You are on page 1of 81

THE EFFECTS OF REINFORCEMENT AND PUNISHMENT

ON RETARDATE ATTENTION AND

DISCRIMINATION LEARNING

by

JOHN LAV7S0N REEVES II, B.A.

A DISSERTATION

IN

PSYCHOLOGY

Submitted to the Graduate Faculty


of Texas Tech University in
Partial Fulfillment of
the Requirements for
the Degree of
DOCTOR OF PHILOSOPHY

Approved

Accepted

August, 19 7 3
ACKNOWLEDGMENTS

I would like to express my appreciation and gratitude

to the members of my committee, Drs. W. F. Landers, C. G.

Halcomb, D. Chatfield, and J. Bodden, for their guidance,

encouragement, and friendship throughout my graduate train-

ing. I am also deeply indebted to Dr. B. J. Locke for his

help, advice, and tolerance, not only during this project,

but also during the years that I was his graduate advisee.

Special thanks is extended to Mr. Gary Byrd for his un-

selfish assistance throughout all phases of this research

and to Mrs. Beverly Sawyer for her eager help. I would

particularly like to thank my wife Carole for her patience,

support, and long hours of help, without which this project

could not have been completed. Finally, my deepest apprecia-

tion is extended to the late Dr. S. L. Campbell whose dis-

cerning insights greatly influenced those close to him. I

will always remember him.

11
TABLE OF CONTENTS

ACKNOWLEDGMENTS 11

• • •
LIST OF TABLES
111
LIST OF ILLUSTRATIONS
iv
yyi. INTRODUCTION
1
Attention and Mental Retardation . . .
2
\/Reinforcement, Punishment, and Atte_ntion
13_
Statement of the Problem
22
II. METHOD
24
Subjects
24
Instrumentation
25
Procedure
31
\ ^11. RESULTS
38
Learning
38
Motivation
40
Attention . .__:u^ — . . . , . ,
42_
sj I V . DISCUSSION
51
V. SUMMARY AND CONCLUSIONS
J 61
REFERENCES
64
APPENDIX
72

111
LIST OF TABLES

Table Page

1. Analysis of Variance for Trials to Criterion . 40

2. Analysis of Variance for Response Rate . . . . 42

3. Analysis of Variance for Observing Responses . 44

4. Analysis of Variance for Observing Time . . . . 46

5. Analysis of Variance for Observing Time


Following Reinforced and Punished Trials . . 48
6. Analysis of Variance for Percent of
Correct Responses 49

IV
LIST OF ILLUSTRATIONS

Figure Page

1. Floor Plan of the Performance Booth 26

2. Display Panel 27

3. Cowboy Stimulus 28

4. Floor Plan of the Experimental Room 32

5. Mean Trials to Criteria Per Group 39

6. Trial Latency (1/2 Seconds) 43


CHAPTER I

INTRODUCTION

Mental retardation has long occupied a position of

serious social concern. Historically, this concern has

tended to reflect an unmistakably pessimistic attitude.

Early investigators viewed retardates as incurably defec-

tive individuals whose hereditary condition produced

degenerate and immoral life styles (Crafts, 1916; Goddard,

1914; Knollin & Terman, 1918). These social indictments

are dramatically illustrated in the following address

delivered by a leading figure in mental retardation in the

early 1900's:

The social and economic burdens of uncomplicated


feeblemindedness are only too well known. The
feebleminded are a parasitic, predatory class,
never capable of self-support or of managing their
own affairs. The great majority ultimately become
public charges in some form. They cause unutter-
able sorrow at home and are a menace and danger
to the comjuunity. Feebleminded women are almost
invariably immoral and . . . usually become car-
riers of venereal disease, or give birth to chil-
dren v.'ho are as defective as themselves . . . Every
feebleminded person, especially the high-grade
imbecile, is a potential criminal, needing only
the proper environment and opportunity for the
development and expression of his criminal tenden-
cies. (Fernald, 1919, cited in Zigler & Harter, 1969)

Such indictments have had profound impact on "modern"

thinking. Contemporary theorists have tended to reflect

the rather pessimistic outlook which characterized their


heritage. Where earlier workers conceptualized the deficit
as residing in the moral character structure, their contem-
poraries view the problem as inhering in some "specifiable"
cognitive or physical deficiency. Thus, prevailing theo-
retical perspectives have relegated the mental retardate's
condition to a variety of such incurable "traits" as the
relative impermeability of the boundaries between regions
of the cognitive structures (Lewin, 1936); primary and sec-
ondary rigidity resulting from subcortical and cortical
malformations (Goldstein, 194 3); inadequate disinhibitory
mechanisms (Siegel & Foshee, 1960) ; disassociation of the
verbal and motor systems (Luria, 19 63); brevity of the
stimulus trace (Ellis, 1963); inadequate neural satiation
(Spitz, 1963); and impaired attention directing mechanisms
(Zeaman & House, 1963) . Although these theorists differ
greatly among themselves as to the exact nature of the im-
pairment, the idea of an "attention deficit" either directly
or indirectly emerges as a dominant theme in most of their
theories (Crosby & Blatt, 1968).

Attention and Mental Retardation


Clinical and educational opinion have long held impair-
ment of attention to be a general characteristic of mentally
retarded individuals (Crosby & Blatt, 1968). At least four
theorists have accounted for the "subnormal" functioning of
the mentally retarded by positing a specific attention
deficit. Luria (1963), drawing from the Pavlovian tradi-
tion, holds that all "higher nervous activity" results
from varying degrees of equilibrium between excitatory and
inhibitory processes. He further hypothesizes that the
mentally retarded (MRs) exhibit a disturbance of this
equilibrium with excitation predominating. This distur-
bance results in an inability to refrain from responding
producing "distractible" and "impulsive" behavior. Further-
more, this deficiency frequently results in an absence of
the Orientation Reaction (OR) to stimuli of low and medium
intensity. Coupled with a failure to habituate the OR to
strong stimuli, the MR continues to respond to such stimuli
regardless of task relevance and exhibits an apparent defect
in attention. Denny's (19 66) "elicitation theory" empha-
sizes the function of rewards in a trial and error situation,
He hypothesizes that the major deficit in trial and error
learning among retarded subjects can be attributed to limi-
tations in their "duration of attention." O'Conner and
Hermelin (1963) have proposed that the learning difficulties
encountered by MRs result from defective acquisition rather
than perceptual or retention disorders. They further specu-
lated that acquisition is impaired because of the MR's abil-
ity to "focus attention" on task relevant dimensions.
Perhaps the most influential current theory in mental
retardation is the "Attention Theory" proposed by Zeaman
and House (1963). Building on Wyckoff's (1952) attention
model, Zeaman and House (19 63) have proposed that the cor-
rect instrumental response in a discrimination learning
task involves a chain of two responses: (a) attending to
the relevant stimulus dimension and (b) approaching and
responding to the correct cue. They further suggest that
it is in the initial ability to attend to relevant dimen-
sions and cues rather than in the final stage of learning
that children of different mental abilities differ. This
can be seen in "backward learning curves" in which children
of varying mental abilities exhibit the same rate of learn-
ing, once learning begins, but differ with respect to the
number of trials required for the child to produce better
than chance performance. This longer "lead-in" to the
onset of learning among retarded subjects is interpreted
as reflecting an inability to attend (Zigler, 1966).
The contributions of these theories notwithstanding,
their development has been predicated more on theoretical/
methodological ease than empirical, behavioral and physio-
logical evidence. The tendency has been for each theorist
to devise and perseverate on idiosyncratic measurement
tactics which optimize the credibility of their own position
This has resulted in a wide range of mutually exclusive tech-
niques designed to assess attention (Fisher, 1972) ; reaction
time (Ellis, 1963); incidental learning (Denny, 1966); ease
of distraction (Crosby, 1972); impulse control (Spivack,
Levine, & Sprigle, 1959); and backward learning curves
(Zeaman & House, 19 63). Attention has been treated as a
single process conveniently linked to such paradigms, rather
than a complex multibehavioral process (Alabiso, 19 72;
Kantor, 1957).

Although widely employed, it is evident that usage of


the term "attention" has not been entirely consistent. In
a recent survey of the literature on attention deficit in
retarded subjects, Alabiso (1972) found that the term at-
tention has generally been used to refer to three separate
behaviors: span, focus, and selectivity.

Span
Attention has been defined as the length of time a
subject can spend on a given task (Pillsbury, 1908). As
early as 1904, Kuhlman (1904) suggested that short atten-
tion span accounted for the retardate's apparent difficulty
in discrimination learning tasks. More recently, research
efforts by Strauss and Kephart (19 55) and Strauss and
Lehtinen (1947) have added credence to the idea that brain
injured children have difficulty attending to a task for
any length of time. Contemporary literature is also replete
with references indicating deficit attention span behaviors
in retarded populations (Alabiso, 1972; Cromwell et al.,
1963; Crosby, 1972; Crosby & Blatt, 1968; Garton, 1964;
Goldstein & Seigle, 1961; Weber, 1963).
While task variables (Moyer & Von Haller Gilmer, 1955) ,
distracting stimuli (Crosby, 1972), developmental level
(Schrater, 1933) , and physical properties of the stimulus
(Berlyne, 19 50) have been found to affect attention span,
recent studies have reported that span may be more closely
related to the reinforcement contingencies associated with
the particular task (Redd, 1972). Staats, Staats, Schultz,
and Wolf (19 62) found that by reinforcing progressively
longer observing responses (attending behaviors), primary
school Ss V7ere able to increase task perseverance by 150%,
resulting in improved academic performance. Martin and
Powers (19 67) reported similar findings with MRs in a bar-
pressing task. They found that reinforcement applied con-
tingently for attending progressively longer to a bar-
pressing task dramatically increased performance outcome.
In accordance with the previous studies, Hewett, Taylor,
and Artuso (19 69) reported that the systematic application
of token reinforcements produced higher levels of task
attention and arithmetic performance with primary school
children. Such results have been replicated employing normal
and retarded Ss on several other occasions (Brov/n & Foshee,
1971; Kerr, 1962; O'Leary & Becker, 1967; Redd, 1972;
Schmidt & Urlich, 1969), demonstrating the controlling
effects of reinforcement on attention span.
However, span is only one aspect of the multibehavioral
field called attention. Several investigators have concerned
themselves with the process by which a subject comes to ob-
serve a particular aspect of a task, i.e., the focus process
(Alabiso, 1972).

