Journal or Personality and Social Psychology
1984. Vol 47. No 2,427-439
Copyright 1984 by the
American Psychological Association, Inc
Construct Accessibility and Depression: An Examination
of Cognitive and Affective Factors
Ian H. Gotlib
C. Douglas McCann
University of Western Ontario
London, Ontario, Canada
Ohio State University
Columbus, Ohio
Two studies in which the parameters of construct accessibility in depression were
examined are reported. In Experiment 1, depressed and nondepressed subjects
were required to name the colors of tachistiscopically presented depressed-content,
neutral-content, and manic-content words. Because of the predicted accessibility
and interference effects of the depressed-content words, the depressed subjects were
expected to demonstrate longer response latencies to these words than to the nondepressed-content words. This response pattern was found for the depressed subjects;
the nondepressed subjects did not demonstrate differential response latencies. In
Experiment 2, a mood-induction paradigm was used to investigate whether the
interference effects obtained in Experiment 1 were due to temporary mood differences between the depressed and nondepressed subjects, or were a function of
more stable depression-associated patterns of information processing. Although
predicted group differences were obtained on a mood adjective checklist, no effects
were found for task performance. These results suggest that transient mood is not
a sufficient explanation for the results obtained in Experiment 1. The implications
of the present findings for the understanding of both construct accessibility and
depression are discussed, and directions for future research are suggested.
There is a rapidly growing body of literature
in which the role of cognitive processes in the
development and maintenance of depression
is examined. These studies are based largely
on Beck's (1967, 1976) cognitive model of
depression, in which depressed persons are
characterized by a "cognitive triad": a negative
view of the self, the world, and the future. This
conceptualization of depression ascribes considerable importance to the role of cognitive
schemata. Schemata are generally viewed as
cognitive structures that affect the encoding,
storage, and retrieval of information (Taylor
& Crocker, 1981). According to Beck, the
schemata used by depressed individuals are
negative in nature, leading the depressed person both to selectively filter out positive information and to perceive negative or neutral
information as being more negative than it
actually is; consequently, the depressed indiWe would like to express our appreciation to Albert N.
Katz and Patricia G. Devine for their helpful comments
on an earlier draft of this article, and to Alexandra Urbanowicz for her help in collecting the data.
Requests for reprints should be sent to Ian H Gotlib,
Department of Psychology, University of Western Ontario,
London, Ontario, Canada N6A 5C2.
vidual is often inaccurate in his or her perception of the environment.
Empirical support for this model has come
from studies in which the differential accuracy
of recall of reinforcement in depressed and
nondepressed subjects is assessed. For example,
Wener and Rehm (1975), Buchwald (1977),
and Gotlib (1981) reported that depressed
subjects underestimated the number of reinforcements they received on a laboratory task.
In addition, Gotlib (1983) found that depressed psychiatric patients recalled an interpersonal evaluation as being significantly more
negative than was actually the case.
These investigations, then, provide support
for Beck's (1967, 1976) conceptualization of
depression, which suggests that this disorder
is characterized by the operation of negative
schemata. The results of a number of other
studies, however, are not consistent with this
formulation. For example, in a study designed
to examine both perception and recall of neutral feedback in depression, Craighead, Hickey,
and DeMonbreun (1979) found no differences
between depressed and nondepressed students.
Furthermore, several investigations have presented data that suggest that it is nondepressed
427
428
IAN H. GOTLIB AND C. DOUGLAS McCANN
individuals who distort their environment,
whereas depressed persons are painfully accurate. Nelson and Craighead (1977) reported
that consistent with predictions derived from
Beck's model of depression, depressed students
recalled receiving a higher rate of punishment
than did nondepressed students. This difference, however, was due to the nondepressed
students underestimating the amount of punishment they had actually received; the depressed students were accurate in their recall.
Lewinsohn, Mischel, Chaplin, and Barton
(1980) reported an analogous finding in which
the degree of concordance between self-ratings
and observer ratings of social competence was
higher in depressed than in nondepressed subjects. Finally, both Alloy and Abramson (1979)
and Golin and his colleagues (e.g., Golin, Terrell, & Johnson, 1977; Golin, Terrell, Weitz,
& Drost, 1979) reported results suggesting that
nondepressed individuals see themselves as
having more control in experimental situations
than is actually the case, whereas depressed
persons are relatively accurate in their assessment of response-outcome contingencies.
There are discrepancies, therefore, in the
results of studies examining the perception and
recall of environmental information in
depression. To understand the nature of these
differences, it is important to note that these
investigators have not attempted to distinguish
among the various motivational (cf. Alloy,
1982) and cognitive processes involved in their
experimental tasks. Therefore, the extent to
which each of these researchers assesses the
operation of cognitive or motivational processes is unclear. Furthermore, studies of cognition in depression are characterized by an
inattention to more molecular aspects of information-processing, which virtually prevents
one from precisely ascertaining the nature of
the deficits demonstrated by depressed individuals. The present studies represent an attempt to address these limitations through a
delineation of the parameters of construct accessibility in depression (cf. Higgins & King,
1981; McCann & Higgins, in press).
It has long been suggested (Bartlett, 1932;
Bruner, 1957) that when exposed to stimulus
input, individuals engage in a process of "effort
after meaning," in which incoming stimuli are
interpreted with reference to the person's
preexistent cognitive structures. Researchers
postulate that both the stimuli attended to and
the interpretations placed on those stimuli are
influenced by the relative accessibility of various constructs or schemata (cf. Higgins,
Rholes, & Jones, 1977; Wyer & Srull, 1981).
Accessibility, in this context, refers to "the
readiness with which a stored construct is utilized in information processing" (Higgins &
King, 1981, p. 71).
Higgins and King (1981) found that the accessibility of constructs is a function of such
factors as the expectations of the individual
and the recency and frequency of activation
of the construct. In addition to these temporary differences in accessibility, it is likely
that individuals demonstrate more stable differences in the accessibility of particular types
of constructs because of such factors as their
previous socialization experiences (cf. McCann
& Higgins, in press). In this context, and consistent with Beck's (1967, 1976) concept of
the schema, both Higgins and King (1981)
and McCann and Higgins (in press) suggested
that depressed individuals may differ from
nondepressed persons in the relative accessibility of negative-content constructs.
In this article we describe two studies conducted to gain a better understanding of the
nature of negative schemata in depressed individuals. In the first study we assess the relative accessibility of positive and negative constructs in depressed and nondepressed university students. The second study is an
attempt to replicate the findings of the initial
investigation through a mood induction paradigm.
Experiment 1
If depressed individuals are in fact characterized by negative schemata through which
they filter environmental information, they
should be naturally "primed" to perceive negative-content stimuli. This priming is viewed
as the result of the relatively greater accessibility of negative constructs for depressed individuals.
