Journal of Personality and Social Psychology Copyright 1982 by the American Psychological Association, Inc.

1982, Vol. 42, No. 6, 1137-1145 0022-3514/82/4206-1137S00.7S

The Causal Dimension Scale: A Measure of How

Individuals Perceive Causes
Dan Russell
College of Medicine, University of Iowa

A shortcoming of previous attribution research has been the assumption that

researchers can accurately translate causal attributions into causal dimensions.
Attributional statements are often ambiguous and even when clearly stated may
be perceived quite differently by the attributor and the researcher. The studies
reported describe the development of the Causal Dimension Scale, a measure
designed to assess how the attributor perceives the causes he or she has stated
for an event. This scale assesses causal perceptions in terms of the locus of
causality, stability, and controllability dimensions described by Weiner. Two
studies are reported that test the reliability and validity of the Causal Dimension
Scale. All three subscales were found to be reliable and valid, and a three-mode
factor analysis confirmed the three-dimensional structure of the scale. Results
also indicated differences in the perception of causes of success and failure, with
attributions for success being perceived as more internal, stable, and controllable
than attributions for failure. The relationship between the Causal Dimension
Scale and other attribution measures (such as locus of control or "attributional
style" measures) is discussed.

Previous attribution research has suffered difficulty is that attributional statements are
from a basic problem that could be termed often ambiguous (see Ross, 1977). In our
the "fundamental attribution researcher er- own research dealing with causal attribu-
ror" (i.e., assuming that the researcher can tions in the sports pages (Lau & Russell,
accurately interpret the meaning of the sub- 1980), we found many statements very dif-
ject's causal attributions). In the traditional ficult to interpret. For example, is the causal
attribution paradigm, an essential step in- explanation "They played better than we
volves the translation by the researcher of did" attributing causality to the attributor's
causal attributions into causal dimensions, own team or to the oppositipn?
such as internal-external or stable-unstable. Moreover, even when the meaning of a
Based on this classification of the subject's causal attribution is clear, the attributor
causal attributions, the investigator can then may perceive the cause quite differently than
test a variety of predictions about the attri- the researcher. As Weiner (1979) has noted,
bution process. the placement of a causal attribution in
The danger in this procedure is that the terms of causal dimensions may vary greatly
researcher and the attributor may not agree from person to person, as well as from sit-
on the meaning of a causal attribution. One uation to situation. For example, one student
may state that his or her failure in a math-
ematics course is due to lack of ability and
This article is based on a doctoral dissertation sub- perceive this cause as stable over time. An-
mitted to the Department of Psychology at the Univer- other student might also view the failure as
sity of California, Los Angeles. I would like to thank caused by ability but believe that ability in
Bernard Weiner for his supervision of this research and
comments on the present manuscript, and committee mathematics can be improved through study.
members Seymour Feshbach, Connie Hammen, Anne Situational variability in attributions can
Peplau, Tara Scanlan, and Deborah Stipek for their also occur. An ability attribution for perfor-
comments and suggestions. mance in an academic subject is undoubt-
Requests for reprints should be sent to Dan Russell,
Graduate Program in Hospital and Health Administra- edly perceived differently than an ability
tion, College of Medicine, University of Iowa, S517 attribution for performance in athletics,
Westlawn, Iowa City, Iowa 52242. where improvements in skill occur through
1137
1138 DAN RUSSELL

