The effect of noncontingent feedback on attributional style

THE EFFECT OF NONCONTINGENT FEEDBACK ON ATTRIBUTIONAL STYLE M. VIRGINIA MARTIN, JOSEPH A. BUCKHALT AND RANDOLPH B. PIPES

Auburn University

MARYRUTH K . NIVENS

Auburn University at Montgomery

JUDD A. KATZ

Auburn University at Montgomery

The relationship between experiences with noncontingency and attributional style was examined in experimental and correlational models. One hundred and twenty-six college student subjects were provided noncontingent, contingent, or no feedback as to the correctness of their responses on a concept discrimination problem. They then completed an attributional style questionnaire, a scale designed to measure life experiences with noncontingency, and a depression inventory. The experimental hypothesis was that exposure to noncontingent outcomes, both in a laboratory and historically, would result in more depressive attributions than would exposure to contingent outcomes or to no outcomes. The manipulation of feedback contingencies produced the predicted effect on attributions for positive and negative events combined into a single composite difference score. A correlation between life experiences with noncontingency and attributions was found only for the positive events measure. However, such life experiences were correlated significantly with depression.

The original learned helplessness theory proposed that experiences with noncontingent, uncontrollable events led to an expectation of future noncontingency that generalized to new situations and resulted in the cognitive, motivational, and emotional deficits characteristic of helplessness and of clinical depression (Seligman, 1975). The theory later was revised to include an attributional component (Abramson, Seligman, & Teasdale, 1978). The effects of experiences with response independent events were now said to depend on the causal explanations that the individual makes for such events. Depressed people were hypothesized to exhibit a bias toward making internal, stable and global attributions for negative outcomes, and external, unstable and specific attributions for positive outcomes in what has come to be called the “depressive attributional style” (Seligman, Abramson, Semmel, & von Baeyer, 1979).

This attributional model of depression has received considerable experimental support (Anderson, Horowitz, & French, 1981; Golin, Sweeney, & Shaeffer, 1981; Metalsky, Abramson, Seligman, Semmel, & Peterson, 1982; Raps, Peterson, Reinhard, Abram- son, & Seligman, 1982), but an important and unresolved issue is that of the relationship between noncontingency and attributional style. The question whether an individual‘s typical attributional responses are influenced systematically by exposure to noncontingent reinforcement or are formed independently of reinforcement history is a critical link between learned helplessness theory and attribution theory. If responses to uncontrollable outcomes were to vary capriciously with attributional biases that are unrelated to prior

This paper is based on a doctoral dissertation submitted by the first author to the Department of Counselor

Requests for reprints should be sent to Joseph A. Buckhalt, Department of Counselor Education, Auburn Education, Auburn University.

University, Alabama 36849.3501. 456

Noncontingent Feedback 457

experiences with noncontingency, then the learned helplessness theory of depression and the attributional model of depression would appear to be functionally unrelated.

This study will examine the effects of exposure to noncontingency on attributional style in experimental and correlational designs. Feedback contingencies (independent variable) will be manipulated, and the effect of this manipulation on attributions (dependent variable) will be measured. Also, the correlation between historical experiences with noncontingency and attributions and between such experiences and depression will be determined. The hypothesis to be tested is that experiences with noncontingent feedback, both in a laboratory setting and in real life, result in a depressive attributional pattern.