The phenomena of priming and construct
accessibility have received considerable attention from investigators concerned with the
mechanisms involved in information processing. In one procedure used to assess differential construct accessibility, the Stroop
CONSTRUCT ACCESSIBILITY AND DEPRESSION
(1935) color word task is used. The original
Stroop procedure requires subjects to name
the color of ink in which color names are
printed. Thus, for example, on seeing the word
red printed in blue ink, the correct response
from the subjects would be "blue." Subjects
typically demonstrate longer reaction times to
naming the ink colors of color words (e.g., red,
green, yellow) than of noncolor words (e.g.,
sock, cup, drawer). Investigators have interpreted this differential response latency in
terms of response competition: The processing
of the content of the color word interferes with
the response of naming the color of ink in
which it is printed.
Warren (1972, 1974) and Conrad (1974,
1978) have used a variation of this task in
demonstrating accessibility or priming effects
that are due to recency of construct activation.
Warren (1972), for example, first aurally presented subjects with three words that all had
the same category name, (e.g., aunt, uncle,
cousin). This was followed by the visual presentation of one of three different words: a
word that was in the original list (e.g., uncle),
the category common to all of the words in
the list (e.g., relative), or an unrelated word
(e.g., coal). As in the original Stroop task, these
words were printed in colored ink, and the
subjects' task was to name the color in which
the word was printed as quickly as possible
after the word was presented. Warren found
that color-naming reaction times were significantly longer to both category names and
words from the list than for unrelated words.
He explained this finding by suggesting that
the category designation of a word is activated
as part of the word's encoding, and both the
category and the word itself, therefore, produce
interference effects on this Stroop procedure
through response competition.
This paradigm can be profitably extended
to the study of depression. If Beck's (1967,
1976) postulation that depressed individuals
are characterized by the primacy of negative
schemata is valid, the relatively greater accessibility of negative constructs in depressed individuals would be expected to produce interference effects for negative-content words
analogous to those found by Warren (1972).
We therefore predicted that depressed subjects
would demonstrate longer response latencies
to naming the colors of depressed-content than
429
neutral- or manic-content words. On the basis
of results of previous work by Kuiper and bis
colleagues (e.g., Deny & Kuiper, 1981), depressed subjects were also expected to recall
a greater number of depressed-content than
neutral- or manic-content words on a subsequent incidental recall task. We predicted that
nondepressed subjects, on the other hand,
would not exhibit either differential reaction
times or differential recall for the three types
of words.
Method
Subjects. Thirty undergraduate students at the University of Western Ontario constituted the final sample of
subjects for this study.1 These subjects were selected on
the basis of their scores on the Beck Depression Inventory
(BDI; Beck, Ward, Mendelson, Mock, & Erbaugh, 1961).
Fifteen subjects had BDI scores of 4 and below, and the
remaining IS subjects all obtained scores of 9 or above
(nondepressed: M = 1.00, SD = 1.25; depressed: M =
15.80, SD = 3.78). The nondepressed group comprised
5 men and 10 women, and the depressed group comprised
2 men and 13 women.
Materials and equipment. One hundred and fifty different words were used as stimuli in the present study.
One hundred of these words were drawn from a list developed by Myers (1980), who obtained ratings of the degree
of self-descnption of each of 400 adjectives from psychiatric
patients diagnosed as either depressed or manic. The 50
adjectives rated both by the depressed patients as most
self-descriptive and by the manic patients as least selfdescriptive were selected as the depressed-content words;
similarly, the 50 adjectives rated both by the depressed
patients as least descriptive and by the manic patients as
most descriptive were chosen as the manic-content words.
Finally, 50 neutral-content words were chosen from the
list presented by Thorndike and Lorge (1944), and were '
matched with the depression- and manic-content words
with respect to frequency of occurrence in English. None
of the neutral-content words were contained in the list of
400 self-descriptive adjectives used by Myers.
Each of the 150 words was printed in capital letters
5
/i6 in. (8 mm) high on a 4 X 6 in. (10.2 X 15.3 cm) white
index card. Each word was printed in one of five colors
of ink (red, yellow, green, brown, or blue). Ten of the 50
words in each content category were printed in each of
the five colors, and the length of the words was matched
as closely as possible for each color.
Procedure. Each subject was tested individually. Sub1
Before we began this study, we decided to include only
those subjects who obtained scores of 4 or below and 9
or above on the Beck Depression Inventory. Each subject's
responses to this inventory were not scored, however, until
after the subject had completed the experiment, in order
that the experimenter remain blind with respect to each
subject's group assignment A total of 54 subjects actually
completed the experiment before data from a final sample
of 30 subjects was collected.
430
IAN H. GOTLIB AND C. DOUGLAS McCANN
Table 1
Group Means and Standard Deviations on Color-Naming Response Latencies to the Stimulus Words
Word type
Depressed
Neutral
Manic
Subject group
M
SD
M
SD
M
SD
Nondepressed
Depressed
694.67
762.40
76.88
94.94
699.67
739.00
82.82
77.44
692.47
746.80
70.53
77.09
Note Ms are in msec
jectsfirstcompleted the BDI and were told to name aloud
the colors of a series of words. The words were presented
in random order for each subject via a Ralph Gerbrands
Co. two-field tachistoscope (Model T-2B-10). An ElectroVoice Inc. (Model 621) microphone was positioned approximately 2V« in. (7 cm) from the subject's mouth. The
microphone was connected to a Lafayette Instruments Co.
(Model 18010) voice-activated relay, which stopped a
Hunter-KJockounter (Model 120) Umer at the initiation
of the subject's vocal response Subjects w?re told that five
colors would be used and were given five practice trials,
which involved naming the colors of the words one, two,
three, four, and Jive, as quickly as possible. When it was
clear that each subject understood the task requirements,
the experimenter presented the 150 stimulus words Each
trial consisted of a 1-s presentation of a fixation cross
followed by a 250-msec blank interval followed by the
presentation of a stimulus word. Onset of the stimulus
started the timer, which was stopped by the subject's vocal
response. Each stimulus remained in view for 1 5 s, regardless of the subject's reaction time. The interval between
stimulus presentations was used by the experimenter to
record the response latency and to reset the equipment
for the next trial Thus this time interval was not held
constant, but was generally around 5 s
After the presentation of the 150 stimulus words, subjects
were asked to write down, in any order, as many of the
words as they could recall, and were given 3 min to complete this task. Following the recall task, subjects were
thanked for their participation and were fully debriefed.
Results
Response latency. As is typically the case
in vocal reaction time tasks, errors were virtually nonexistent, occurring on only 1.3% of
the trials given (57 errors in 4,500 trials). These
errors consisted of calling out the word name,
prematurely tripping the voice key with a hesitation sound such as "ah" or "er," or responding too quietly to trip the voice key. The
trials on which errors occurred were excluded
from further analysis.