Focus
Wachtel (1967) has likened focus to a beam of light
which is more intense at the middle while becoming weaker
at its periphery. Research seems to suggest that MRs exhibit
deficit focus behaviors. Hagen and Huntsman (1971) employed
an "incidental learning" paradigm in which a child's atten-
tion was directed to a central picture on a stimulus card
which also contained an incidental figure. After removal
of the ten stimulus cards, the Ss were required to identify
both the central and incidental pictures. Institutionalized
MRS correctly identified significantly fewer central aspects
and a significantly greater number of incidental aspects
than normal Ss. These results support the hypothesis that
MRS are unable to focus their attention on the relevant
aspects of a task and ignore the irrelevant. Jenson (19 63)
compared acquisition rates of mentally retarded, "average,"
and "above average" subjects on learning tasks requiring
8
"concentrated visual focus." He reported that acquisition
rates covaried significantly with intellectual development,
further support for the thesis that MRs exhibited deficit
focus behaviors.

Some researchers have attempted to abandon inferential


assessments in favor of procedures allowing a more objective
measure of focusing behavior. These investigators have
chosen to quantify attention (focus of attention) by measur-
ing overt observing responses (ORs). White and Plum (19 64)
employed eye photography to map eye-movements of children
while they learned a visual discrimination. They found
that the amount of stimulus scanning was positively related
to ease of learning a discrimination. In another experiment,
Wright and Dahler (19 66) reported that grade school chil-
dren "shaped" to emit observing responses (ORs) during pre-
training sessions subsequently exhibited significantly better
oddity problem performance than Ss deprived of OR training.
They accounted for their results by suggesting that the ORs
produced by pretraining increased the probability that the
Ss would focus their attention on relevant stimuli. Wright
and Smothergill (19 67) provided a more precise analysis of
this hypothesis. They devised a two-choice discrimination
task that required the Ss to operate a lever in order to
bring either of two discriminative stimuli into focus. The
number of lever movements constituted the OR data. Groups
receiving reinforcement contingent on observing responses
not only emitted more ORs but demonstrated a marked increase
in discrimination performance when compared to controls not
receiving reinforcement for ORs. Their interpretation of
these results was consistent with the hypothesis proposed
by Wright and Dahler (1966). O'Donnell (1969) utilized a
two-choice discrimination learning paradigm with normal and
retarded Ss. The discrimination apparatus consisted of a
box with two spring operated doors that concealed the dis-
criminative stimuli. Observation of the discriminative
stimuli required the Ss to open one of the doors, thus
defining an OR. Half of the Ss received "OR training."
During the OR training sessions, Ss received candy rein-
forcement for opening the doors in the presence of a neutral
stimulus. Ss reinforced for emitting ORs required signifi-
cantly fewer trials to criteria than those not receiving OR
training. Further, normal Ss engaged in more OR behavior
and performed the discrimination at significantly more ef-
ficient rates than retarded Ss. He concluded that the
superior performance of the normal Ss resulted from the fact
that they had acquired ORs to a "higher level" than the MRs.
He further posited that the poorer OR performance of the
MRS provided support for an "attention-deficit" hypothesis
of retardate discrimination learning. Thus, the discrimina-
tion learning research indicates that: 1) MRs appear to
10

exhibit deficit focusing behaviors (ORs) ; 2) ORs can be

induced by and seem to be a function of the contingent

application of reinforcement; and 3) this process results

in increased performance.

Attention requires not only remaining on-task and


observing the stimuli, but the ability to systematically
differentiate between stimuli and choose the correct dimen-
sion. This process is defined as selectivity (Alabiso,
1972) .

Selective Attention

Selective attention has been conceptualized as a "two-

link" process (Alabiso, 1972; Crosby & Blatt, 1968; Zeaman

& House, 1963) . The first link requires the Ss to select

the correct stimulus (e.g., form). The second link involves

approaching the correct dimension (e.g., circle). Acquisi-

tion of the two-link chain requires the Ss to eliminate

certain competing response patterns in selection. Harlow

(1959) referred to these modes of selectivity as "error-

factors." Harlow (19 59) described four principle error-

factors which might occur regardless of their correctness

(i.e., whether they are reinforced or not): 1) Stimulus

Perseveration refers to the tendency to repeat incorrect

choices on subsequent trials of the same problem; 2) Posi-

tion Preference is defined as the tendency to consistently


11
respond to the left or right position in a discrimination
task; 3) Response Shift is defined as the tendency to try
out or explore both stimulus objects regardless of correct-
ness; and 4) Differential Cue Errors refer to the relative
frequency of errors on successive trials in which the cor-
rect stimulus object changes position from the previous
trial when compared to errors on successive trials in which
the stimulus remains in the same position. Differential Cue
Errors are attributed to the operation of interacting re-
sponse tendencies when there is ambiguity between the object
rewarded and the position rewarded on any particular trial
of a nonspatial discrimination.
When the error-factors have been reduced or eliminated,
more appropriate or contingent modes of selectivity can be
adopted to solve a discrimination problem. For example, a
contingent solution to a two-choice simultaneous size dis-
crimination task requires that if an S selects the correct
dimension (i.e., the larger stimulus) on a given trial, he
should stay with that object on the next trial. If his re-
sponse was incorrect (i.e., choice of the smaller stimulus),
he should shift to the larger object of the pair (Harter,
Brown, & Zigler, 1971). Thus, the S is "attending" to the
consequences of his behavior and selecting dimensions accord-
ingly, thereby learning the discrimination more rapidly.
MRS have been described as exhibiting particularly deficient
12
modes of contingent selectivity. Ellis, Girardeau, and
Pryer (1962) reported that position preferences were most
characteristic of retardate performance and could account
for most of their errors in a two-choice object-quality
discrimination problem. In a similar experiment, Kaufman
and Peterson (1958) found that retarded Ss exhibited a sig-
nificantly greater percentage of stimulus-perseveration
errors than did normal controls. They concluded that stimu-
lus perseveration errors are generally characteristic of the
slower learning approach of MRs but not normals. In another
research effort, Harter, Brown, and Zigler (1971) found that
retarded subjects generally utilized significantly more
position preference strategies than contingent modes of
selection in a size-discrimination problem. Normals tended
to respond more contingently and learned the task more
rapidly.
One can only speculate as to "why" MRs tend to engage
in position and stimulus perseveration errors. Some investi-
gators have speculated that such selection preferences have
developed through the S's past experience or reinforcement
history (Gerjuoy & Winters, 1968; Kantor, 1959). Researchers
have demonstrated that, by differentially manipulating the
schedules of reinforcement associated with the various
strategies, older retardates will tend to adopt response
strategies on the basis of their relative probability of
reinforcement (Gerjuoy & Gerjuoy, 19 64, 1965; Gerjuoy &
Winters, 1968; Schusterman, 1964; Weir, 1964). Thus, the
selectivity literature supports the hypothesis that: 1)
retardates exhibit deficit habits of attention and 2) selec-
tivity, like span and focus, is a partial function of the
contingent application of reinforcement.

To date, researchers have relied heavily on the system-


atic application of reinforcement. Reinforcement has been
generally viewed as the most powerful tool available to the
practitioner for managing span, focus, and selectivity as
well as subsequent performance of normal and retarded chil-
dren (Martin & Powers, 1967). The educational implications
of this position for mental retardation are obviously pro-
found. If the generally inferior learning performance
exhibited by MRs reflects deficit habits of attention rather
than learning ability (Zeaman & House, 1963), then an appro-
priate remediation tactic is generally to be found in the
systematic application of reinforcement.