practice. The typical attribution paradigm Study 1

does not allow for these variations in the
perception of causes. This study had two goals: (a) generating
To address these problems, this article a set of items for the Causal Dimension
describes the development of the Causal Scale, and (b) conducting reliability and va-
Dimension Scale, a measure designed to as- lidity tests of the generated items. A set of
sess how the attributor perceives the causal items was written to assess separately each
attributions he or she has stated. First, how- of the three causal dimensions described by
ever, another issue must be addressed: Do Weiner (1979): locus of causality, stability,
people actually organize their thinking in and controllability. The conceptual defini-
terms of the causal dimensions described by tions provided by Weiner of the locus of cau-
attribution researchers and theorists? sality and stability dimensions were used for
Research supporting relationships be- writing items. That is, locus of causality was
tween causal dimensions and the conse- defined as referring to whether the cause was
quences of the attribution process, such as something about the attributor (internal) or
affective reactions and expectancies of fu- outside the attributor (external), whereas
ture success, suggests that people do process stability was defined as referring to whether
information concerning causality in terms of the cause was constant over time (stable) or
causal dimensions. For example, how can variable over time (unstable). For control-
ability attributions for success lead the at- lability, the definition was modified slightly
tributor to anticipate future successes, unless to allow both internal and external causal
he or she recognizes that ability is stable or factors to be considered controllable. A con-
unlikely to vary over time? Other research trollable cause was therefore defined as one
also suggests that causal dimensions underlie that could be changed or affected by some-
individuals' perceptions of causes. Passer one, either the actor or other people. Based
(1977) asked subjects to judge the similarity on these definitions of the dimensions, a set
of 15 causes of success and failure. Analyses of semantic differential scales was developed
revealed dimensions underlying these judg- to measure perceptions of causes along each
ments corresponding to locus of causality dimension.
and controllability. A multidimensional scal-
ing analysis of causes of loneliness by Mich- Method
ela, Peplau, and Weeks (Note 1) found di- Participants in this study were 189 undergraduate
mensions labeled as locus of causality and students (117 females, 72 males) who served as subjects
stability underlying perceptions of these to fulfill a requirement for an introductory psychology
course. The experimental design was 2 X 2 X 2 X 2 X
causes, with controllability emerging as a 2 factorial, with the factors being sex of subject, out-
third nonorthogonal dimension. Finally, a come (success or failure), locus of causality, stability,
factor-analytic study by Meyer (1980) of and controllability of the attribution. Sex of subject and
causal attributions for success and failure outcome were between-subjects factors, whereas the
outcomes revealed factors corresponding to three attribution dimensions were within-subjects fac-
tors. The eight different causal attributions used to
the locus of causality, stability, and con- manipulate the causal dimensions are shown in Table
trollability dimensions underlying impor- 1. An additional factor, order of the questionnaire ma-
tance ratings of the causal explanations.1 terials, was included to control for order effects, with
Thus, it may be possible to assess directly five random orders being used. No significant main ef-
how the attributor perceives his or her own
causal attributions in terms of causal di- 1
Falbo and Beck (1979) have also reported a mul-
mensions. Rather than having the researcher tidimensional scaling analysis of causes of success and
code the attributional statements into di- failure, in which the dimensions described by Weiner
mensions, the attributor does the coding. were not found. However, Falbo and Beck's interpre-
Assuming that such a Causal Dimension tations of the dimensions derived from their analysis
appears suspect, based on a replication by Weiner (Note
Scale is reliable and valid, responses on this 2). Because of this and other problems with the Falbo
measure should accurately reflect the mean- and Beck study, it is not discussed here (see Weiner,
ing of causal attributions to the attributor. Note 2, for more details).
 CAUSAL DIMENSION SCALE 1139