Response latencies for each subject were averaged across the 50 words within each of the
three content categories. Means and standard
deviations for the color-naming latencies for
the depressed and nondepressed subjects on
the depressed-, neutral-, and manic-content
words are presented in Table 1. A twoway repeated measures analysis of variance
(A NOVA) was performed on the data with subject group (depressed, nondepressed) and word
content type (depressed, neutral, manic) as
factors. The analysis revealed a significant effect only for the Group X Word Type interaction, F(2, 56) = 3.27, p < .05. Neither of
the main effects attained statistical significance:
for group, F{1, 28) = 3.54, p > .05; for word
type, F(2, 56) = 1.77, p > .1. Newman-Keuls
post hoc tests revealed that whereas the nondepressed subjects did not demonstrate significantly different response latencies to the
three types of words, the depressed subjects
took significantly longer (p < .05) to name the
colors of the depressed-content words than
they did to name the colors of either the neutral- or the manic-content words; those two
times did not differ significantly from each
other.
Recall of words. Means and standard deviations for the number of words recalled from
each of the three content categories by subjects
in the two groups are presented in Table 2. A
two-way repeated measures ANOVA conducted
on these data yielded a significant effect only
for word type, .F(2, 56) = 7.85, p < .001; neither the main effect for group nor the Group X
Word Type interaction attained statistical significance (both .Fs < 1). Newman-Keuls post
hoc tests indicated that all subjects recalled
significantly more depressed-content than either neutral- or manic-content words (p < .05).
Discussion
The results of this study provide empirical
support for an important aspect of Beck's
CONSTRUCT ACCESSIBILITY AND DEPRESSION
Table 2
Group Means and Standard Deviations on
Number of Stimulus Words Recalled
Word type
Depressed
Neutral
Manic
Subject group
M
SD
M
SD
M
SD
Nondepressed
Depressed
3.67
3.53
1.54
2 17
2.47
1.60
1.85
1.35
2.73
240
1.67
1.92
(1967, 1976) model of depression. Depressed
university students evinced longer color-naming response latencies to depressed-content
than to neutral- or manic-content words. In
contrast, nondepressed students did not demonstrate differential reaction times to the three
types of words. In addition to providing support for Beck's concept of a "negative schema"
in depression, the obtained results are consistent with McCann and Higgins's (in press)
position that depressed individuals differ from
nondepressed persons in the relative accessibility of positive and negative cognitive constructs.
The absence of differential response latencies to the three types of words for the nondepressed subjects suggests that the accessibility of nondepressed words for nondepressed
individuals is not as well or specifically organized as that of depressed persons for depressed words. Another possibility, however,
is that the type of words used to examine category accessibility in nondepressed individuals
was not appropriate for assessing this construct. The words used in this study were based
on ratings obtained by depressed and manic
psychiatric patients, and it is likely that the
manic self-descriptive adjectives were not salient to nondepressed university students, and
therefore were not differentially accessible. It
would be important in future work to obtain
adjectives more descriptive of nondepressed
students and to attempt to replicate the results
obtained with the depressed subjects in this
study.
The finding that depressed and nondepressed subjects did not differ with respect to
their recall of the three types of words warrants
comment. A number of studies have demonstrated that depressed individuals, relative
to nondepressed subjects, recall (a) receiving
fewer reinforcements on laboratory tasks
431
(Gotlib, 1981; Nelson & Craighead, 1977), (b)
evaluations to be more negative than is actually
the case (Gotlib, 1983), and (c) a greater number of depressed-content adjectives that they
have rated as self-referent (Derry & Kuiper,
1981). The discrepancy between the findings
of these investigations and those of our study
likely involves the different levels of stimulusprocessing demanded in these studies. Craik
and Lockhart (1972) have suggested that the
greater is the depth to which a stimulus is
processed, the greater is the probability that
it will later be recalled. Consistent with this
formulation, depressed individuals appear to
demonstrate a bias in their recall of stimuli
only if they are required to explicitly attend
to the information. Subjects in this study did
not explicitly process the stimulus words, and
this fact may account for the absence of
depression-associated differences in recall.
Parenthetically, this explanation may also account for the present finding of a response
latency effect in the absence of recall differences. Craik and Tulving (1975), for example,
have demonstrated that latency and recall
measures are not necessarily concordant, and,
consistent with a levels-of-processing formulation, Shiffrin and Schneider (1977) suggest
that stimuli that are processed automatically
(as in the present study) are not necessarily
encoded in long-term memory and available
to recall. Clearly, further research is required
in order to examine more systematically the
effects of levels of processing on response interference and stimulus recall in studies of
depression.
A related point involves the finding that
subjects in both groups recalled significantly
more depressed-content than either neutralor manic-content words. These three types of
words were matched with respect to frequency
of occurrence in English, and this dimension,
therefore, cannot account for this finding.
However, Rubin (1980) has demonstrated that
other characteristics of words, such as emotionality, associative frequency, pronounceability, and imagery, all affect the ease with
which the words are recalled. Therefore, although the three types of words were matched
with respect to frequency, it is likely that they
differed along one or more other dimensions,
resulting in enhanced recall by all subjects of
the depressed-content words.
432
IAN H. GOTLIB AND C. DOUGLAS McCANN
Finally, it is not clear whether the interference effects observed in the depressed subjects
are due to a temporary dysphoric mood state
or are a function of more stable patterns of
information-processing in depressed individuals. This question is especially important in
light of Beckys (1967, 1976) formulation of
cognition as causal in depression. If the accessibility effects are simply due to sad mood,
Beck's proposition that cognitive styles in depressed individuals are stable and precede
dysphoric affect would be challenged. Experiment 2 was designed to examine this issue,
attempting to replicate the results of Experiment 1 through a mood induction procedure.
differential moods associated with depressed
and nondepressed states may serve to increase
the accessibility of mood congruent cognitive
categories. Thus although these two explanations are concordant with respect to the importance of the role played by category accessibility in producing the results obtained
in Study 1, they differ in the role ascribed to
temporary (or chronic) mood differences. Although the mood interpretation is weakened
somewhat by the failure in Study 1 to find
evidence for increased accessibility of positive
or neutral categories for the nondepressed
subjects, it is compatible with a large body of
literature in which the effects of mood on a
variety of cognitive and memory processes are
demonstrated (e.g, Bower, 1981; Clark & Isen,
Experiment 2
The results of Study 1 suggest that, whereas 1982; Isen, Shalker, Clark, & Karp, 1978;
depressed individuals have more negative con- Teasdale & Fogarty, 1979; Zajonc, 1980).