Reinforcement, Punishment, and Attention


y^
There is, indeed, a general consensus affirming rein-
forcement manipulations as the instrument of choice in
promoting enhanced attention and subsequent learning. How-
ever, there is less agreement whether such results are best
achieved via positive reinforcement, punishment, or
14
reinforcement-punishment combinations. Recent work in com-
plex discrimination learning tends to support the contention
that best results are achieved when a punishment component
is included in the motivational procedures.
A large number of studies employing children as subjects
have concerned themselves with the question of the relative
effectiveness of reinforcement and punishment in the acqui-
sition of discrimination learning tasks. In the prototypic
experiment, the Ss are required to choose between two or
more stimulus alternatives over a series of trials (White-
hurst, 19 69) . One group receives jpositive reinforcement
contingent upon choosing the "correct" alternative and no
feedback (referred to as a "blank" trial) following an in-
correct response (condition R). Another group receives a
punishing stimulus contingent upon choosing alternative(s)
designated as incorrect and receives a blank trial following
the correct response (condition P). A third group of sub-
jects receives both a positive reinforcement for a correct
response and a punishing stimulus for the incorrect alterna-
tive (condition R-P). All groups of Ss are required to
achieve an arbitrary performance criterion and then the
efficiency of acquisition is compared (Whitehurst, 1969) .
These experiments have generally afforded results indi-
cating that punishment, either alone or in combination with
reinforcement, leads to better acquisition performance than
15
does reinforcement alone (Marshall, 19 65; Massey & Insalaco,
1969; Spence, 1972; Spence & Seger, 1967; Whitehurst, 1969).
While the literature affirms the consistency of these find-
ings, the process by which punishment serves to enhance
acquisition performance is still at issue.
Earlier explanations tended to be motivational in na-
ture (Brackbill & O'Hare, 19 58; Buss & Buss, 19 56; Curry,
19 60). These theoretical positions suggested that the
effect of punishment was to act as an additional source of
drive which enhanced learning. Such conjectures, however,
have been recently subplanted by theories emphasizing at-
tentional mechanisms. Penny (1967) employed a simultaneous
somatosensory discrimination learning task to study the
effects of punishment on focus of attention. Kindergarten
children were divided into there groups (R-P, P, and R) and
required to reach through a semicircular hole in front of
the apparatus in order to gain access to the stimulus ob-
jects. Each stimulus consisted of 7.62 cm. wooden objects
differentiated on the basis of form, i.e., a circle and a
triangle. Ss were instructed to touch the blocks as many
times as they wished before making a response choice. The
frequency witli which stimulus objects were examined consti-
tuted the measure of focus (OR). Response choice consisted
of lifting the block selected. Under the correct block, Ss
in the R and R-P groups received a M&M candy, while Ss on
16
the P group received a "blank" trial. If the block desig-
nated as incorrect was chosen, Ss in the P and R-P group
received a 98 db tone punishment and Ss in the R group
received a blank trial. The experiment was terminated when
the S attained a 9 out of 10 consecutive correct trials
criterion. Analysis of the data revealed that group P re-
quired the fewest trials to criterion followed by group R-P
and R, respectively. Group P emitted a significantly greater
number of observing responses than the other groups while
there were no significant differences between the remaining
groups. For all groups, the larger number of ORs covaried
with faster learning. Penny hypothesized that punishment
decellerated response rate and increased the number of ORs
emitted, resulting in better attention to the stimulus. On
the other hand, reinforcement tended to increase response
rate and decrease stimulus scanning, resulting in poor at-
tention to the relevant dimension.
Penny's (19 67) basic findings were replicated in an
experiment conducted by Witte and Grossman (1971). They,
too, employed a somatosensory discrimination learning para-
digm that allowed quantification of the Ss' observing
responses. Although their procedures were virtually iden-
tical to Penny's, Witte and Grossman did not find perfor-
mance differences between the groups receiving a tone
punisher (R-P and P). The punishment groups did, however.
17
exhibit a significantly greater number of correct responses
than the R group which received token reinforcements when
correct and a blank trial when incorrect. The punished
groups tended to engage in more OR behaviors than the R
group, a finding consistent with Penny's results. In order
to test the hypothesis that punishment served to accentuate
subject motivation, the stimuli were presented on scale plat-
forms which allowed for measurement of the force with which
Ss made their choices. They noted that others (Holton, 19 61)
have employed this method to measure the possible motivating
effects of frustration. Witte and Grossman obtained no sig-
nificant force differences suggesting that the differences
in discrimination performance between the groups were due to
attentional rather than motivational variations. To the ex-
tent that the OR is an adequate measure of focus of atten-
tion, these studies suggest that the punished Ss were more
"attentive" and, as a result, exhibited better performance.
Analogous findings have been reported regarding selective
attention (Harter, Brown, & Zigler, 1971; Meyer & Offenbach,
1962; Stevenson, Weir, & Zigler, 1959).

Stevenson, Weir, and Zigler (1959) presented preschool


children with a three-choice discrimination problem under
two reinforcement conditions: group R-P received a token
for correct responses and had to relinquish a token for in-
correct responses, and group R received the tokens for
18
correct responses and a blank-trial following incorrect re-
sponses. They found that group R-P exhibited a signifi-
cantly greater number of correct responses than group R.
The authors hypothesized that the penalty for incorrect
responses had a facilitating effect on performance by
increasing Ss' attention to relevant cues. Although they
failed to actually quantify selection behaviors, their ob-
servations suggested that the S's greatest difficulty lay
in their failure to rely on the stimuli as a basis for their
responses. Rather, the Ss tended to engage in error-factor
modes of selectivity. The penalty for incorrect responses
was assumed to have increased S's tendency to focus on
stimulus attributes (adopt contingent modes of selection)
before making a response.
In another experiment concerning selectivity, Meyer
and Offenbach (19 62) used primary school children as Ss to
examine the effects of verbal reinforcement and punishment
on the performance of a complex discrimination task. Rein-
forcement consisted of the word "Right" and punishment, the
word "Wrong." Ss were divided into three treatment condi-
tions: "Right"--Blank (R) , "Wrong"—Blank (P) , and "Right-
Wrong" (R-P). The discriminative stimuli consisted of
geometric blocks that could be varied over a variety of
dimensions, i.e., color, shape, and size. Only one of these
dimensions was designated as relevant (correct) while the
19
others were irrelevant. Their results indicated that while
all groups exhibited substantial learning, the two groups
receiving punishment were significantly better than the
reinforcement only group (R). Their interpretation of
these results focused on the cue or informational proper-
ties of reward as opposed to punishment. For example,
Meyer and Offenbach might have simultaneously presented
the S with a large red square stimulus and a small green
circle stimulus. They suggested that it seemed reasonable
to assume that Ss approached such a problem with a specific
response set. The S might adopt the response set that the
correct or relevant dimension was size while the other
dimensions (color and shape) were irrelevant to the task.
Acquisition for the punished groups occurred as response
sets to the irrelevant dimensions were eliminated, resulting
in closer attention to the remaining dimensions. Conversely,
acquisition for group R occurred when a response to a par-
ticular dimension was confirmed. However, they noted that
all dimensions, relevant and irrelevant, were present on
each stimulus presentation, resulting in a high probability
that an incorrect response set would also be confirmed and
strengthened. Therefore, reinforcement served to distract
the Ss from attending to the relevant dimension by rewarding
responses to irrelevant dimensions.
20
Such interpretations seem founded on somewhat tenuous
bases (Meyer & Offenbach, 1962; Stevenson et al., 1959).
The nature of their experimental paradigms did not allow
for the empirical assessment of the Ss' selecting behaviors.
Rather, their conclusion that reinforcement and punishment
exhibit differential effects on selectivity relied mostly
on logical speculation rather than empirical inference.
In an attempt to promote an empirically-based assess-
ment of these speculations, Harter, Brown, and Zigler (19 71)
employed a two-choice size discrimination task which allowed
for the quantification of selection strategies. Normal and
retarded Ss were either reinforced for correct responses
(received a marble token) and punished for incorrect choices
by relinquishing a marble (group R-P), or reinforced for a
correct response and given a blank trial for incorrect re-
sponses (group R). Unfortunately, a group P was not in-
cluded. The groups were given 50 acquisition trials and
the number of correct responses tabulated. Their procedure
allowed them to use a modified version of Levine's (1963)
method for categorizing selection responses into contingent
patterns (object contingent hypothesis and position con-
tingent hypothesis) and noncontingent patterns (object
preferences and position preferences). Analysis of their
data revealed that normals performed the discrimination
better than retardates. The data further indicated that
21
normal and familial retarded Ss showed better learning than
organically retarded Ss in the R-P condition compared to
the R condition. The strategy data indicated that organi-
cally retarded Ss tended to utilize more position preference
behaviors while both normal and familial retarded Ss showed
more contingent strategy behavior in the R-P condition com-
pared to the R condition. While the organic S tended to use
less position selection strategies in the R-P condition
compared to the R condition, they did not significantly
increase their use of contingent modes of selection in the
former condition as did the familials and normals. They
interpreted these results in accordance with Stevenson
et al. (1959) and Meyer and Offenbach's (1962) contention
that the effect of including punishment in the reinforcement
procedures was to "heighten S's attention" to response con-
tingent modes of responding.
The research reviewed suggests that punishment acts
to improve discrimination performance by increasing atten-
tion to the stimulus (a large number of focusing ORs) and/or
eliminating irrelevant responses V7hile accelerating more
contingent patterns of selectivity. While focus and selec-
tivity have been investigated, researchers have failed to
study the effects of punishment on attention span. With the
tendency for researchers to treat attention as a single
process linked to their paradigms, studies integrating span,
focus, and selectivity have been lacking. In his recent
survey of the literature, Alabiso (1972) failed to uncover
a single study considering the integration of these atten-
tional processes. It would seem that further understanding
of the molar determinants of attention, particularly regard-
ing the effects of punishment, necessitates an integrative
approach, since "focus presupposes span while selectivity
requires the preceding two" (Alabiso, 19 72).
In summary, the above research indicates that the in-
ferior discrimination performance exhibited by MRs has been
attributed to deficit patterns of attention such as span,
focus, and selectivity. However, discrimination learning
among both retarded and normal subjects can be enhanced
through the use of punishment. Research suggests that pun-
ishment increases discrimination learning by altering Ss'
focus and selection behaviors rather than affecting their
motivation. The effects of punishment on span has not been
investigated.

\/ statement of the Problem


The present experiment proposed to study the effects
of three combinations of reinforcement and punishment (R-P,
P, and R) on retardate attention and discrimination learning
A two-choice discrimination task which allowed simultaneous
trial-by-trial measurement of span, focus and, to an extent.
selectivity,was employed in order to provide an integrated
assessment of attention. Moreover, the addition of a re-
sponse rate measure allowed the analysis of the possible
additional motivating effects of punishment. The research
hypotheses tested were:
1. That groups receiving punishment would learn the
discrim.ination faster;
2. That punishment and reinforcement would differen-
tially affect measures of span, focus, and selec-
tivity; and
I
\ 3. That the increased performance of the punished
groups would be related to enhanced span, focus,
and selectivity but not motivation.
CHAPTER II

METHOD

Subjects

Subjects were 16 male and 11 female moderately retarded

students from the Lubbock State School for the mentally re-

tarded, Lubbock, Texas. Only Ss residing in dormitories

designated specifically for "moderately retarded" students

were chosen for the present research. Individuals exhibit-

ing gross motor and perceptual disturbances were excluded.

All Ss had previously engaged in research totally unrelated

to the present study. The Ss were divided into three sub-

groups of nine, each subgroup being assigned to a different

reinforcement combination: 1) group R-P whose Ss received

a reinforcement for correct responses and a punishment for

incorrect responses, 2) group R received a reinforcement

for correct responses and a blank trial for incorrect re-

sponses, and 3) group P received punishment for incorrect

responses and a blank trial for correct responses. The

subgroups were matched in means and variance on chronologi-

cal age, sex, and IQ (as entered in the school's records

and based on the Wechler Adult Intelligence Scale). The

overall mean IQ was 66.5 and mean age was 18.9 years. Ap-

pendix A provides a detailed summary of the Ss' individual

statistics per treatment condition.

24
25
Instrumentation
Experimental Chamber

The experimental chamber measured 1.32 meters wide by

1.32 meters in length and stood 1.82 meters high (Figure 1 ) .