METHOD Subjects

Volunteers between the ages of 17 and 26 were recruited from local college cam- puses. The total sample of 126 included 62 females and 64 males. Subjects were assigned randomly, by sex, to one of three feedback conditions. There were 41 subjects in a noncontingent feedback group, 44 subjects in a contingent feedback group, and 41 subjects in a no feedback group. Instruments

Levine Discrimination Problem. The experimental task consisted of five discrimination problems (Levine, 197 1) of a type popular in learned helplessness research (Hiroto & Seligman, 1975; Pittman & Pittman, 1979, 1980). In each problem, two stimulus patterns were presented on each of a series of 10 5 x 7 index cards. The patterns were composed of various combinations of five geometric dimensions, each with a complementary value. One pattern on the card contained a combination of values from each of the five dimensions, and the other pattern contained the complementary values. The dimensions and their values were: (1) letter, A or T; (2) letter color, shaded or clear; (3) border that surrounded the letter, circle or square; (4) dot above the letter, one or two; and ( 5 ) underline, solid or broken. The experimenter arbitrarily selected one value (the letter T, for example) from each I0-card problem as correct, and the task for the subjects was to find this value. By selecting the right or left side of the cards, and receiving feedback as to the correctness of the choices, a solution could be ascertained.

Attributional style was measured by use of the Attributional Style Questionnaire (Peterson et al., 1982). This instrument contains written descriptions of 12 hypothetical events, half of which are positive events and half of which are negative events. Subjects are asked to imagine themselves in each situation and to write down the one major cause of each event if it had happened to them. This cause then is rated on a 7-point scale for each of three attributional dimensions: Internal- external, stable-unstable, and global-specific. High ratings indicate greater internality, stability, and globality of attributions. Composite scores for positive outcomes and for negative outcomes are created by averaging rating on the six positive and six negative outcome scales. A single difference score that utilizes all of the information produced by the Attributional Style Questionnaire is obtained by subtracting negative events composite scores from positive events composite scores in a manner described by Blaney, Behar, and Head (1980) and by Hamilton and Abramson (1983). The authors report coefficient alphas of .75 and .72 for the composite attributional style scales for positive and negative events, respectively, and 5-week test-retest liability coefficients of .70 and .64 for the two scales.

Attributional Style Questionnaire.

458 Journal of Clinical Psychology, September 1987, Vol. 43, No. 5

Success-Failure Scale. In an attempt to assess subjects’ historical experiences with noncontingent outcomes, the Success-Failure Scale was created. This is a 10-item instrument on which subjects are asked to rate, on a scale of 1-7, the degree to which they perceive their total life experiences as having been successful or unsuccessful in five major life areas: (1) relationship with mother; (2) relationship with father; (3) relationship with friends; (4) sexual relationships; and ( 5 ) academic experiences. They then are asked for similar estimates of the degree to which they were able to control outcomes in these areas. High rating signified more success and control. The rationale for this approach to assessing an individual’s history of exposure to noncontingency is derived from the literature, in which success and failure are confounded most often with contingency and noncontingency (Sargent & Lambert, 1978), and in which the perception of control has been shown to be a more important determinant of how the world im- pinges on us than the fact of control (Glass, Singer, & Friedman, 1969; Langer, 1983).

The internal reliability for the scale, computed by use of Cronbach’s coefficient alpha (Cronbach, 1951), was .47 for items 1-5, which measure success, .54 for items 6-10, which measure control, and .68 for the entire scale.

The Beck Depression Inventory (Beck, 1978) is composed of 21 items, each of which contains four statements weighted 0-3 in terms of in- tensity. Subjects are asked to choose the statement that best describes their feelings during the past week, and the sum of the highest rating for each of the items yields a total depression score. The possible range of scores is 0-63; higher scores indicate higher levels of depression. The 1978 revision of the Beck Inventory is similar to the earlier version, for which Beck (1 967) reported a Spearman-Brown corrected split-half reliability coefficient of .93 and a correlation of every item with the total score at significance levels of .01 or greater. Also, the scale is reported to show correlations of .65 and .67 with clinical judgments and .75 with MMPI Depression scale scores. Design

The experimental analysis employed a 2 x 3 randomized block design in which sex was an assigned independent variable and feedback contingency was a manipulated independent variable that contained three levels: Noncontingent feedback, contingent feedback, and no feedback. The dependent variable was composite difference scores on the Attributional Style Questionnaire.