tent categories accessible to them than positive
Study 2 was designed to examine the poscontent categories, nondepressed individuals sibility that the effects obtained in Study 1
are not characterized by differential accessi- were due simply to mood differences between
bility. It becomes important now to examine the depressed and nondepressed subjects. To
the processes underlying this depressed-non- this end, we used the Velten Mood Induction
depressed difference in category accessibility. Procedure (Velten, 1968) to produce positive,
Higgins and his colleagues (e.g., Higgins & neutral, and negative moods. A substantial
King, 1981; McCann & Higgins, in press) have body of recent research provides empirical
demonstrated that temporary increases in ac- support for the contention that such mood
cessibility can be produced by both the recency induction procedures do in fact manipulate
and frequency of category activation. Higgins subjects' moods. Hale and Strickland (1976)
suggests further that analogous "chronic" ef- and Strickland, Hale, and Anderson (1975),
fects may be operative in a number of indi- for example, found that subjects in a depresvidual difference variables, and specifically sion-induction condition, relative to subjects
cites depression as an example. In the case of who underwent an elation-induction, indicated
depressed persons, it is possible that chronic feeling more depressed, anxious, and hostile,
increases in negative category accessibility were less expansive in graphic expression,
could result from such factors as the frequency wrote more slowly, and performed more poorly
with which negative labels have been applied on the Digit Symbol Test. These results have
to their behavior, both by themselves and by been extended and supported by several other
others. Such a cognitive interpretation is clearly investigations (e.g., Alloy, Abramson, & Viscompatible with other current cognitive and cusi, 1981; Natale, 1977; Natale & Hantas,
interpersonal analyses of depression (e.g., Beck, 1982; Riskind, Rholes, & Eggers, 1982).
1967; Gotlib & Robinson, 1982; Lewinsohn, Therefore, if the mood interpretation of the
1974).
effects of Study 1 is correct, similar results
Within this context, the depression-asso- should be obtained in Study 2. If, however,
ciated patterning of response latencies ob- the effects of Study 1 are due to cognitive diftained in Study 1 can be interpreted as re- ferences associated with more chronic charflecting the operation of differential category acteristics of depression, no such effects should
accessibility, a cognitive process difference. be obtained in Study 2.
There is, however, an equally viable alternative
explanation for these results: The cognitive difMethod
ferences are viewed as reflecting mood differences between depressed and nondepressed
Subjects Thirty undergraduate students at the Uniindividuals. According to this interpretation, versity of Western Ontario received course credit for par-
433
CONSTRUCT ACCESSIBILITY AND DEPRESSION
ticipatiog in the present study.2 Subjects were randomly
assigned to one of three experimental conditions: induced
depression, induced elation, or neutral mood. There were
2 men and 8 women in each of the three conditions.
Procedure. Each subject was tested individually. Subjects were told that the study concerned the effects of
mood on perceptual processes. Each subjectfirstcompleted
the BDI and then participated in Velten's (1968) mood
induction procedure. Subjects in the elation-induction
condition were asked to readfirstsilently, then aloud, each
of SO statements progressing from relative mood neutrality
(e.g., "Today is neither better nor worse than any other
day") to elation (eg., "God, I feel great"). Subjects in the
depression-induction condition read SO statements progressing from relative mood neutrality to depression (e.g.,
"I want to go to sleep and never wake up"). Subjects in
the neutral condition read SO neutral statements (e.g.,
"The review is concerned with the first three volumes")
to control for the effects of reading statements. All statements were typed on separate 4 X 6 in. (10.2 X 1S.3 cm)
white index cards; subjects were first shown a card containing Velten's standard instructions encouraging them
to try to respond to the feeling suggested by each statement
Subjects were asked to read the statements at their own
speed and were given 7.S min to complete this task; the
experimenter was out of the room. Subjects were further
instructed that if theyfinishedreading the statements before
the experimenter returned, they were to reread some of
the statements.
After the 7.5 min had elapsed, the experimenter returned
to the room and asked the subjects to complete the Brief
Form of the Multiple Affect Adjective Check List (MAACL;
Zuckerman & Lubin, 1965). In order to heighten or reinstate the induced mood, subjects were then given 3.S min
to reread the last 25 statements. Following this reinduction,
subjects were told that they were to name aloud the colors
of a series of words. As in Experiment 1, the 150 stimulus
words were presented in random order, and the subjects'
response latency to naming the color of each word was
recorded.
Following the presentation of the ISO words, subjects
were asked to write down, in any order, as many of the
words as they could recall, and were given 3 min to complete this task. Following this recall task, subjects were
thanked for their participation in the study and were fully
debriefed, hi addition, subjects in the depression-induction
condition read the SO elation-induction statements before
leaving.
Results
Mood. An ANOVA conducted on the BDI
scores of subjects in the three experimental
conditions proved nonsignificant, F\2, 27) <
1. (The mean BDI score for all subjects was
4.03, with a SD of 2.85). Thus there was no
difference in level of depression among subjects
in the three experimental conditions before
the mood induction procedure.
Means and standard deviations for the responses of subjects in the three induction
groups to the depression, anxiety, and hostility subscales of the MAACL are presented in
Table 3
Induction Group Means and Standard Deviations
on the Three Subscales of the MAACL
MAACL subscale
Depression
Anxiety
Hostility
Condition
M
SD
M
SD
M
SD
Depression
Neutral
Elation
7.90
3.10
2.30
5.30
2.08
2.00
3.70
0.70
0.80
5.77
0.82
1.23
7.10
4.40
4.60
2.77
2.88
2.12
Note, MAACL = Multiple Affect Adjective Check List
Table 3. A multivariate analysis of variance
conducted on the data yielded a significant
effect, F\6, 50) = 2.55, p < .05. Subsequent
univariate ANOVAS produced significant effects
for the depression and hostility subscales: for
depression, F{2, 27) = 7.55, p < .005, and for
hostility, ^ 2 , 27) = 3.33, p < .05; the effect
for the anxiety subscale did not reach significance, F{2, 27) = 2.45, p > .1. NewmanKeuls post hoc tests indicated that subjects in
the depression-induction group reported feeling significantly more depressed (p < .01) and
more hostile (p < .05) following the induction
procedure than did subjects in either the neutral- or the elation-induction groups, who did
not differ significantly from each other.
Response latency. As was the case in Experiment 1, subjects made very few errors on
the color-naming task (54 errors in 4,500 trials,
for a 1.2% error rate). The trials on which
errors occurred were excluded from further
analysis.
Response latencies for each subject were averaged across the 50 words within each of the
three content categories. Means and standard
deviations for the color-naming latencies for
subjects in the three mood induction groups
on the depressed-, neutral-, and manic-content
words are presented in Table 4. A two-way
repeated measures ANOVA conducted on these
data failed to reveal significant effects: for
group, F(2, 27) < 1; for word type, F\2, 54) =
2
Because this study was concerned with the effects of
induced mood, it was necessary that subjects be initially
nondepressed. Therefore, the BDI was administered to
each subject at the beginning of the experimental session,
and the data from those subjects who obtained scores of
9 or above were excluded from further analysis. The final
sample consisted of 30 subjects.