fects or interactions were associated with this variable, Table 1

so it will not be discussed further. Specific Causal Attributions Used to
Each subject completed a questionnaire that consisted Manipulate Causal Dimensions
of descriptions of eight different achievement situations,
followed by the semantic differential scales. The Stability
achievement situations consisted of an outcome (success
or failure) and one of the eight causal attributions shown Controllability Stable Unstable
in Table 1. The following is an example of one of the
situations, for failure due to lack of ability: Internal
Imagine that you have received a very low score in Controllable Stable effort Unstable effort
a class that is very important to you. You feel the Uncontrollable Ability Mood
reason you received this low score is your lack of
ability in the subject. External
While imagining themselves in each situation, the stu-
dents evaluated the cause of the success or failure out- Controllable Other's stable Other's unstable
come on 12 semantic differential scales. An example of effort effort
one of the items assessing the locus of causality dimen- Uncontrollable Task difficulty Luck
sion was "Reflects on you-reflects your situation." Stu-
dents rated the extent to which they felt the cause was
internal or external on this scale, by circling a number
from 1 to 9. The other 11 items that were used are to the stability items, these rating scales were
shown in Table 2. In total, each student made 96 ratings, found to differentiate stable from unstable
evaluating the eight causal attributions on all 12 se- causes, with the stability main effect ac-
mantic differential scales.
counting for 18-19% of the variance in these
items. For the controllability items, con-
Results and Discussion founding by the locus of causality dimension
is apparent, with main effects for the latter
To test the validity of the individual se- dimension accounting for substantial por-
mantic differential scales, each item was tions of the variance in these rating scales.
subjected to separate analyses of variance. Only one of the controllability rating scales
If a given item is indeed assessing the causal (Unintentional-intentional) appears to ad-
dimension it was designed to measure, the equately assess the controllability dimen-
main effect for that dimension should be very sion.
large. So, for example, an item designed to From these validity tests, it appears that
assess the locus of causality dimension should the locus of causality and stability dimen-
produce significantly different ratings for in- sions are assessed reasonably well. The three
ternal versus external causes. Moreover, items assessing these two dimensions can
each item should also have discriminant va- therefore be combined into subscales. One
lidity. That is, an item designed to assess concern is the reliability of these three-item
locus of causality should not also differen- scales. A coefficient a value of .88 was found
tiate stable from unstable causes or con- for both subscales.2 Thus, these two mea-
trollable from uncontrollable causes. sures also appear to be reliable. However,
Thus, the main effects for the three causal the scales assessing the controllability di-
dimensions provide validity tests for the in- mension are problematic. An examination
dividual semantic differential scales. The re-
sults are presented in Table 2, along with 2
the variance accounted for by each main In calculating the alpha coefficients, interitem cor-
effect. Considering the three locus of cau- relations were computed by collapsing across the ex-
perimental conditions. Since each subject made eight
sality items first, the results clearly indicate ratings on each rating scale (one for each causal attri-
that these items adequately distinguish be- bution) the data were treated as if there had been 189 X
tween internal and external causes. The lo- 8 = 1152 observations. Because the experimental con-
cus of causality main effect accounts for 46- ditions created a good deal of variation in the ratings
on the semantic differential scales, this procedure
59% of the variance in these items, while seemed appropriate, as correlations among the items will
very little of the variance is explained by the also reflect covariation in ratings on the items across the
other two causal dimensions. Turning next experimental conditions.
1140 DAN RUSSELL

of the individual rating scale items revealed three-item subscales. All but one of the con-
the apparent source of the problem. The con- trollability items was found to be con-
trollability items were found to be primarily founded by the locus of causality dimension.
of two types: (a) internal-controllable scales Two more controllability scales were there-
(e.g., Not under your control-under your fore constructed based on these findings and
control) and (b) external-controllable scales added to the single adequate controllability
(e.g., Under control by others-not under item. The resulting nine-item Causal Di-
control by others). These items obviously mension Scale was employed in the next
confound the locus of causality and con- study, in an attempt to replicate the findings
trollability dimensions, as indicated by the from Study 1 and to test the adequacy of
large main effect for locus of causality found the controllability items once again.
for these items. The single item that was
found to adequately measure controllability Study 2
(Unintentional-intentional) can refer to the
person performing the achievement task or Method
other people. Controllability is therefore Participants in this study were 99 undergraduates
specified independently of locus of causality. (38 females, 61 males) who participated to satisfy a
requirement for an introductory psychology course. The
To summarize these findings, the locus of experimental design was identical to that employed in
causality and stability dimensions appear to Study 1, with the same scenarios used to manipulate
be adequately assessed by their respective causal dimensions (see Table 1). Students evaluated the

Table 2
Analysis of Variance Results and Variance Accounted for by Each Main Effect
Effect