In the correlational analysis, the 10 Success-Failure scores were predictor variables, and Attributional Style Questionnaire scores for positive events, Attributional Style Ques- tionnaire scores for negative events, and Beck Depression Inventory scores were criterion variables in three multiple regression equations.

In addition, Success-Failure scores were divided into low (mean - .5 standard deviations) and high (mean + .5 standard deviations) groups and entered as an assigned independent variable in a multivariate analysis of variance in which Attributional Style Questionnaire scores for positive and negative events were the dependent variables. Procedure

Subjects were presented individually with the concept discrimination task. An ex- planation and practice problem were provided. On the five test problems, subjects in the noncontingent feedback group were told that .5 of their choices were correct and .5 were incorrect on a predetermined schedule (Example: IICIICCICC). Regardless of whether they chose the right or left pattern, for half of the cards they were told “that’s correct,” and for half they were told “that’s incorrect.” At the completion of each 10-card problem, subjects were asked for the solution and always were told that it was incorrect.

Subjects in the contingent feedback group were given the same instruments, but were provided with accurate feedback. Twelve of the 44 subjects in this group solved

Beck Depression Inventory.

Noncontingent Feedback 459

all of the five problems. Twenty-four subjects missed only one problem. Seven subjects missed two problems, and one subject failed to find solutions for four of the five problems.

Subjects in the no feedback group were asked to select a value in each problem and to report whether it appeared on the right or left side of the cards. No feedback was given.

Upon completion of the concept discrimination problems, subjects completed the Attributional Style Questionnaire, the Success-Failure Scale, and the Beck Depression Inventory. Four subjects omitted parts of the Attributional Style Questionnaire, and these were entered as missing values in the statistical analyses.

RESULTS All data analysis employed statistical procedures from SPSS" (1983). Cell means

and standard deviations for the analysis of variance are presented in Table 1. There was no main effect for sex and no Group x Sex interaction. There was, however, a significant main effect for feedback group, F(2, 116) = 3.27, p < .05, which a post hoc TukeyB test located at the comparison of noncontingent and contingent feedback groups. Examination of cell means reveals that subjects who received noncontingent feedback made more internal, stable, and global attributions for negative outcomes and more external, unstable, and specific attributions for positive outcomes than did subjects who received contingent feedback.

Table 1 Cell Means and Standard Deviations for Attributional Style Questionnaire Composite Diyerence Scores by Feedback Group and Sex

Feedback group Sex Noncontingent Contingent No feedback

Females M 1.53 3.54 3.66

SD 4.30 2.16 2.03 N 20 21 20

Males M 3.69 5.06 2.97

SD 3.44 2.60 2.15 N 20 20 21

When all of the Success-Failure Scale scores were entered into multiple regression equations, a significant correlation was found with Beck Depression Inventory scores, R = .54, F(10, 1 1 1) = 4.61, p < .001. But the multiple correlations between Success- Failure and Attributional Style Questionnaire scores for positive events and between Success-Failure and Attributional Style Questionnaire scores for negative events failed to attain statistical significance.

Separating Success-Failure scores into low and high groups in a MANOVA design revealed a significant main effect for this factor, multivariate F(2, 73) = 5.48, p < .01, which a follow-up univariate analysis located at the dependent variable - Attributional


Style Questionnaire scores for positive events. As indicated by cell means (Table 2), subjects who described themselves as successful and in control of outcomes (high Success- Failure scores) gave more internal, stable, and global explanations for positive events than did subjects who rated themselves low on these scales. No main effect for feedback group appeared in this analysis, and a significant Group x Success-Failure interaction that appeared in the univariate analysis of attributional scores for positive events could not be interpreted in the absence of a significant multivariate interaction effect.