434
IAN H. GOTLIB AND C. DOUGLAS McCANN
Table 4
Induction Group Means and Standard Deviations on Color-Naming Response Latencies
to the Stimulus Words
Word type
Neutral
Depressed
Manic
Condition
M
SD
M
SD
M
SD
Depression
Neutral
Elation
778.50
762.90
744.20
146.55
87.49
86.97
770.10
750.80
736.90
134.50
83.33
86.25
765.50
747.20
746.60
122.74
84.71
92.46
Note. Ms are in msec.
1.49, p > .1; for Group X Word Type, F{4,
54) < 1. Thus subjects in the three mood induction groups were indistinguishable with respect to their color-naming latencies; furthermore, subjects did not respond differentially
to the three types of words.
Recall of words. Means and standard deviations for the number of words recalled from
each of the three content categories by subjects
in the three mood induction groups are presented in Table 5. A two-way repeated measures ANOVA conducted on these data failed
to yield significant effects (all Fs < 1). Subjects
in the three groups, therefore, did not differ
in the number of words recalled from each of
the three content categories.
with depression. Thus the procedure of Study
2 was modeled after that of Study 1, but instead
of assessing depressed and nondepressed subjects, it induced depression and elation.
The first issue of concern involves the success of the present procedure in inducing the
intended moods. Researchers have shown that
Velten's (1968) mood induction procedure is
successful in effecting changes in such variables
as speed and expansiveness of writing (Hale
& Strickland, 1976), time and strength of retrieving pleasant and unpleasant memories
(Natale & Hantas, 1982; Teasdale & Fogarty,
1979), and mood and self-devaluation (e.g.,
Frost, Graf, & Becker, 1979; Riskind et al.,
1982). The results of the analyses of the MAACL
subscales in the present study suggest that appropriate moods were in fact induced: Subjects
Discussion
who underwent the depression induction reIn Study 1, depressed subjects evidenced ported more depressed affect following the ingreater response latencies to depressed-content duction procedure than did subjects who parwords, whereas nondepressed subjects did not ticipated in the elation induction.
demonstrate differential reaction times. Study
If the differential response latencies obtained
2 was designed to assess the possibility that in Study 1 were due solely to differences in
these results were due to mood-related in- mood between the depressed and the nondecreases in category accessibility rather than to pressed subjects, then similar results would be
more stable cognitive differences associated expected in this second study with depressionand elation-induced subjects. The results of
this study indicated that this predicted pattern
Table 5
was not obtained. No group differences were
Induction Group Means and Standard Deviations
found with respect to response latencies to the
on Number of Stimulus Words Recalled
three types of words. Interestingly, an examWord type
ination of the mean response latencies for this
study presented in Table 4 suggests that subDepressed
Neutral
Manic
jects in all the induction groups responded to
the stimulus words with reaction times similar
Condition
M
SD
M
SD
M
SD
to those demonstrated by the nondepressed
Depression 2.30 1.34 2.90 1.66 1.80 1.23
subjects in Study 1 (see Table 1). Given the
Neutral
1.90 1.29 1.60 0.84 2.00 1.15
fact that all subjects in the second study were
Elation
1.80 1.87 2.50 1.72 2.10 1.37
nondepressed before the mood manipulation,
CONSTRUCT ACCESSIBILITY AND DEPRESSION
these results suggest that the group effects obtained in Study 1 are not due solely to transient
mood differences between the depressed and
nondepressed subjects. This is not to state, of
course, that mood plays no role in increasing
the accessibility of mood-congruent categories,
but simply that mood is not a sufficient explanation for the original pattern of results.
Several previous investigations have reported results suggesting that mood induction
procedures can affect the relative accessibility
of positive and negative life experience memories (e.g., Teasdale & Fogarty, 1979; Teasdale
& Taylor, 1981). It may be instructive, then,
to discuss briefly our finding that mood induction did not affect construct accessibility.
Two explanations may account for this discrepancy. First, whereas previous researchers
have examined memory for life experiences,
response interference that was due to construct
activation was assessed in the present study.
Tulving (1972) has distinguished between these
two kinds of information processing systems
and has argued that memory for experiences
(episodic memory) is dependent to a much
greater extent on context (e.g., mood) than is
memory for meaning and associations (semantic memory). The present results are fully
consistent with this formulation. Second, previous investigations have examined intentional
activation of memory, and it may be that this
type of activation is more susceptible to
changes in mood than is the automatic activation examined in this study. It remains for
future research to examine these two possibilities more explicitly.
435
The finding in Study 1 that all subjects recalled more depressed-content than either
neutral- or manic-content words was not replicated in Study 2. The most likely explanation
for this discrepancy involves the method used
to induce mood in Study 2. The Velten (1968)
procedure requires that subjects read a series
of 50 statements printed on index cards. It
may be that this procedure of reading and
encoding words immediately before undertaking the Stroop task produced a form of
proactive interference for recall of the stimulus
words. Interestingly, the direction of the difference in the mean number of words recalled
by subjects in the two studies is consistent
with this explanation (Study 1: M = 8.2; Study
2: M = 6.3). In the future researchers might
explore this possibility by examining differential effects of the Velten procedure and other
less explicitly verbal mood induction techniques (e.g., hypnosis, free gifts) on cognitive
processing.
Onefinalissue relevant to the present study
concerns the possible role played by demand
characteristics (cf. Orne, 1962). Several researchers have recently questioned whether the
experimental effects obtained through mood
induction procedures are the result of an actual
change in affect or, rather, are due to demand
characteristics (e.g., Buchwald, Strack, &
Coyne, 1981; Polivy & Doyle, 1980). For an
experimental effect to be caused by demand
characteristics, it is necessary that subjects be
aware of the expected results of a manipulation
and act in such a way as to confirm the perceived hypotheses. The only explicit demands
of the Stroop procedure used in the present
study seem to be similar to those inherent in
the writing speed task used by Hale and
Strickland (1976)—that is, for the depressioninduced subjects to respond more slowly than
normal and the elation-induced subjects to respond more quickly than normal, regardless
of word type. This pattern, a main effect for
group, was not found in this study, and it is
unlikely, therefore, that demand characteristics
played a significant role in affecting the obtained results.
The results of this study have important
implications for Beck's (1967, 1976) cognitive
model of depression. Beck suggests that dysfunctional cognitions or negative schemata
cause the depressed mood and other affective
and somatic symptoms of depression. The difficulties of unambiguously demonstrating the
fundamental causal role of cognition in
depression have recently been discussed in detail by Coyne and Gotlib (1983). The present
results suggest that depression-associated differences in construct accessibility are not a
consequence of depressed affect. Although this
finding does not constitute a direct test of
General Discussion
Beck's causality-of-cognition hypothesis, it is
nonetheless consistent with the position that
The research reported in this article was
negative cognition preceeds rather than follows designed to examine the relative accessibility
depressed affect.
of negative and positive trait constructs in de-
436
LAN H. GOTLIB AND C. DOUGLAS McCANN
pressed and nondepressed individuals. On the
basis of the work of Higgins and King (1981)
and McCann and Higgins (in press), we hypothesized that depressed individuals would
be characterized by a greater accessibility of
negative than of positive cognitive constructs.