The interior of the chamber was lined with acoustical tiling

and the floor carpeted. The chamber was a double walled ply-

wood structure containing sound dampening material between

the walls. A one-way glass mirror, 60.9 centimeters by 91.4

centimeters, to the S's left permitted the E to observe the

S throughout the experimental session. Lighting was provided

by a fluorescent light mounted on the ceiling of the experi-

mental chamber. White dampening noise was delivered through

a 30 centimeter speaker placed on the floor of the chamber.

Display Panel
The display panel is illustrated in Figure 2. Mounted

91.4 centimeters from the floor, it was constructed of a

60.9 by 60.9 by .63 centimeter plywood sheet painted dark

blue. The display panel sat at a 45 degree angle directly

in front of the S.

The Stimulus
The stimulus, a cowboy, was drawn on a translucent

sheet of smoked mylar with black India ink (Figure 3) . The

cowboy stimulus \;as divided into two distinct body sections

of equivalent area, the trunk and the legs. The two body
26

One-way Window

Indirect Lighting

Fig. 1.—Floor plan of the performance booth.


27

>- 3
o
m 3
^ ^
O
STl

(1)
o
a
CM

>i

w
•H
Q
I
I

rcj

•H
fin

c
E
W
"c
3
Q.
28

Body
dimension
Stimulus
Lights

W^^=:*!miKajB«aw*^ ll

^ N
Legs /
\
dimension
' \ \ R )( G ) I Y

/
kv\
.-^^
. 9 ^wy^.-Tfasut^tymlmt ^^U5,^acn^»fcs:v=fii^|i»^
L.
Fig. 3.—Cowboy Stimulus.
29
sections were centered over a pair of 28 volt panel lights
(General Electric number 1819), one red and one green,
positioned 1.27 centimeters below the surface. When en-
gaged, the green or red light would illuminate its respective
body section. On any trial, one of four possible combina-
tions of lights was randomly presented with the restriction
that the same light would not appear more than twice in a
row. For example, a red light could appear under the trunk
of the cowboy and green light under his legs, or a green
light under both his trunk and legs.

Observing Response
Two double-pole relay micro-switches served as observing
buttons. They were placed in a column 27.9 centimeters from
the left edge of the display panel and 6.35 centimeters
apart, in line with one pair of stimulus panel lights. In
order to observe or activate the stimulus panel lights of a
particular body section, the S was required to press the
corresponding observation button. The observation buttons
also operated a BRS electronic timer (Model TI-9 03) which
timed the duration S observed the stimulus panel lights of
each body section on every trial. The observation times
were recorded on a Grayson-Stadler model E12505A print
out counter. In addition, each press of the observation
buttons operated a corresponding BRS model C-P/901 panel
30
counter, which allowed a trial-by-trial assessment of the
number of observing responses for each body section.

Response Function

Response micro-switches were placed on the left side


of the display panel (Figure 2). A "happy face" and a "sad
face" decal (7.62 centimeters in diameter) were positioned
below the left and right response buttons, respectively.
Ss were instructed to engage the "happy" micro-switch when
they thought the "happy" stimulus panel light was shining
or to press the "sad" micro-switch when they thought the
"sad" stimulus panel light was on. A response resulted in
the appropriate reinforcement contingency and advanced the
program to the next trial, containing a new array of stimu-
lus lights. The stimulus combinations were programmed on
standard BRS electromechanical programming equipment.

Reinforcement
Reinforcement was delivered in the form of points which
could be exchanged for prizes. The reinforcement delivery
unit consisted of a four digit numerical counter positioned
five centimeters above the response buttons with an amber
signal light mounted directly above it. For groups receiv-
ing reinforcement, a correct response advanced the counter
by one and simultaneously illuminated the amber signal light
for one second. Those groups receiving punishment for
31

incorrect responses received a 110 db, 3000 cps (HZ) tone

for one second. The punishing stimulus was produced by a

MAICO Precision Hearing Test instrument and delivered

through an Ampex Model 414 speaker located approximately

60.9 6 centimeters from the S's left ear. The one second

reinforcement and punishment periods also served as an inter-

trial interval and all response buttons were inoperative

during this interval.

Procedure

Ss were individually scheduled for a single testing

session. The Ss were escorted into the experimental labor-

atory which housed the testing chamber, relay programming

equipment, and an attractive display of prizes (Figure 4 ) .

The prizes included giant suckers, shaving lotion, makeup,

candy, gum, and jewelry. The "grand prize," a Snoopy-Doll,

was awarded to the subject achieving the best score at the

termination of the research project. After exhibiting the

prize table, the E informally presented the following in-

structions :
In the little room I have a light game you can play
if you want to [E pointed at the experimental cham-
ber]. I will show you how to play in just a minute.
If you do well, you can have your choice of any
prize on this table. The one who plays the game
the best will win this "Snoopy-Doll." Even if you
do not win the "Snoopy," you can still win something
else if you try real hard. Do you see anything here
that you want?
32

Prize Table

Performance Booth

Programming Equipment

Fig. 4.—Floor plan of the experimental room.


33
When the S chose an item from the table, E praised him or
her for a good choice and emphasized the desirability of
the selected item. The E then continued the instructions:

If you would like to win the prize, you will have


to stay for a little longer and play the light game.
If you do not want to play, you do not have to.
Would you like to play the light game?

If the S responded negatively, he was instructed to return


to the cottage or to his work station. If the S's answer
was affirmative, he was directed into the experimental cham-
ber with on-going 25 db white noise for masking purposes.
The following instructions were then presented by E in a
casual and informal manner:

Please listen to me carefully. We are going to


play a game in which you will have to figure out
what makes the cowboy happy [E pointed to the happy
face] or what makes the cowboy sad [E pointed to
the sad face]. What are you going to have to
figure out in this game?

If the S failed to respond or responded incorrectly (i.e.,

did not say "what makes the cowboy happy or sad"), E re-

peated the above instructions and continued:

Now look at these buttons [E pointed at the observ-


ing buttons]. Press them and see what happens.
When you press the top button, a colored light ap-
pears under the cowboy's body [E demonstrated].
When you press the second button, colored lights
appear under his legs [E pressed the second button].
Try it again. Sometimes during this game the
lights will change from red to green under the dif-
ferent parts of the cowboy. Watch them carefully
[E then demonstrated how the lights changed colors].
Now show me what you do in order to see what color
is under his body, and under his legs. Now both
parts of the cowboy at once.
34
If the S was having difficulty, the E repeated the instruc-
tions, guiding the S's hands to the appropriate observing
button.

In order to minimize confounding variance associated

with response strategies totally unrelated to the solution

of the experiment, Hovland (19 52) has suggested informing

each S about the types of dimensions and specifying the

different values of each dimension. In accordance with

Hovland's suggestion, the following instructions proceeded:

Now comes the hard part of the game, where you have
to figure out what makes the cowboy happy and what
makes him sad. As you can see, the cowboy is wear-
ing a mask and we cannot see if he is smiling or
frowning. So here is how you can see if he is happy
or not, even when he is wearing a mask. First you
have to see what colors are under the different
parts of the cowboy by pushing these buttons [E
pointed at the observing buttons]. If you look
carefully at the colors under the cowboy's body and
under his legs, one of them will tell you if he is
happy and another one if he is sad [E pointed at
happy and sad response buttons]. If you think one
of these lights is the happy light, press this but-
ton [E pointed at the happy face response button].
If you think that the sad light is shining, press
the sad face button. Remember,the color of the
lights change from time to time, watch them care-
fully. This is the hard part of the game and you
have to work very carefully to figure out which
light makes the cowboy happy and which makes him
sad. Remember, only one of these lights is the
happy light and only one of these lights is the sad
light.

All Ss received the preceding instructions. In addi-

tion, Ss received the following instructions which varied

in accordance with the treatment condition to which they


35

were assigned. Instructions for the R-P condition were as


follows:

If you make the right choice, you will get a point


and this light will come on [E pointed at the counter
and amber reinforcement light, then demonstrated
its function]. If you are wrong, however, you do
not get any points, but you will hear this loud
noise [E pointed to the speaker and operated the
tone]. If you make a lot of points, you will get
the prize.

Ss in group P received the following instructions:


If you are wrong, you will hear this loud noise [E
pointed at speaker and operated the tone]. When
you press the buttons and nothing happens, that
does not mean that you were right or wrong [E
pointed out the response buttons]. Remember, try
to keep the tone from coming on. If you can keep
the noise from coming on very much, you will get
the prize.

After the instructions were presented, E remained in

the experimental chamber for five trials, coaxing the Ss

and answering any questions that arose. At the termination

of the fifth trial, the E said:


Good, now I am going to leave you for a while. Do
as good as you can and you will win the prize you
chose on the table. Remember, the person who does
the best will win the "Snoopy." I will be right
outside if you need me.

The experimental session was terminated after a cri-

terion run of 9 out of 10 correct responses was achieved,

or 250 trials without a criterion run. The E then praised

the S for his "outstanding" performance and gave him the

prize.
36

The E arbitrarily designated the legs of the cowboy


stimulus to be the salient dimension. A green value dis-
played beneath the legs was designated as a "happy" re-
sponse whereas a red value under the legs was a "sad"
response. Correct responses were contingent on both observ-
ing the salient stimulus value and responding appropriately
to it. For example, S was required to press the "happy"
response button when a green light appeared in the legs
region, or press the "sad" button when the red light ap-
peared in the legs. An incorrect response resulted when
the S engaged either response button without making any
observing response or failed to observe the values of the
salient dimension (the legs). An incorrect response was
also designated when the S observed the values of the
salient dimension and emitted an inappropriate response
to it.
The procedure for any given subject involved the follow-
ing sequential components on each trial. Subsequent to
response consequation of the preceding trial, S depressed
one or both observing buttons on an ad lib basis, then
depressed one of the two response keys, i.e., the "happy"
or "sad" button. At this point, any scheduled response
consequences were administered for one second and a new
trial begun. Trials were continued until criterion was met
at which point S was interviewed to assess his ability to
37
verbalize the stimulus properties correlated with reinforce-
ment (see Appendix B ) . The experimental regimen was com-
pleted at this point for the participating subject.
CHAPTER III

RESULTS

The analyses performed are presented under three major


headings: Learning, Motivation, and Attention. Unless
specified under the various headings, all analyses performed
attained an adequate power of 0 = .75 or better (Cohen,
1971). The figures and tables that follow summarize the
analyses.