Locus of Causality Stability Controllability

Item

Locus of causality
Reflects on you-reflects 678.69*** .46 50.03*** .02 36.78*** .02
your situation
Outside of you-inside of 720.65*** .59 40.71*** .02 29.86*** .01
you
Something about you- 741.54*** .47 1.77 .00 2.60 .00
something about others
Stability
Permanent-temporary 74.92*** .08 346.58*** .19 113.51*** .05
Variable over time- 80.05*** .07 290.30*** .18 89.81*** .07
stable over time
Changeable-unchanging 26.47*** .03 276.35*** .18 54.70*** .04

Controllability
Not under your control- 973.45*** .49 47.79*** .02 240.02*** .08
under your control
You are responsible-you 855.21*** .56 29.58*** .01 78.79*** .03
are not responsible
Uninfluenceable- 4.89* .00 24.97*** .01 2.91 .00
influenceable
Someone else is 846.31*** .42 .00 3.03 .00
responsible-no one else
is responsible
Under control by others- 786.29*** .40 .00 3.97* .00
not under control by
others
Unintentional-intentional 69.53*** .08 102.61*** .05 333.53*** .23

*/><.05. **p<.01. ***/>< .001.

 CAUSAL DIMENSION SCALE 1141

Table 3
Analysis of Variance Results and Variance Accounted for by Each Main Effect
Effect

Locus of causality Stability Controllability

Item F w2 F w2 F a,2

Locus of causality
Reflects on you-reflects 425.03*** .54 4.96* .00 10.40** .01
your situation
Outside of you-inside of 356.20*** .56 1.05 .00 2.68 .00
you
Something about you- 397.26*** .50 .00 13.04*** .01
something about others
Stability
Permanent-temporary 24.84*** .05 142.06*** .19 18,80*** .02
Variable over time-stable 41.52*** .08 82.16*** .15 11.68*** .01
over time
Changeable-unchanging 7.95** .02 106.24*** .14 2.11 .00
Controllability
Uncontrollable by you or 17.87*** .04 7.06** .01 82.32*** .14
other people-controllable
by you or other people
Intended by you or other 17.66*** .02 46.14*** .01 169.44*** .24
people-unintended by
you or other people
No one is responsible- 3.06 .00 41.12*** .03 201.71*** .26
Someone is responsible

* p < .05. **p < .01. ***p < .001.

causes in each situation using the revised rating scales on the rating scales. Since the data formed
developed in Study 1. Five different random orders of an Individuals X Experimental Conditions X
the questionnaire materials were again used. No sig-
nificant order effects were found. Rating Scales matrix, a three-mode factor
analysis was conducted (Tucker, 1966). This
procedure allows the derivation of factors for
Results and Discussion each mode (individuals, conditions, and rat-
Validity tests for the individual Causal ing scales), as well as a core matrix relating
Dimension Scale items were performed, the factor structure for each mode to all oth-
identical to those reported in Study 1. These ers. Attention will be focused on the factor
results are presented in Table 3. For each structure found to underlie the rating scales.3
item, the largest main effect was found for Three principle component factors were ex-
the dimension the item was designed to as- tracted for the rating scales and rotated by
sess. Results for the locus of causality and varimax procedures to simple structure. The
stability rating scales were very similar to ^suiting factor-loading matrix is shown in
the findings from Study 1. The controllabil- Table 4. As can be seen, the factor structure
ity rating scales also appeared valid, with the for the scale very clearly corresponds to the
controllability main effect accounting for three causal dimension subscales. Also shown
14-26% of the variance. Main effects for the in Table 4 are the alpha coefficients for the
other two causal dimensions were generally three subscales. Consistent with the factor
quite small. All three causal dimensions analysis results, all three scales were found
therefore appeared to be adequately assessed to be internally consistent.
by the final nine-item measure.
To examine the factor structure of the fi- 3
More detailed information on the factor analysis
nal scale, a factor analysis was performed results may be obtained by writing the author.
1142 DAN RUSSELL