The color-naming response latencies of depressed and nondepressed subjects on a modified Stroop task were assessed as a function
of the valence of the stimulus words (i.e., positive, neutral, and negative). As predicted, depressed subjects exhibited significantly longer
response latencies to negative than to positive
or to neutral words, reflecting the operation
in the depressed subjects of interference effects
for the negative words. In contrast, nondepressed subjects evinced no such differential
pattern in their reaction times. These results
were interpreted as supporting the hypothesized accessibility differences between depressed and nondepressed individuals.
Study 2 was designed to further clarify the
relation between depression and construct accessibility. It was suggested that rather than
being due to cognitive factors associated with
depression, the interference effects obtained
in Study 1 may have been a function of differences in mood between the depressed and
nondepressed subjects. In order to assess this
possibility, Velten's (1968) mood induction
procedure was used in Study 2 to experimentally induce depressed and elated mood states.
In this study, initially nondepressed subjects
participated in the Study 1 color-naming task
after being exposed to a depressed-, elated-,
or a neutral-mood induction procedure. Although subjects in these three experimental
conditions responded differentially to a mood
adjective check list, no group differences were
found in task performance. These results suggest that the significant depression-associated
effects demonstrated in Study 1 are more likely
the result of cognitive than of affective differences between the depressed and nondepressed
subjects. The results of these two studies together provide support for Beck's (1967, 1976)
cognitive model of depression and particularly
for Beck's conceptualization of negative schemata operational in this disorder.
(e.g., Alloy, 1982), and cognitive (e.g., Beck,
1976) antecedents, concomitants, and consequences. Perhaps the greatest proportion of
research in this area has been addressed to an
assessment of Beck's cognitive model of
depression. According to this model, depressed
individuals are characterized by the presence
of a cognitive triad by which they view themselves, the world, and the future in negative
terms. The depressed individual's thinking is
seen to reflect automatic-and systematic cognitive distortions (e.g., selective abstraction and
arbitrary inference) that compound his or her
current depressed state. Beck postulated that
these negative cognitions are the result of the
operation of negative cognitive structures, or
schemata, that constitute "the basis for
screening out, differentiating, and coding
stimuli that confront the individual" (Beck,
Rush, Shaw, & Emery, 1979, p. 13). Although
a great deal of research has been cited in support of this model, little effort has been devoted
to examining the nature of the mechanisms
responsible for the pervasiveness of the depressed persons' cognitive distortions. The
present studies represent an attempt to address
this issue within the context of recent work in
social cognition.
Social perception is often viewed as a process
wherein the perceiver encodes incoming stimuli in terms of preexistent cognitive structures
that affect the encoding, storage, and retrieval
of information (cf. Taylor & Crocker, 1981).
Temporary experimentally induced increases
in the accessibility of certain structures or
constructs have been shown to dramatically
affect how incoming stimuli are processed (see
McCann & Higgins, in press, and Wyer &
Srull, 1981, for reviews of this work). It has
been suggested, however, that individuals may
also evidence more stable differences in their
accessibility of cognitive constructs (cf. Higgins, King, & Mavin, 1982; McCann & Higgins, in press). One specific area to which this
suggestion has been applied is that of depression. Higgins and his colleagues have postulated that depressed individuals may be characterized by highly accessible negative cognitive constructs, and the results of the present
Recent interest in factors involved in the studies provide strong empirical support for
etiology and maintenance of depression has this hypothesis.
resulted in the examination of its interpersonal
It is instructive to consider the present re(e.g., Coyne, 1976; Gotlib, 1982), motivational sults in relation to studies in which different
CONSTRUCT ACCESSIBILITY AND DEPRESSION
methodologies are used to assess schematic
processing in depressed and nondepressed individuals. Most notably, Deny and Kuiper
(1981) and Kuiper and MacDonald (1982)
have conducted a series of investigations designed to delineate the parameters of depressive
self-schemata. Their paradigm requires depressed and nondepressed subjects to make
explicit ratings of depressed- and nondepressed-content adjectives on a number of different scales, and the time taken to make each
rating is assessed. They found that depressed
psychiatric patients make judgments about
depressed-content adjectives as quickly as
nondepressed individuals make judgments
concerning nondepressed-content adjectives.
In the present studies, subjects were not asked
to attend to the content of the stimulus words;
indeed, they were implicitly required to ignore
it. The fact that even under these instructions
depressed individuals exhibited interference
effects for negative-content adjectives provides
convergent evidence for the strength of construct accessibility in depression.
Also of interest is the absence in Study 1
of differential response latencies in the nondepressed subjects. As discussed earlier, these
results may be due to the organization of accessible constructs in nondepressed individuals, or to the relative lack of salience to these
subjects of the stimulus words used in this'
study. Although it remains for future research
to address these issues more explicitly, these
results are nonetheless important in light of
a number of theoretical formulations whereby
nondepressed persons interpret their environment in a self-enhancing manner. Lewinsohn
et al. (1980), for example, have described an
"illusory warm glow" evidenced by nondepressed individuals. Similarly, both Miller and
Ross (1975) and Bradley (1978) have discussed
the concept of "self-serving attributions," and
Golin and his colleagues (e.g., Golin et al.,
1977, 1979) have demonstrated an "illusion
of control" in nondepressed persons. Researchers postulate that these "positivity
biases" are ego-enhancing or ego-protective
(cf. Alloy, 1982), and if this is in fact the case,
one would not expect nondepressed individuals to demonstrate biases on tasks that do
not involve threats to the subjects' egos. The
task used in the present study clearly conforms
to this criterion, and the absence of longer
437
response latencies by the nondepressed subjects to the positive or manic words adds credence to this motivational formulation.
The present analysis suggests that the negative cognitions, distortions, and inferences
attributed to depressed individuals may be due
at least in part to the relatively greater accessibility of negative than positive constructs activated in their processing of environmental
information. Moreover, this differential accessibility appears to be a function not simply
of depressed mood but of more stable, cognitive factors associated with depression. These
findings may have important implications for
approaches to therapy with depressed persons.
A number of theorists (cf. Costello, 1972;
Ferster, 1974) have conceptualized depression
as resulting from a low rate of positive reinforcement, and Lewinsohn (1974) has used
this formulation in suggesting that therapy with
depressed persons focus on increasing the
amount of positive reinforcement received.