^ Learning
Trials to Criterion

The number of trials to criterion were arbitrarily

designated as 9 out of 10 consecutive correct responses.

A randomized block analysis of variance with three treat-

ment conditions was used to analyze these data (Kirk, 1968).

The treatment conditions were R-P, R, and P. In order to

approach homogeneity of variance, a reciprocal transforma-

tion was performed on the individual scores. The results

of this analysis are recorded in Table 1.

Table 1 shov.'s the F ratios for the treatment condi-

tions, none of which achieved significance. Although not

significant, inspection of Figure 5 does indicate differen-

tial performance between the experimental groups. Group

R-P and group P clearly required less trials to reach

38
39

160
cc
Lli
t 150
CC
o
o 140
1-

mJ
130
<
cc
h- 120

H
R P P R
TREATMENT GROUP

Fig. 5.—Mean trials to criteria per group


40
criterion than group R (mean TTC = 130.556, 138.778, and
155.778, respectively).

TABLE 1
ANALYSIS OF VARIANCE FOR TRIALS TO CRITERION

Source df F

Between Treatments 2 2.757


Between Blocks 8 2.40 6
Residual 16

A power analysis (Cohen, 1971) yielded an adequate but


rather low power (0 = .65), suggesting that a larger sample
of subjects might have afforded significant results.
In order to control for possible error variance asso-
ciated with the slightly heterogeneous Full Scale Intelli-
gence scores between the groups, an analysis of covariance
was applied to trials to criterion data (Kirk, 19 68). The
analysis revealed that the covariate measure. Full Scale
Intelligence scores, did not significantly contribute to the'
total variance (F = 1.208 df 2/15 P<.05).

Motivation

Response Rate Following Reinforced


and Punished Trials
Response rates or response latencies were computed by
electromechanically recording the duration of time elapsing
41
between the termination of reinforcement or punishment, and
a response by the S in half seconds units. , The response
rate measure encompassed both the observing responses and
response time for each S. A Log-10 transformation was
applied to the response rate scores following reinforced
trials and those scores follov;ing punished trials. The trans
formed data were analyzed using a split plot factorial (SPF
3.2; Kirk, 19 68) with three treatment levels and two levels
of response rates (response rates following reinforced
trials and response rate following punished trials).
Table 2 illustrates the results of the analysis. In-
terestingly, overall differences in group response rates
failed to emerge (mean trial response rates in one-half
seconds were: R-P ^ .911, R = .907, and P = 1.036). Fur-
thermore, the analysis revealed that mean response rates
following reinforcement and punishment were equal. Response
rate for trials following reinforcement was .971 seconds,
and trials following punishment was .9 32 seconds. Even
though Figure 6 seems to show a much slower response rate
for group P than for the other groups, the analysis revealed
a non-significant group x trials interaction, indicating no
performance differences.
42

TABLE 2
ANALYSIS OF VARIANCE FOR RESPONSE RATE

Source SS df MS

Treatment (A) 0356 0178 .6811


Subj. W. Groups 6272 24 .0261
Reinf. or Punish. Trials (B) 0083 .0083 1.9253
AB 0038 .0019 .4424
B X Subj. W. Groups 1032 24 .0043
Total 7781 53

Attention
Focus of Attention
Observing Response

The observing response (OR) was a quantification of the

number of times the Ss viewed the cowboy stimulus. The OR

was obtained by electromechanically counting the nun^er of

observation button depressions for the respective body and

leg dimensions that were emitted by the Ss per trial.

The mean number of observing responses per trial for

eacli of the two dimensions of the cowboy stimulus was com-

puted for each subject. A split plot factorial analysis of

variance (SPF 3.2) with three treatment conditions (R-P, R,

and P) and two repeated measures (ORs for leg dimension and
43

>• 1.05
O P
LLI 1.00 "
R

.9 5 R P
R P
.90 -

.85 -
O
O
.00 V

r^

1
REINFORCEMENT PUNISHMENT
TREATMENT CONDITION

Fig. 6.—Group latencies on trials follov/-


ing reinforcement and trials follov/ing
punishment.
44

ORs for body dimension) was applied to the data (Kirk,


19 68) . A summary of the source table is presented in
Table 3.

TABLE 3
ANALYSIS OF VARIANCE FOR OBSERVING RESPONSES

Source SS df MS

Treatment (A) 3.5593 1.7796 .4991


Subj. W. Groups 85.5695 24 3.5654
Dimension (B) 3.6029 3.6029 8.1919*

A X B .6223 .3112 .7075


B X Subj. W. Groups 10.5556 24 -4398

Total 103.9096 53

*Signifleant beyond the .01 level.

The analysis yielded a non-significant treatments main

effect indicating no differences in the average number of

observing responses between the treatment conditions. The

mean OR per trial for groups R-P, R, and P was .84, .77,

and .73, respectively. Highly significant differences,

however, were found favoring a larger number of ORs toward

the salient leg dimension over the non-salient body dimen-

sion of the cowboy stimulus (leg = .90 and body = .84 ORs).

The non-significant treatment x dimension interaction was


45

further evidence of equivalent OR performance between the


experimental groups, with all groups orienting more towards
the salient dimension.

(/Attention Span
Observing Time

The observation time data provided information regard-

ing the average duration of salient and non-salient dimen-

sion observations per trial. The measure was obtained by

recording the duration of time the Ss depressed each observa-

tion button per trial in tenths of a second. A Log-10

transformation was applied to these data in order to ap-

proximate homogeneity of variance. A split plot factorial

analysis of variance with three treatment conditions and

two repeated measures (salient and non-salient dimensions)

was used to analyze the transformed data (SPF 3.2; Kirk,

1968) .

As depicted in Table 4, the experimental manipulations

did not significantly alter the mean observation time of

the Ss (group R-P = .559; R = .765, and P = .923). The non-

significant treatment x dimension interaction further evi-

denced similar group observing time performance for the

salient and non-salient dimensions. Although approaching

significance, a reliable overall dimensions effect failed

to emerge for the observing time between salient and


46

non-salient dimensions (x = .820 and x = .640 seconds, re-

spectively) .

TABLE 4

ANALYSIS OF VARIANCE FOR OBSERVING TIME

Source SS df MS

Treatment (A) .4737 2 2369 1.2883

Subj. W. Groups 4.4125 24 1839

Dimension (B) .0322 1 0322 4.1945

A X B .0019 2 0009 .1201

B X Subj. W. Groups .1842 24 0077

Total 5.1045 53

Observation Time Following Reinforced


d aiid Punished Trials
In order to extend the examination of the effects of

reinforcement and punishment on the observation time of the

Ss, the following procedures were utilized. For each S the

mean observation time on the trials following a reinforcement

were computed as were the mean observation times following

punished trials. These data were analyzed using a split plot

factorial (SPF 3.2, 2; Kirk, 1968) with three between treat-

ment levels, two within levels of trials following reinforce-

ment and trials following punishment and two within levels

of salient and non-salient dimensions.


47
Inspection of Table 5 revealed no overall treatment
differences in observing time. The analysis of observing
time following the reinforced trials also failed to reach
significance. These results indicate that the "reinforcers"
and "punishers" utilized in this experiment did not succeed
in producing differential effects. Further evidence of the
equivalent effects of the reinforcement and punishment pro-
cedures on observing time are demonstrated by the non-
significant treatment x reinforcement interaction seen in
Table 5. In accordance with previous analysis, a non-
significant overall dimension effect was obtained. This
finding suggests that the Ss tended to observe the salient
dimension no longer than the non-salient one. As can be
seen in Table 5, the remaining interactions also failed to
attain significance.

Selective Attention
Response Strategies
Unlike Harter, Brown, and Zigler's (1971) study, the
design of the present experiment did not allow for a precise
analysis of contingent and noncontingent response strategies
However, inferences regarding strategies at a molar level
were accessible.
The percentage of successes following a successful
trial and the percentage of successes following an unsuccess
ful trial were computed for each treatment condition.
48

TABLE 5
ANALYSIS OF VARIANCE FOR OBSERVING TIME FOLLOWING
REINFORCED AND PUNISHED TRIALS

Source SS df MS

Treatment (A) 1.2628 2 .6314 1.3965

Error 10.8516 24 .4522

Reinf. or Punish. Trials (B) .0025 1 .0025 .2201


AB .0059 2 .0029 .2636

Error .2666 24 .0111


Dimension (C) .0502 1 .0502 3.6909
j^C .0001 2 .0001 .0047

Error .3263 24 .0136

BV .0001 1 .0001 .0021

ABC .0197 2 .0099 .4692

Ej-ror .5045 24 .0210


Total 13.2902 107 .1242

Quantification of the data in this manner allowed one to

infer whether reinforcement or punishment differentially

affected selection strategies.

The reciprocal of the percent correct response scores

were then subjected to an analysis of variance (SPF 3.2;

Kirk, 1968). Groups R-P, R, and P were the three between

measures and percent of successful responses following


49

reinforcement and percent of successful responses following

punishment, the two repeated measures. A summary of the

analysis is illustrated in Table 6.

TABLE 6
ANALYSIS OF VARIANCE FOR PERCENT OF CORRECT RESPONSES

Source SS df MS

Treatment (A) 6.38 2 3.19 4.642*


Subj. W. Groups 16.51 24 .69
Reinf. and Punish. (B) .03 1 .03 .038

A X B .52 2 .26 .322


B X Subj. W. Groups 19.48 24 .81

*Signifleant P<.05.