Table 2 Cell Means and Standard Deviations for Dependent Measures by Feedback Group and Success-Failure

N

~~~

ASQ-Ba ASQ-Gb

Noncontingent feedback group High S-F‘ 13 M 12.84 16.95

SD 2.72 2.03 LOW S-F 10 M 12.75 14.27

SD 1.49 2.28

Contingent feedback group High S-F 18 M 12.04 16.33

SD 1.89 1.92

Low S-F 1 1 M 12.14 16.53

SD 1.47 1.37

No feedback group High S-F 13 M 13.00 16.61

SD 1.75 1.54

LOW S-F 15 M 12.52 14.90 SD 2.35 1.79

aAttributional Style Questionnaire scores for negative outcomes. bAttributional Style Questionnaire scores for positive outcomes. ‘Success-Failure Scale scores.

Finally, the correlation between the Attributional Style Questionnaire scores for positive events and Beck Depression Inventory scores was - .33 (p < .001), and between Attributional Style Questionnaire scores for negative events and Beck Depres- sion Inventory scores the correlation was .17 (p < .05).

DISCUSSION According to the reformulated learned helplessness theory of depression, the dimen-

sions of an individual’s reaction to uncontrollable events are defined by the causal explanations that are made for those events. Internal, stable, and global explanations for negative outcomes, and external, unstable, and specific explanations for positive outcomes predispose the individual to become clinically depressed. If the theory is indeed correct, it follows logically that a causal relationship must exist between exposure to response independent events and depressive explanatory tendencies. In this study, such a relationship was demonstrated experimentally. Subjects exposed to noncontingent feed-

Noncontingent Feedback 46 1

back as to the correctness of their responses on concept discrimination problems subse- quently made significantly more depressive attributions than did subjects who had received contingent feedback.

This finding is relevant not only to the theory of learned helplessness, but also within the broader context of the state-trait issue in attribution theory. Attributional style measures assume, implicitly and explicitly, that the attributional characteristics of in- dividuals may properly be considered to be personality traits that show temporal and situational stability (Peterson et al., 1982). The frequency with which certain attributional patterns are reported in relation to such factors as self-esteem (Ickes & Layden, 1978), sex differences (Deaux & Farris, 1977; Nicholls, 1975; Simon & Feather, 1973) and depression (Metalsky et al., 1982; Raps et al., 1982; Seligman et al., 1979) supports the viewpoint that these are enduring cognitive habits. Other researchers, however, find little evidence for cross-situational consistency in attributions (Cutrona, Russell, & Jones, 1985). The results of this study, in which small, but statistically significant changes in attributional style were produced after a brief exposure to noncontingent outcomes in a laboratory environment, suggest that attributions may indeed be reactive and, therefore, subject to experimental manipulation and a causal interpretation of findings not possible in correlational research.

No correlation was found between life experiences with noncontingency and attributional style when all data were included in multiple regression analyses, but a comparison of subjects who reported the most dissimilar experiences (high Success-Failure compared with low Success-Failure scores) revealed that those who reported the least success and control in their lives made significantly more depressive explanations for positive events. This rather weak effect may be due to deficiences in the instrument used to measure historical experiences with noncontingency.

That the Success-Failure Scale possesses some merit, though, is evidenced in its statistically significant correlation with Beck Depression Inventory scores. Subjects who reported having experienced low levels of success and control in their past also reported a more depressed mood. One must observe the caution that correlation does not imply causality and that a depressed mood may just as well have influenced subjects’ perception of life experiences as vice versa, or some third cause could account for both effects.