The present results suggest that this approach
to therapy may be less effective than postulated
because depressed individuals may not be
primed to perceive reinforcement in their environment. The present findings also suggest
that the efficacy of cognitive therapies that are
based upon rational thought-testing procedures (e.g., Beck et al., 1979; Ellis, 1973) will
be increased if these therapies are complemented by techniques designed explicitly to
alter the relative accessibility to depressed individuals of positive and negative constructs.
It remains for researchers in the future to examine these postulations.
Finally, two limitations of the present results
warrant comment. First, subjects in these two
studies were university students, some of whom
were experiencing mild depression. A number
of investigators have cautioned that empirical
results obtained with this population may not
be generalizable to more clinical forms of
depression (e.g., Depue & Monroe, 1978;
Gotlib, 1984; Gotlib & Asarnow, 1979). Before
implications of the present findings for the
understanding of clinical depression can be
advanced, therefore, it is important that these
results be replicated with a sample of clinically
depressed individuals. Second, the studies reported in this article did not address the issue
of the specificity of differential construct accessibility to depression. It is possible that
438
IAN H. GOTLIB AND C. DOUGLAS McCANN
greater accessibility to negative than to positive
constructs is not specific to depression, but
rather is characteristic of global psychopathology. In the future, researchers must examine
this issue, assessing construct accessibility in
individuals demonstrating forms of psychopathological behavior other than depression.
The results of these investigations, in conjunction with those of the present studies,
promise to further elucidate the nature of cognitive functioning in depression, leading to a
more comprehensive understanding of this
disorder.
Conrad, C. (1978). Some factors involved in the recognition
of words. In J. W. Cotton & R. L. Klatzky (Eds.), Semantic factors in cognition (pp. 103-120). New \brlc:
Erlbaum.
Costdlo, C. G. (1972). Depression: Loss of reinforcers or
loss of reinforcer effectiveness? Behavior Therapy, 3,
240-247.
Coyne, J. C. (1976). Toward an interactional description
of depression. Psychiatry, 39, 28-40.
Coyne, J. C , & Gotlib, I. H. (1983). The role of cognition
in depression: A critical appraisal. Psychological Bulletin,
94. 472-505.
Craighead, W. E., Hickey, K. S., & DeMonbreun, B. C.
(1979). Distortion of perception and recall of neutral
feedback in depression. Cognitive Therapy and Research,
3. 291-298.
Craik, F. I. M., & Lockhart, R. S. (1972). Levels of processing: A framework for memory research. Journal of
References
Verbal Learning and Verbal Behavior, 11, 671-684.
Craik, F. I. M., & Tulving, E. (1975). Depth of processing
Alloy, L. B. (1982, August). Depression- On the absence
and the retention of words in episodic memory. Journal
of self-serving cognitive biases. Paper presented at the
of Experimental Psychology. General, 104, 268-294.
90th Annual Meeting of American Psychological AsDepue, R. A., & Monroe, S. M. (1978). Learned helpsociation, Washington, D.C.
lessness in the perspective of the depressive disorders:
Alloy, L. B., & Abramson, L. Y. (1979) Judgment of
Conceptual and definitional issues. Journal of Abnormal
contingency in depressed and nondepressed students.
Psychology, 87, 3-20.
Sadder but wiser? Journal ofExperimental PsychologyDerry, P. A., & Kuiper, N. A. (1981). Schematic processing
General. 108. 441-485.
and self-reference in clinical depression. Journal of AbAUoy, L. B., Abramson, L. Y., & Viscusi, D. (1981). Innormal Psychology, 90, 286-297.
duced mood and the illusion of control. Journal ofPerEllis, A. E. (1973). Humanistic psychotherapy: The rasonality and Social Psychology, 41, 1129-1140.
lional-emotive approach. New York: Julian Press, Inc.
Bartlett, F. C. (1932). Remembering Cambridge, MA:
Ferster, C. B. (1974). Behavioral approaches to depression.
Cambridge University Press.
In R. J. Friedman & M. M. Kate (Eds.), The psychology
Beck, A. T. (1967). Depression Causes and treatments
of depression: Contemporary theory and research (pp.
Philadelphia: University of Pennsylvania Press.
29-53). Washington, DC: V. H. Winston.
Beck, A. T. (1976). Cognitive therapy and the emotional
disorders. New York: International Universities Press.
Frost, R. Q, Graf, M., & Becker, J. (1979). Self-devaluation
Beck, A. T, Rush, A. J., Shaw, B. X, & Emery, G. (1979).
and depressed mood. Journal of Consulting and Clinical
Cognitive therapy of depression. New York: Guilford.
Psychology, 47, 958-962.
Beck, A. T., Ward, C. H., Mendelson, M., Mock, J., & Gohn, S., Terrell, T, & Johnson, B. (1977). Depression
Erbaugh, J. (1961). An inventory for measuring depresand the illusion of control. Journal of Abnormal Psysion. Archives of General Psychiatry, 4, 561-571.
chology, 86, 440-442.
Bower, G. H. (1981). Mood and memory. American Psy- Golin, S., Terrell, T, Weitz, J., & Drost, P. L. (1979). The
chologist, 36. 129-148.
illusion of control among depressed patients. Journal
Bradley, G. W. (1978). Self-serving biases in the attribution
of Abnormal Psychology, 88. 454-457.
process: A ^examination of the fact or fiction question.
Gotlib, I. H. (1981). Self-reinforcement and recall: DifJournal of Personality and Social Psychology, 36. 56ferential deficits in depressed and nondepressed psy71.
chiatric inpatients. Journal ofAbnormal Psychology, 90,
Brunei; J. S. (1957). On perceptual readiness. Psychological
521-530.
Review, 64, 123-152.
Gotlib,
I. H. (1982). Self-reinforcement and depression in
Buchwald, A. M. (1977). Depressive mood and estimates
interpersonal interaction: The role of performance level.
of reinforcement frequency. Journal of Abnormal PsyJournal of Abnormal Psychology, 91, 3-13.
chology. 86, 443-446.
Buchwald, A. M., Strack, S., & Coyne, J. C. (1981). De- Gotlib, I. H. (1983). Perception and recall of interpersonal
feedback: Negative bias in depression. Cognitive Therapy
mand characteristics and the Velten mood induction
and Research, 7, 399-412.
procedure. Journal of Consulting and Clinical PsyGotlib, I. H. (1984). Depression and general psychopachology. 49. 478-479.
thology in university students. Journal of Abnormal
Clark, M. S., & Isen, A. M. (1982). Toward understanding
Psychology, 93, 19-30.
the relationship between feeling states and social behavior. In A. H. Hastorf & A. M. Isen (Eds.), Cognitive Gotlib, I. H., & Asarnow, R. F. (1979). Interpersonal and
socialpsychology (sip. 73-108). New 'Vbrk: Elsevier North
impersonal problem-solving skills in mildly and clinically
Holland, Inc.
depressed university students. Journal of Consulting and
Clinical Psychology. 47, 86-95.