The analysis disclosed a significant treatment main

effects (F = 4.646, df = 2/24, P<.05) indicating that the

experimental groups significantly differed regarding total

number of successful responses (strategies). A Tukey's HSD

multiple comparison further revealed that the significant

differences existed only between group R-P (x = 60% correct)

and group P (x = 46% correct) with group R (x = 53% correct)

falling between them and, although approaching significance,

not reliably different from either. Overall differences

also failed to emerge between percent of successful responses


50

following reinforced trials and percent successes following


punished trials (52% and 54%, respectively).
CHAPTER IV

DISCUSSION

This research was primarily concerned with an analysis


of punishment effects on learning among retarded Ss in a
two-choice visual discrimination task. More specifically,
the study attempted to address the potential facilitating
effects exerted by supplementing or replacing conventional
reinforcing consequences with punishment. While learning
rate was the primary focus, additional concern v/as directed
toward the motivational and attention elements or deter-
minants of such learning. Unlike previous research, the
paradigm selected provided an integrated analysis of atten-
tion through the simultaneous measurement of span, focus,
and selectivity. The following discussion is an attempt
to relate the present findings to the literature and pro-
pose future research.
Analysis of the trials to criterion data revealed that
groups R-P and P failed to perform the discrimination sig-
nificantly better than group R, though trends favoring the
punished groups were observed (see Figure 5). However, the
rather low power obtained by this analysis suggests that
the statistically unreliable results may belie a relatively
credible trend. Given a greater number of replications than

r 1
52
the present subject pool allowed, the power analysis indi-
cates that reliable effects would have emerged with addi-
tional subjects. Therefore, these data are taken as partial
support for the studies reviewed in Chapter I, which found
superior trials to criterion performance for groups R-P and
P over R.

The motivational theories of Brackbill and O'Hare


(1958), Buss and Buss (1956), and Curry (1960) hypothesized
that punishment enhances discrimination performance by acting
as an additional source of drive. However, a motivational
interpretation of the TTC trends favoring groups R-P and P
would seem unlikely since the present analysis failed to
uncover overall response rate differences. Inspection of
Figure 6 would, in fact, indicate that group P was "less
motivated" when compared to other groups, a finding disso-
nant with the motivational theories. The present response
rate findings are in agreement with those of Witte and
Grossman (1971) . They reported insignificant differences
in motivation between punished and non-punished groups as
measured by response force. These data are also consistent
with the response rate measures obtained by Hamilton (19 69)
in a simple marble dropping task. While the response rate
scores provided little support for a motivational interpre-
tation, the E's observations did indicate that Ss receiving
the tone punishment were much more "concerned" and
53
"emotional" than those Ss who received a blank trial for
incorrect responses in group R.
Penny (19 67) and Witte and Grossman (1971) provided an
alternative "attentional" interpretation to the motivational
theories. They proposed that punishment enhanced perfor-
mance by increasing the number of observing responses (ORs)
emitted, resulting in a more "attentive" S. The analysis
of the focus of attention data in the present study also
failed to support the "attentional" interpretation of the
effects of punishment. No reliable group differences were
found with regard to focusing behavior (ORs).
Although researchers have not studied the effects of
punishment on span, the present findings suggest that punish-
ment, either alone or in combination with reinforcement,
failed to significantly increase Ss' observation time.
Therefore, like focus, the lack of group differences pre-
clude involing variations in attention span as an explana-
tion of the performance trends observed. However, it should
be noted that Ss under all treatment conditions tended to
focus more and attend longer to the relevant dimension of
the task than to the irrelevant.
As previously discussed, the selective attention re-
search suggests that punishment acts to enhance performance
in two ways: 1) by eliminating incorrect selection responses
or response sets (Meyer & Offenbach, 1962), and 2) by acting
54

as an additional source of information "heightening" S's


attention to more contingent or correct modes of selection
(Harter et al., 1971; Stevenson et al., 19 59). The selec-
tivity findings in the present research provided partial
support for this two-part hypothesis. These data must,
however, be interpreted with extreme caution. The paradigm
did not allow a precise measurement of contingent and non-
contingent strategies. The selection data provided evidence
concerning only the number of correct responses emitted on
trials following reinforcement and trials following punish-
ment. Thus, only inferential distinctions regarding con-
tingent and non-contingent modes of selection were feasible.
With this caution in mind, the analysis revealed significant
overall group main effects, indicating that group R-P engaged
in a reliably greater number of contingent (correct) re-
sponses than group P. Group R fell midway between group R-P
and P and closely approached being significantly different
from both. The present findings were comparable with those
of Harter et al. (1971), who also reported a non-significant
trend in which organically retarded Ss engaged in more re-
sponse contingent modes of selection in group R-P than in
group R. While Harter failed to include a group P in her
research, the present experiment provided evidence suggesting
that punishment alone greatly diminished the number of cor-
rect selection responses when compared to group R-P and R.
55
These results are somewhat surprising considering that
group P's performance on trials to criterion was superior
to group R and comparable to group R-P.

One explanation of these unexpected findings may lie


in the contrived nature of a forced choice paradigm. It is
suggested that the increase in nonreinforced responding to
the correct stimulus dimensions that occurred in group P was
merely a function of the fact that Ss were forced to choose
on each trial. Following an analysis by Azrin and Holtz
(1966), Whitehurst (1969) proposed that increased responding
in the presence of the correct stimulus in group P does not
necessarily indicate that learning of the desired discrim-
ination has occurred. Rather, "the elimination of a re-
sponse by punishment is not known to result in an increase
of unpunished responses unless those responses are concur-
rently under the control of some reinforcement procedure"
(Azrin & Holtz, 1966, p. 431). In group P, a decrease in
punished responding necessarily resulted in an increase in
"correct" responding.
Post session interviews with the Ss provided additional
support for this formulation. This evidence is summarized
in Appendix B. The inverview data indicates that all of the
Ss in group R-P (n = 9) and group R (n = 9) were able to
correctly identify the "happy" and "sad" light stimuli fol-
lowing the termination of their session. However, only
56

four Ss in group P identified the stimuli correctly. These


findings yielded a highly reliable Chi Square (X^ = 12.3,
df_ = 2, P<.01). The results suggested that most of the Ss
in group P had, in fact, not learned the "correct" selec-
tions but, rather, had learned which selections to avoid.
Such an account is consistent with fewer number of correct
selection responses emitted by this group. Therefore, these
data indicate that groups R-P and R learned the correct
selections better than group P. Thus, rather than "heighten-
ing" S's attention to more contingent modes of selection,
punishment appears to act only to eliminate incorrect selec-
tion strategies.
Meyer and Offenbach (1962) suggested that the exclusion
of punishment in group R deprives the S of a means of elimi-
nating incorrect selection responses, resulting in inferior
trials to criterion performance. Acquisition for group R
occurs as a particular selection is confirmed. However, it
must be remembered that relevant and irrelevant dimensions
are present on every trial. When an S makes a rewardable
response in this situation, it is highly probable that an
incorrect response set will be confirmed and, therefore,
strengthened. Eventually, learning occurs as the S responds
to a different dimension in an effort to maximize reward.
The differences between group R-P and group R seem to reflect
the effects of punishment on selectivity. Punishment acts
57

as an additional cue informing S that a particular selection


mode should be eliminated.

Clearly, the superior trials to criterion performance


and selectivity findings for group R-P as well as the lack
of correct selection responses in group P concur with White-
hurst's (1969) and Azrin and Holtz's (1966) formulation that
"Punishment is not a method of teaching new behavior; it is
a method of eliminating behavior." If punishment is to be
used effectively in discrimination learning, the "correct"
stimulus dimensions must be concurrently under the control
of a reinforcement procedure (Whitehurst, 19 69).
As with any experimental effort, the heuristic impact
of these findings poses a greater range of issues for future
work than v/ere resolved by the completed study. One impli-
cation of considerable import relates to the precision with
which future discrimination paradigms address attentional
phenomena. Past approaches have generally failed to incor-
porate measures affording a precise analysis of the selection
strategies adopted by the experimental subject. In this
writer's view, the delineation of contingent and noncontin-
gent modes of selectivity and the error factor forms presented
constitute essential prerequisites to any comprehensive ac-
count of the interaction between reinforcement/punishment
combinations and discrimination learning processes. As re-
viewed in Chapter I, such a prototypic paradigm has been
58
provided in the Harter, Brown, and Zigler study (1971).
That method entailed a trial-by-trial analysis of ongoing
response strategies and their susceptibility to modifica-
tion via reinforcing and punishing consequences. Inclusion
of such measures in future work are strongly recommended
to afford access to salient determinants of effective dis-
crimination performance.
The experimental tradition followed by the present
research is one in which relatively modest behavioral conse-
quences liave been utilized and assumed to hold reinforcing
or punishing properties for the experimental subjects at
issue. Further, the tacit assumption has been that the
motivational value of reinforcing stimuli was no more marked
than the suppressive motivational value of punishing stimuli.
Such a priori assumptions are no substitute for empirical
delineation of the nature and extent to which response conse-
quences liold motivational value for the subjects under in-
vestigation. Accordingly, future work should incorporate
empirically-based measures of motivational stimulus com-
ponents .
One class of potentially culpable elements in the
present work's failure to obtain positive findings may be
present in context and difficulty level of the task employed.
Presentation of the experimental task was couclied in a rather
ponderous, game-like analogy and Ss were required to make
59
multiple responses (ORs, "Happy" or "Sad" response choices)
to multiple stimulus dimensions (three binary valued aspects
of color, body form, and position). With a host of similarly
idiosyncratic task formats and a degree of variability in
task complexity present in the accumulated literature, some
greater degree of task standardization would seem appropriate
in future work to facilitate cross comparison of different
studies.
Closer attention to subject attributes seems warranted
in view of the Harter, Brown, and Zigler (1971) finding
that differential discrimination outcomes were associated
with subjects differing in the etiological base of their
retarded condition. Thus, while limited subject records
precluded such an analysis in the present study, the etio-
logical classification of mentally retarded subjects should
either be controlled or systematically selected for compari-
son in future work. The inclusion of normal control subjects
does not, however, seem essential in future studies. Rather,
this writer would concur with Estes' (1971) judgement that
the focus should now be placed on acceleration of retardate
performance rather than identification of his divergence
from norms for unimpaired subjects.