Table 4
Factor Analysis Results for the Causal Dimension Scale Items
Factor loadings

Locus of
Item causality Stability Controllability

Reflects on you-refiects your situation .558 .056 -.004

Outside of you-inside of you .534 .018 .093
Something about you-something about others .621 -.068 -.071
Permanent-temporary .052 .529 .005
Variable over time-stable over time .011 .577 .040
Changeable-unchanging -.064 .602 -.053
Uncontrollable by you or other people- .046 -.045 .575
controllable by you or other people
Intended by you or other people-unintended .031 .109 .548
by you or other people
No one is responsible-someone is responsible -.086 -.058 .593
a coefficient .867 .837 .730
Note. These are the factor loadings following a varimax rotation.

As in Study 1, scores were computed for for sex of subject were found. Significant
the three causal dimension subscales by sum- differences were found for the achievement
ming the responses to the individual seman- outcome. Overall, subjects tended to view
tic differential scales. Correlations were the causes of success as more internal, F(l,
computed among the subscale scores, col- 95) = 3.31, p < .10, more stable, F(l,
lapsing across the experimental conditions 95) = 24.33, p < .001, and more controlla-
(see Footnote 2). As suggested by the factor ble, F(l, 95) - 81.8, p < .01. These differ-
analysis results, the subscales were only ences between the evaluations of causes
moderately related to one another, the cor- following success and failure outcomes sug-
relations ranging from .19 to .28. Analyses gest that a process similar to hedonic bias
of variance were performed on the subscale is influencing the ratings of causes (see Brad-
scores. The between-subjects factors were ley, 1978; Zuckerman, 1979). In contrast to
sex of subject and outcome and the within- previous research on hedonic bias, the cur-
subjects factors were the three attribution rent findings deal with how individuals per-
dimensions. As would be expected for the ceive causal attributions, and not which
attribution dimensions, the largest effects causal attributions are used to explain suc-
were the main effects for the dimension mea- cess and failure. These findings indicate that
sured by the respective causal dimension specific causal attributions are viewed dif-
subscales (see Table 5). For the between- ferently following success and failure. So, for
subjects factors, no significant main effects example, ability attributions are perceived

Table 5
Analysis of Variance Results for the Causal Dimension Subscales
Effect
Locus of causality Stability Controllability
2
Scale F <J F <u F a;2
Locus of causality 541.63*** .62 <1 .00 <1 .00
Stability 28.94*** .05 156.14*** .20 14.28*** .01
Controllability 22.81*** .03 43.98*** .04 238.50*** .29
*** p < .001.
 CAUSAL DIMENSION SCALE 1143

as more internal, stable, and controllable administration format is designed for set-
following success than following failure. Fi- tings in which the investigator is assessing
nally, no significant interactions between sex both the respondent's causal explanation for
of subject and the achievement outcome an event and the respondent's perceptions of
were found for the ratings on the three the causes he or she has stated. Applications
causal dimension subscales. of the measure to settings in which the causal
To summarize the findings from Study 2, attributions are experimentally manipulated
the three subscales that form the final would, of course, eliminate the need for re-
Causal Dimension Scale appear to measure spondents to state causal attributions.
the dimensional properties of causes identi- Although the results of the current studies
fied by Weiner (1979). A three-mode factor clearly support the validity of the measure,
analysis confirmed the three-factor structure some precautionary comments are also in
of the Causal Dimension Scale and all three order. The validity of the measure in as-
subscales were found to be internally con- sessing causal dimensions in real-world set-
sistent. Finally, the ratings of causes on the tings needs to be established. A variety of
Causal Dimension Scale following success other factors may influence responses to the
and failure outcomes suggest that a process Causal Dimension Scale in actual achieve-
similar to hedonic bias may influence how ment settings, which could adversely affect
causes are perceived by the individual. the validity of the measure. Other evidence
suggests that the scale is valid in assessing
General Discussion causal dimensions in actual achievement set-
The items, recommended administration tings (see Russell, Note 3). A similar issue
format, and scoring for the Causal Dimen- arises in applying the Causal Dimension
sion Scale are presented in Table 6. This Scale to settings other than achievement.