Conrad, C. (1974). Context effects in sentence comprehension: A study of the subjective lexicon. Memory and Gotlib, I. H., & Robinson, L. A. (1982). Responses to
Cognition, 2, 130-138.
depressed individuals: Discrepancies between self-report
CONSTRUCT ACCESSIBILITY AND DEPRESSION
and observer-rated behavior. Journal ofAbnormal Psychology, 91. 231-240.
Hale, W. D., & Strickland, B. R. (1976). The induction
of mood states and their effect on cognitive and social
behaviot Journal of Consulting and Clinical Psychology,
44. 155.
Higgins, E. G., & King, G. A. (1981). Accessibility of
social constructs: Information-processing consequences
of individual and contextual variability. In N. Cantor
& J. F. Kihlstrom (Eds.), Personality, cognition, and
social interaction (pp. 69-121). Hillsdale, NJ: Erlbaum.
Higgins, E. T, King, G. A., & Mavin, G. H. (1982). Individual construct accessibility and subjective impression
and recall. Journal of Personality and Social Psychology,
43, 35-47.
Higgins, E. X, Rholes, W. S., & Jones, C. R. (1977). Category accessibility and impression formation. Journal
of Experimental Social Psychology, 13, 141-154.
Isen, A. M., Shalker, T. E, Clark, M., & Karp, L. (1978).
Affect, accessibility of material in memory, and behavior
A cognitive loop? Journal ofPersonality and Social Psychology, 36, 1-12.
Kuiper, N. A., & MacDonald, M. R. (1982). Self and other
perception in mild depressives. Social Cognition, 3, 223239.
Lewinsohn, P. M. (1974). A behavioral approach to
depression. In R. J. Friedman & M. M. Katz (Eds.),
The psychology ofdepression: Contemporary theory and
research (pp. 157-178). Washington: V. H. Winston.
Lewinsohn, P. M., Mischel, W., Chaplin, W., & Barton,
R. (1980). Social competence and depression: The role
of illusory self-perceptions? Journal of Abnormal Psychology, 89, 203-212.
McCann, D., & Higgins, E. T. (in press). Individual differences in communication: Social cognitive determinants and consequences. In H. E. Sypber & J. L. Applegate (Eds.), Understanding interpersonal communication: Social cognition and strategic processes in
children and adults. Beverly Hills, CA: Sage.
Miller, D. T, & Ross, M. (1975). Self-serving biases in the
attribution of causality: Fact or fiction? Psychological
Bulletin, 82. 213-225.
Myers, J. (1980). Adjectives rated as self-descriptive by
depressed and manic psychiatric patients Unpublished
manuscript, University of Calgary, Calgary, Alberta,
Canada.
Natale, M. (1977). Effects of induced elation-depression
on speech in the initial interview. Journal of Consulting
and Clinical Psychology, 45, 45-52.
Natale, M., & Hantas, M. (1982). Effect of temporary
mood states on selective memory about the self. Journal
of Personality and Social Psychology. 42, 927-934.
Nelson, R. E., & Craighead, W. E. (1977). Selective recall
of positive and negative feedback, self-control behaviors,
and depression. Journal of Abnormal Psychology, 86,
379-388.
Orne, M. T. (1962). On the social psychology of the psychological experiment: With particular reference to demand characteristics and their implications. American
Psychologist, 17, 776-783.
Polivy, J., & Doyle, C. (1980). Laboratory induction of
mood states through the reading of self-referent mood
statements: Affective changes or demand characteristic?
Journal of Abnormal Psychology. 89, 286-290.
439
Rislrind, J. H., Rholes, W. S., & Eggers, J. (1982). The
NWten mood induction procedure: Effects on mood and
memory. Journal of Consulting and Clinical Psychology,
SO. 146-147.
Rubin, D. C (1980). 51 properties of 125 words: A unit
analysis of verbal behavior. Journal of Verbal Learning
and Verbal Behavior, 19, 736-755.
Shim-in, R. M., & Schneider, W. (1977). Controlled and
automatic human information processing: n. FV" ^»iimi
learning, automatic attending, and a general theory.
Psychological Review, 84, 127-190.
Strickland, B. R., Hale, W. D., & Anderson, L. K. (1975).
Effect of induced mood states on activity and self-reported affect. Journal of Consulting and Clinical Psychology. 43. 587.
Stroop, J. R. (1935). Studies of interference in serial verbal
reactions. Journal of Experimental Psychology, 18, 643662.
Taylor, S. E., &. Crocker, J. (1981). Schematic bases of
social information processing. In E. T. Higgins, C. P.
Herman, & M. P. Zanna (Eds.), Social cognition: The
Ontario symposium (Vol. 1, pp. 89-134). Hillsdale, NJ:
Erlbaum.
Teasdale, J. D., & Fogarty, S. J. (1979). Differential effects
of induced mood on retrieval of pleasant and unpleasant
events from episodic memory. Journal of Abnormal
Psychology, 88, 248-257.
Teasdale, J. D., & Taylor, R. (1981). Induced mood and
accessibility of memories: An effect of mood state or
of mood induction procedure? British Journal ofClinical
Psychology, 20, 34-48.
Thorndike, E. L., & Lorge, I. (1944). The teachers' word
book of30,000 words. New York; Bureau of Publications.
Tulving, E. (1972). Episodic and semantic memory. In E.
Tulving & W. Donaldson (Eds.), Organization of memory
(pp. 381-403). New York: Academic Press.
Velten, E. A. (1968). A laboratory task for induction of
mood states. Behavior Research and Therapy, 6, 473482.
Warren, R. E. (1972). Stimulus encoding and memory.
Journal of Experimental Psychology, 94, 90-100.
Warren, R. E. (1974). Association, directionality, and
stimulus encoding. Journal of Experimental Psychology,
102, 151-158.
Wener, A. E., & Rehm, L. P. (1975). Depressive affect A
test of behavioral hypotheses. Journal of Abnormal Psychology, 84, 221-227.
Wyer, R. S., & SruU, T. K. (1981). Category accessibility:
Some theoretical and empirical issues concerning the
processing of social stimulus information. In E, T. Higgins, C. P. Herman, & M. P. Zanna (Eds.), Social cognition: The Ontario symposium (Vol. 1, pp. 161-197).
Hillsdale, NJ: Erlbaum.
Zajonc, R. B. (1980). Feeling and thinking: Preferences
need no inferences. American Psychologist, 35, 151175.
Zuckerman, M., & Lubin, B. (1965). Manualfor the Multiple Affect Adjective Check List. San Diego, CA: Educational and Industrial Testing Service.
Received January 21, 1983
Revision received April 25, 1983