Finally and most importantly, the limited range of


attentional features incorporated in studies of attention
should be expanded to include the full range of relevant
60

processes. Alabiso's (1972) review noted the complete


absence of studies simultaneously considering the promotion
of increased span, focus, and selectivity with retarded Ss.
In the present study, these several aspects of attention
tended to vary independently of one another. Since each
constitutes a salient element of the discrimination process,
it would seem that any viable technology of attention train-
ing must be predicated upon an integrated analysis of each
component.
CHAPTER V

SUMMARY AND CONCLUSIONS

Retardates have been found to be particularly slow in

the acquisition of complex discriminations. Although para-

metric data are conspicuously lacking, there is a general

consensus affirming the potential facilitating effects

exerted by supplementing or replacing conventional reinforc-

ing consequences with punishment. However, there is less

agreement whether such results are best explained via moti-

vational or attentional processes. The present study was

designed to gain information on the effects of varying

combinations of reinforcement and punishment on retardate

motivation, attention, and discrimination learning.

Moderately retarded Ss, with a mean IQ of 66.5, per-

formed a two-choice discrimination problem and were either

reinforced for correct responses and given no feedback for

incorrect responses (group R ) , punished for incorrect re-

sponses and given no feedback for correct responses (group

P ) , or reinforced for correct responses and punished for

incorrect responses (group R-P).

Analysis of the results yielded the findings described

below:

61
62
1. Non-significant trends indicated that groups ex-
posed to punishment were superior to the unpunished
group in discrimination performance.
2. Motivational accounts of discrimination performance
were unsupported since differential response rates
were not obtained between groups exposed to diver-
gent reinforcement/punishment combinations.
3. Variations in attention span and focus failed to
account for performance trends favoring punished
groups.
4. Paradoxically, punishment alone seemed to enhance
ultimate acquisition of the discrimination while
diminishing the overall percentage of correct
selection responses. Attentional selectivity
measures were actually lower among subjects receiv-
ing only punishment.
5. It appears that the acquisition process for a
punished subject is different from that for subjects
exposed to reinforcement contingencies. Post-
session interviews indicated that the reinforced
subjects acquired both positive and negative re-
sponse discriminations while subjects under punish-
ment alone acquired only the negative discrimina-
tion. Thus, the divergence exhibited in performance
seemed one of contrast between simple avoidance and
concurrent acquisition and avoidance learning.
63
A substantial proportion of contemporary theory in
mental retardation accords central importance to attentional
phenomena. Moreover, increasing emphasis has been given
the prospect of generating a viable technology of attention
training for the mentally retarded. Divergent behavioral
referents have been utilized among different investigators
(span, focus, and selectivity), yet all seem pertinent to
some area of training application. Most studies assign
rather singular qualities to the attentional process empha-
sizing some components to the exclusion of others. The
present work suggests that the several classes of behavioral
referents for attention may vary independently of one an-
other. It would appear then that any substantial applied
technology must be founded upon empirical groundwork
incorporating an integrated analysis of attention as a com-
plex, multidimensional process.
REFERENCES

Alabiso, F. Inhibitory functions of attention in reducing


hyperactive behavior. American Journal of Mental
Deficiency, 1972, 77, 259-282.

Azrin, N. H., & Holtz, W. C. Punishment. In W. K. Honig


(Ed,), Operant behavior: Areas of research and appli-
cation. New York: Appleton, 19 66, 380-447.
Berlyne, D. A. Stimulus intensity and attention in relation
to learning theory. Quarterly Journal of Experimental
Psychology, 1950, 2, 71-75.

Brackbill, Y., & O'Hara, J. The relative effectiveness of


reward and punishment for discrimination learning in
children. Journal of Comparative and Physiological
Psychology, 1958, 51, 747-751.

Brown, L., & Foshee, J. C. Comparative techniques for in-


creasing attending behavior of retarded students.
Education and Training of the Mentally Retarded, 19 71,
6, 4-11.
Bryant, P. E. The transfer of positive and negative learning
by normal and severely subnormal children. British
Journal of Psychology, 19 65, 56, 81-86. (a)
Bryant, P. E. The transfer of sorting concepts by moderately
retarded children. American Journal of Mental Defi-
ciency, 1965, 70, 291-300. (b)
Buss, A. H., & Boss, E. The effect of verbal reinforcement
combinations on conceptual learning. Journal of Experi-
mental Psychology, 1956, 52, 283-287.
Cohen, J. Statistical power analysis for the behavioral
sciences. Nev/ York: Academic Press, 19 59.

Crafts, L. W. Bibliography of feeblemindedness in relation


to juvenile delinquency. Journal of Juvenile Research,
1916, 1, 195-208.
Cromwell, R. L., Baumeister, A., & Hawkins, W. F. Research
in activity level. In N. R. Ellis (Ed.), Handbook of
mental deficiency. New York: McGraw-Hill, 19 63.

64
65
Crosby, K. G. Attention and distractibility in mentally
retarded and intellectually average children. American
Journal of Mental Deficiency, 1972, 77, 46-53.
Crosby, K. G., & Blatt, B. I. Attention and mental retarda-
tion. Journal of Education, 1968, 150, 67-81.
Curry, C. The effects of verbal reinforcement combination
on learning in children. Journal of Experimental Psy-
chology, 1960, 59, 434.
Denny, M. R. A theoretical analysis and its application to
training the retarded. In N. R. Ellis (Ed.), Inter-
national review of research in mental retardation.
Vol. 2. New York: Academic Press, 19 66, 1-27.
Ellis, N. R. The stimulus trace and behavioral inadequacy.
In N. R, Ellis (Ed.), Handbook of mental deficiency.
New York: McGraw-Hill, 19 63, 134-158.
Ellis, N. R., Girardeau, F. L., & Pryer, M. W. Analysis
of learning sets in normal and severely defective
humans. Journal of Comparative and Physiological
Psychology, 1962, 55, 860-865.
Estes, W. K. Learning theory and mental development. New
York: Academic Press, 19 70.
Fernald, W. E. After-care study of the patients discharged
from Waverly for a period of twenty-five years. Un-
graded, 1919, 5, 25-31.
Ferritor, D. E., Buckholdt, D., Hamblin, R. L., & Smith, L.
The noneffects of contingent reinforcement for attend-
ing behavior on work accomplished. Journal of Applied
Behavior Analysis, 1972, 1, 7-18.
Fisher, L. Attention deficit in brain damaged children.
American Journal of Mental Deficiency, 1970, 74, 502-
508.
Goddard, H. H. Feeblemindedness, its causes and consequences.
New York: MacMillah, 1914.
Goldstein, H., & Seigle, D. M. Characteristics of educable
mentally handicapped children. In J. H. Rothstein (Ed.),
Mental retardation. New York: Holt, Rinehart and
Winston, 1961, 204-230.
66
Goldstein, K. Concerning rigidity. Character and Personal-
ity, 1942-1943, 11, 209-226.

Garton, M. D. Teaching the educable mentally retarded.


Springfield, 111.: Charles C. Thomas, 1964.

Gerjuoy, H., & Gerjuoy, I. R. Choice-sequence patterns in


binary-choice "learning" by retardates. American
Journal of Mental Deficiency, 1964, 69, 425-431.

Gerjuoy, I. R,, & Gerjuoy, H. Binary-choice sequences of


retardates, normal children and college students under
random--and pattern--set instructions. ?imcrican
Journal of Mental Deficiency, 1965, 69, 854-859.
Gerjuoy, I. R., & Winters, J. J. Development of lateral
and choice-sequence preferences. In N. R. Ellis (Ed.),
Handbook of mental deficiency. New York: McGraw-Hill,
1968, 31-63.

Hagen, J. W., & Huntsman, N. J. Selective attention in


mental retardates. Developmental Psychology, 19 71, 5,
151-160.

Harlow, H. F. Learning set and error factor theory. In


S. Koch (Ed.), Psychology: A study of a science. New
York: McGraw-Hill, 1959, 2, 492-537.

Harter, S., Brown, L., & Zigler, E. The discrimination


learning of normal and retarded children as a function
of penalty conditions and etiology of the retarded.
Child Development, 1971, 42, 517-536.

Hewett, F. M. , Taylor, F. D., & Artuso, A. A. The Santa


Monica Project: Evaluation of an engineered classroom
design with emotionally disturbed children. Exceptional
Children, 1969, 35, 523-529.
Holton, R. B. Amplitude of an instrumental response follow-
ing the cessation of reward. Child Development, 19 61,
32, 107-116.
Hovland, C. I. A. "Communication analysis" of concept learn-
ing. Psychological Review, 1952, 59, 461-472.

Jensen, A. L. Learning ability in retarded average and


gifted children. Merrill Palmer Quarterly, 19 63, 9,
123-140.
67
Kantor, J. R. Interbehavioral psychology. Granville, Ohio:
The Principia Press, Inc., 1959.

Kaufman, M. E., & Peterson, W. M. Acquisition of a learning


set by normal and mentally retarded children. Journal
of Comparative and Physiological Psychology, 19 58, 51,
619-621. ~~

Kerr, N. Applications of behavioristic techniques and field


theoretical concepts in somatopsychology. Unpublished
doctoral dissertation. University of Houston, 1962.

Kirk, R. E. Experimental design: Procedures for the be-


havioral sciences. Belemont, Calif.: Brooks/Cole
Publishing Co., 19 68.

Knollin, H. W., & Terman, L. W. A partial psychological


survey of the prison population of San Quentin, Cali-
fornia. In Surveys in mental deviation. Sacramento,
Calif.: State Printing Office, 1918.
Kuhlman, E. Experimental studies in mental deficiency.
American Journal of Psychology, 1904, 15, 391-446.

Lawlis, F. G., & Chatfield, D. Multivariate approaches for


the behavioral sciences: A brief text. Unpublished
manuscript, Texas Tech University, 19 72.

Levine, M. Mediating processes in humans at the outset of


discrimination learning. Psychological Review, 19 63,
70, 254-276.
Lewin, K. A dynamic theory of personality. New York:
McGraw-Hill, 19 36.
Luria, A. R. The mentally retarded child. Oxford: Perga-
mon, 1963. (a)
Luria, A. R. Psychological studies of mental deficiency
in the Soviet Union. In N. R. Ellis (Ed.), Handbook
of mental deficiency. New York: McGraw-Hill, 19 63,
353-387. (b)
Marshall, H. H. The effect of punishment on children: A
review of the literature and a suggested hypothesis.
Journal of Genetic Psychology, 1965, 106, 23-33.