Table 6
The Final Causal Dimension Scale
Instructions: Think about the reason or reasons you have written above. The items below concern your impressions
or opinions of this cause or causes of your outcome. Circle one number for each of the following scales.
1. Is the cause(s) something that:
Reflects a n aspect 9 8 7 6 5 4 3 2 1 Reflects an aspect of the
of yourself situation
2. Is the cause(s):
Controllable b y 9 8 7 6 5 4 3 2 1 Uncontrollable by you
you or other or other people
people
3. Is the cause(s) something that is:
Permanent 9 8 7 6 5 4 3 2 1 Temporary
4. Is the cause(s) something:
Intended b y y o u 9 8 7 6 5 4 3 2 1 Unintended by you or
or other people other people
5. Is the cause(s) something that is:
Outside o f y o u 1 2 3 4 5 6 7 8 9 Inside of you
6. Is the cause(s) something that is:
Variable over 1 2 3 4 5 6 7 8 9 Stable over time
time
7. Is the cause(s):
Something about 9 8 7 6 5 4 3 2 1 Something about others
you
8. Is the cause(s) something that is:
Changeable 1 2 3 4 5 6 7 8 9 Unchanging
9. Is the cause(s) something for which:
N oo n ei s 1 2 3 4 5 6 7 8 9 Someone is responsible
responsible
Note. A total score for each of the three subscales is arrived at by summing the responses to the individual items
as follows: (1) locus of causality—Items 1, 5, and 7; (2) stability—Items 3, 6, and 8; (3) controllability—Items
2,4, and 9. High scores on these subscales indicate that the cause is perceived as internal, stable, and controllable.
1144 DAN RUSSELL

