
R Sq Linear = 0.025
■o
a>
4-
2 -
R Sq Linear = 0.327
1 2 3 4 5
State Perceived Stereotype Threat
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
45
Figure 2.
12 -
No Threat
10- Threat
Cued Recall Performance
8 - o o R Sq Linear = 0.016
C ’o
6-
4-
R Sq Linear = 0.181
2 -
1 2 3 4 5
Trait Perceived Stereotype Threat
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
46
Figure 3.
1.50-
No Threat
1 . 00 -
Threat
0 .5 0 -
CO R Sq Linear = 0.047
• _
(2 o.oo-
c
O)
2 -0 .5 0 -
- 1 . 00 -
R Sq Linear = 0.331
-1 .5 0 -
1 2 3 4 5
Trait Perceived Stereotype Threat
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 4.
1 .5 0 -
No Threat
Threat
1 . 00 -
Change in Negative Affect
R Sq Linear = 0.35!
0 .5 0 -
0 . 00 -
R Sq Linear = 0.058
- 0 .5 0 -
0 1 2 3 4
Completed Stereotype Activation Items
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.