Martin, G. M., & Powers, R. B. Attention span: An operant


conditioning analysis. Exceptional Children, 19 67, 33,
565-570.
68
Massey, P. S., & Insalaco, C. Aversive stimulation as
applied to discrimination learning in mentally retarded
children. American Journal of Mental Deficiency, 19 69,
74, 269-272: "
Meyer, W. J., & Offenbach, S. I. Effectiveness of reward
and punishment as a function of task complexity.
Journal of Comparative and Physiological Psychology,
1962, 55, 532-534.
Moyer, K. E., & Von Haller Gilmer, B. Attention spans of
children for experimentally designed toys. Journal of
Genetic Psychology, 1955, 87, 187-201.
Nie, N., Bent, D. H., & Hull, C. H. Statistical package
for the social sciences. New York: McGraw-Hill, 1970.
0*Conner, N., & Hermelin, B. Speech and thought in severe
subnormality. New York: Pergamon, 19 63.
O'Donnell, J. P. Observing-response training and subsequent
discrimination performance. Child Development, 1969,
40, 191-199.
O'Leary, K. D., & Becker, W. C. Behavior modification of
an adjustment class: A token reinforcement program.
Exceptional Children, 1967, 33, 637-642.
Penny, R. K. The effect of reward and punishment on chil-
dren's orientation and discrimination learning. Journal
of Experimental Psychology, 1967, 75, 140-142.
Pillsbury, W. C. Attention. New York: MacMillan, 19 08.
Redd, W. H. Attention span and generalization of task-
related stimulus control: Effects of reinforcement
contingencies. Journal of Experimental Child Psychol-
ogy, 1972, 13, 527-539.
Schmidt, G. W., & Ulrich, R. E. Effects of group contingent
events upon classroom noise. Journal of Applied Be-
havior Analysis, 1969, 2, 171-179.
Schrater, H. A. A method for measuring the sustained atten-
tion of preschool children. Journal of Genetic Psy-
chology, 1933, 42, 339-371.
69
Schusterman, R. J. Strategies of normal and mentally re-
tarded children under conditions of uncertain outcome.
American Journal of Mental Deficiency, 1964, 69, 66-75.
Schusterman, R. J. The use of stragegies in two-choice
behavior of children and chimpanzees. Journal of
Comparative and Physiological Psychology, 19 63, 56,
96-100.

Siegel, P. S., & Foshee, J. G. Molar variability in the


mentally defective. Journal of Abnormal and Social
Psychology, 1960, 61, 141-143.

Spence, J. T. Verbal and nonverbal rewards and punishments


in the discrimination learning of children of varying
socioeconomic status. Developmental Psychology, 19 72,
6, 381-384. '~^

Spence, J. T. Verbal discrimination performance as a func-


tion of instructions and verbal reinforcement combina-
tion in normal and retarded children. Child Develop-
ment, 1966, 37, 269-281.

Spence, J. T. Verbal reinforcement combinations and


concept-ido.ntification learning: The role of nonrein-
forcement. Journal of Experimental Psychology, 19 70,
85, 321-329.
Spence, J. T., & Segner, L. L. Verbal versus nonverbal
reinforcement combinations in the discrimination learn-
ing of middle- and lower-class children. Child Devel-
opment , 1967, 38, 29-38.
Spitz, H. H. Field theory in mental deficiency. In N. R.
Ellis (Ed.), Handbook of mental deficiency. New York:
McGraw-Hill, 1963, 11-40.
Spivack, G., Levine, M., & Sprigle, H. Intelligence test
performance and the delay function of the ego. Journal
of Consulting Psychology, 1959, 23, 428-431.

Staats, A. W., Staats, C. K., Schutz, R. E., & Wolf, M. M.


The conditioning of testual responses using "extrinsic"
reinforcers. Journal of tJie Experimental Analysis of
Behavior, 1962, 5, 33-40.
Stevenson, H. W., Weir, M. W., & Zigler, E. F. Discrimina-
tion learning in children as a function of motive-
incentive conditions. Psychological Report, 19 59, 5,
95-98.
70
Strauss, A., & Kephart, N. C. Psychopathology and education
of the brain-injured child, II. New York: Grune &
Stratton, 1955.
Strauss, A,, & Lehtinen, L. B. Psychopathology and educa-
tion of the brain-injured child. New York: Grune &
Stratton, 1947.
Trebasso, T., & Bower, G. H. Attention in learning, theory
and research. New York: John Wiley and Sons, Inc.,
1966.
Wachtel, P. M. Conceptions of broad and narrow attention.
Psychological Bulletin, 1967, 68, 417-429.
Weber, E. W. Mentally retarded children and their education.
Springfield, 111.: Charles C. Thomas, 1963.
Weir, M. W. Developmental changes in problem-solving strate-
gies. Psychological Review, 1964, 71, 473-490.
Weisberg, P. Operant procedures with the retardate: An
overview of laboratory research. In N. R. Ellis (Ed.),
International review of research in mental retardation,
New York: Academic Press, 1972.
White, S. H., & Plum, G. E. Eye movement photography during
children's discrimination learning. Journal of Experi-
mental Child Psychology, 1964, 1, 327-338.
Whitehurst, G. J. Discrimination learning in children as a
function of reinforcement condition, task complexity,
and chronological age. Journal of Experimental Child
Psychology, 1969, 7, 314-325.
Witte, K. L., & Grossman, E. E. The effects of reward and
punishment upon children's attention, motivation and
discrimination learning. Child Development, 1971, 42,
537-542.
Wright, J. C , & Dalher, M. Shaping children's observing
behavior for an oddity problem learning set. Paper
presented at the meeting of the Psychonomics Society,
Saint Louis, October 19 66.
Wright, J. C , & Smothergill, D. Observing behavior and
children's discrimination learning under delayed rein-
forcement . Journal of Experimental Child Psychology,
1967, 5, 430-440.
71

Wyckoff, L. B. The role of the observing response in dis-


crimination learning. Part I. Psychological Review,
1952, 59, 431-432.

Zeaman, D., & House, B. J. The role of attention in retar-


date discrimination learning. In N. R. Ellis (Ed.),
Handbook of mental deficiency. New York: McGraw-Hill,
1963, 159-223.

Zigler, E. Research in personality structure in the retar-


date. In N. R. Ellis (Ed.) , International reviev: of
research in mental retardation. New York: Academic
Press, 1966, 77-108.

Zigler, E. F., & Harter, S. The socialization of the men-


tally retarded. In D. A. Goslin (Ed.), Handbook of
socialization theory and research. New York: Academic
Press, 1969, 1065-1102.
f^^ p p i v j « I . .—*4 m^

APPENDIX

A. Individual Subject's Age, IQ, and Length of Institu-


tionalization
B. Post Experimental Session Interview

w
T-

72
73

APPENDIX A: INDIVIDUAL SUBJECT'S AGE, IQ, AND LENGTH


OF INSTITUTIONALIZATION

IQ (WAIS^ ) Length of
Institutionalization
Subject Sex Age FSIQ PIQ VIQ (Months)

Group R-P
S. W. F 19 70 65 77 48
E. B. M 22 68 72 68 54
J. R. F 18 55 47 66 12
B. J. M 21 79 92 72 67
R. 0. M 30 69 82 63 79
G. R. F 21 58 59 62 115
T. T. M 25 72 b b 159
H. P. F 18 72 77 70 115
K. N. M 20 53 56 54 6

Mean 21.5 66.2 68.7 66.5 42

Group R
R. S. M 15 59 68 58 10
B. G. F 21 76 69 83 34
J. W. M 20 66 65 70 36
A. H. F 18 74 94 60 16
R. 0. M 18 71 73 72 19
B. W. F 17 72 78 71 89
J. F. M 18 65 74 62 34
D. P. F 17 43 51 46 102
J. C. M 17 61 64 63 b

Mean 17.8 65.2 70.6 65 37.7

Group P 60
M. M. F 17 71 69 75
J. H. M 16 75 86 69 12
F. W. M 18 50 58 51 17
S. W. F 16 64 71 63 16
A. S. F 16 62 64 66 9
D. S. M 24 76 102 58 28
V. M. M 17 49 b b 94
74

IQ (WAIS^ ) Length of
Inst itutionalization
Subject Sex Age FSIQ PIQ VIQ (Months)

D. C. M 16 61 b b 77
J. F. M 18 46 51 51 16
Mean 17.5 68.3 71.5 61.8 36.5

Grand Mean 18.9 66.5 70.2 64.4 51.4

Wechsler Adult Intelligence Scale


Information not available.
75
APPENDIX B: POST EXPERIMENTAL SESSION INTERVIEW

Post experimental session interviews were conducted


with each S. After the Ss had achieved a criterion run, the
E entered the performance booth. The dialogue proceeded as
follows:

You did very good and won the prize. After every-
one else has finished playing the game, I will let
you know who won the "Snoopy." Now, can you tell
me which light was the happy light [E waited for
a reply]? Can you tell me which light was the sad
light [E again waited for a reply]?

If the Ss answered by saying "the happy light was the

green one in the legs and the sad light was the red one in

the legs" or some variation of this, the S was correct.

Thus, the S received one point for the purposes of a Chi

Square analysis. However, if the S's response was extremely

vague or totally incorrect, no point was awarded for the

Chi Square analysis. The total number of Ss responding cor-

rectly and incorrectly for each group were recorded in a

3x2 contingency table and a Chi Square performed. A sum-

mary of the Chi Square analysis is provided below.

CONTINGENCY TABLE

Correct Incorrect
1 - " • • ' • •'

Group R-P 9 0
Group R 9 0
Group P 4 5

X^ = 12.312 df = 2 P<.01
76
2
The X analysis indicated that a significantly greater
number of Ss in group P did not know the correct dimensions
when compared to groups R-P and R.

You might also like