Although the measure may be valid in as- measure developed by Seligman et al. (1979)
sessing the perceptions of causes in achieve- assesses the extent to which an individual
ment contexts, the validity of the scale also generally perceives achievement or affilia-
needs to be established in other settings tive events as internally or externally caused.
where attributions occur. By contrast, the Causal Dimension Scale
Another issue is the construct validity of assesses the respondent's perceptions of
the measure. If the Causal Dimension Scale causes in a particular situation. Although
does in fact assess the dimensional properties attributional styles or general beliefs con-
of causes described by Weiner (1979), then cerning locus of control may influence how
scores on the measure should be related to an individual perceives causes in a specific
other variables as predicted from Weiner's situation, causal perceptions are also greatly
model. So, for example, scores on the locus influenced by situational factors (see Wei-
of causality subscale should be related to ner, 1979). Future research needs to address
affective reactions following success and fail- trait and situational influences on causal per-
ure. Additional research of mine provides ceptions by examining the impact of attri-
some preliminary evidence for the construct butional styles or general causal beliefs and
validity of the Causal Dimension Scale, in- situational factors on the attribution process,
dicating strong relationships between scores employing measures such as the Causal Di-
on the locus of causality subscale and affec- mension Scale.
tive reactions to success and failure (Russell,
Note 3, Note 4). Although these validity Reference Notes
findings are encouraging, more research em-
1. Michela, J., Peplau, L. A., & Weeks, D. Perceived
ploying the measure to test other predictions dimensions and consequences of attributions for
from Weiner's model is clearly needed. loneliness. Unpublished manuscript, University of
Finally, the relationship between the California, Los Angeles, 1980.
Causal Dimension Scale and other measures 2. Weiner, B. On causes and causal dimensions: A re-
in the attributional domain should be men- ply to Falbo and Beck. Unpublished manuscript,
University of California, Los Angeles, 1979.
tioned. A number of measures of locus of 3. Russell, D. The impact of causal attributions on
control beliefs have been developed for a emotional reactions to a midterm examination.
wide range of life situations (Rotter, 1966) Manuscript in preparation, 1980.
and more specifically for achievement 4. Russell, D. Causal attributions and emotional ex-
perience: Towards a cognitive model of emotion in
(Crandall, Katkovsky, & Crandall, 1965) achievement settings. Paper presented at the Annual
and health-related situations (Lau & Ware, Convention of the American Psychological Associ-
1981; Wallston & Wallston, 1980). Mea- ation, Montreal, September 1980.
sures more closely related to attribution re-
search have also been devised (Laird & Ber- References
glas, 1975; Lefcourt, von Baeyer, Ware, & Bradley, G. W. Self-serving biases in the attribution
Cox, 1979). Recently an attributional style process: A reexamination of the fact or fiction ques-
measure has been developed by Seligman, tion. Journal of Personality and Social Psychology,
Abramson, Semmel, and von Baeyer (1979) 1978, 36, 56-71.
to assess the respondent's perceptions of the Crandall, V. C., Katkovsky, W., & Crandall, V. J. Chil-
dren's beliefs in their own control of reinforcements
causes of hypothetical achievement and af- in intellectual-academic achievement situations. Child
filiative events. This measure employs a for- Development, 1965, 36, 91-109.
mat very similar to the Causal Dimension Falbo, T., & Beck, R. C. Naive psychology and the
Scale. It asks respondents to state a causal attributional model of achievement. Journal of Per-
attribution for the hypothetical event and sonality, 1979, 47, 185-195.
Laird, J. D., & Berglas, S. Individual differences in the
then to rate the cause on a set of semantic effects of engaging in counter-attitudinal behavior.
differential scales assessing the attributional Journal of Personality, 1975, 43, 286-304.
dimensions of locus of causality, stability, Lau, R. R., & Russell, D. Attributions in the sports
and globality. All of these previous measures pages. Journal of Personality and Social Psychology,
1980, 39, 29-38.
are designed to assess the individual's gen- Lau, R. R., & Ware, J. F. Refinements in the mea-
eral or cross-situational perceptions of cau- surement of health-specific locus of control beliefs.
sality. For example, the attributional style Medical Care, 1981, 19, 1147-1158.
 CAUSAL DIMENSION SCALE 1145

Lefcourt, H. M., von Baeyer, C. L., Ware, E. E., & Seligman, M. E. P., Abramson, L. Y., Semmel, A., &
Cox, D, J. The Multidimensional-Multiattributional von Baeyer, C. Depressive attributional style. Journal
Causality Scale: The development of a goal specific of Abnormal Psychology, 1979, 88, 242-247.
locus of control scale. Canadian Journal of Behav- Tucker, L. R. Some mathematical notes on three-mode
ioral Science, 1979, 11, 286-304, factor analysis. Psychometrika, 1966, 31, 279-311.
Meyer, J. P. Causal attribution for success and failure: Wallston, K. A., & Wallston, B. S. Health locus of
A multivariate investigation of dimensionality, for- control scales. In H. M. Lefcourt (Ed.), Advances and
mation, and consequences. Journal of Personality innovations in locus of control research. New York:
and Social Psychology, 1980, 38, 704-718. Academic Press, 1980.
Passer, M. W. Perceiving the causes of success and
failure revisited: A multidimensional scaling ap- Weiner, B. A theory of motivation for some classroom
proach. Unpublished doctoral dissertation, University experiences. Journal of Educational Psychology,
of California, Los Angeles, 1977. 1979, 71, 3-25.
Ross, L. The intuitive psychologist and his shortcom- Zuckerman, M. Attribution of success and failure re-
ings: Distortions in the attribution process. In L. Ber- visited, or: The motivational bias is alive and well in
kowitz (Ed.) Advances in experimental social psy- attribution theory. Journal of Personality, 1979, 47,
chology (Vol. 10). New York: Academic Press, 1977. 245-287.
Rotter, J. B. Generalized expectancies for internal ver-
sus external control of reinforcement. Psychological
Monographs, 1966, 80(1, Whole No. 609). Received November 20, 1980