Finally, the stronger correlation between Success-Failure scores and Attributional Style Questionnaire scores for positive than for negative outcomes, plus the higher correlation between depression and attributions for positive than for negative events, suggest that for the subjects of this study, positive outcomes were more vulnerable to the effects of experience than were negative outcomes.

CONCLUSION The findings of this study provide positive support for the attributional reformula-

tion of learned helplessness model of depression and for the experimental hypothesis. Exposure to noncontingent feedback in a laboratory setting was shown to produce characteristic depressive attributions. No correlation was found between explanatory style for negative events and life experiences with noncontingency, but there was evidence that such experiences were related to attributions for positive events. A strong relationship between depression and noncontingent experiences was found.

REFERENCES ABRAMSON, L. Y . , SELIGMAN, M. E. P. , g, TEASDALE, J. D. (1978). Learned helplessness in humans: Criti-

que and reformulation. Journal of Abnormal Psychology, 87, 49-74. ANDERSON, C., HOROWITZ, L. M., ~r FRENCH, R. S. (1981). Attributionalstyles of the lonely and depressed.

Paper presented at the annual meeting of the Southeastern Psychological Association, Houston. (ERIC Document Reproduction Service No. ED 205 873)


BECK, A. T. (1967). Depression: Clinical, experimental and theoretical aspects. New York: Harper & Row. BECK, A. T. (1978). BLANEY, P. H., BEHAR, V., & HEAD, R. (1980). Two measures of depressive cognitions: Their association

CRONBACH, L. J . (1951). Coefficient alpha and the internal structure of tests. Psychometrika, 16, 297-334. CUTRONA, E., RUSSELL, D., & JONES, R. D. (1985). Cross-situational consistency in causal attributions. Journal

DEAUX, K . , & FARRIS, E. (1977). Attributing causes for one’s own performance: The effects of sex, norms,

GLASS, D. C., SINGER, J . E., & FRIEDMAN, L. N. (1969). Psychic cost of adaptation to an environmental

GOLIN, S., SWEENEY, P. D., & SHAEFFER, D. C. (1981). The causality of causal attributions in depression:

HAMILTON, E. W., & ABRAMSON, L. Y. (1983). Cognitive patterns and major depressive disorder: A longitudinal

HIROTO, D. S. , & SELIGMAN, M. E. P. (1975). Generality of learned helplessness in man. Journal of Per-

ICKES, W., & LAYDEN, M. A. (1978). In J. H. Harvey, W. J . Ickes, & R. F. Kidd (Eds.), New directions

LANCER, E. J . (1983). LEVINE, M. (1971).

METALSKY, G. I . , ABRAMSON, L. Y., SELIGMAN, M. E. P., SEMMEL, A., & PETERSON, C. (1982).

Beck Inventory. Philadelphia: Center for Cognitive Therapy.

with depression and with each other. Journal of Abnormal Psychology, 89, 678-682.

of Personality and Social Psychology, 47, 1043-1058.

and outcome. Journal of Research in Personality, 11, 59-72.

stressor. Journal of Personality and Social Psychology, 12, 200-210.

A cross-lagged panel correlational analysis. Journal of Abnormal Psychology, 90, 14-22.

study in a hospital setting. Journal of Abnormal Psychology, 92, 173-184.

sonality and Social Psychology, 31, 31 1-327.

in aftribution research (Vol. 2, pp. 119-152). Hillsdale, NJ: Erlbaum. The psychology of control. Beverly Hills, CA: Sage.

Hypothesis theory and nonlearning despite ideal S-R reinforcement contingencies. Psychological Review, 78, 130-140.

Attributional styles and life events in the classroom: Vulnerability and invulnerability to depressive mood reactions. Journal of Personality and Social Psychology, 43, 612-617.

NICHOLLS, J. G. (1975). Causal attributions and other achievement related cognitions: Effects of task outcome attainment value, and sex. Journal of Personality and Social Psychology, 31, 379-389.

PETERSON, C., SEMMEL, A., VON BAEYER, C., ABRAMSON, L. Y., METALSKY, G. I., & SELIGMAN, M. E. P. (1982). The attributional style questionnaire. Cognitive Therapy and Research, 6, 287-300.

PITTMAN, N. L., & PITTMAN, T. S. (1979). Effects of amounts of helplessness training and internal-external locus of control on mood and performance. Journal of Personality and Social Psychology, 37, 38-47.

PITTMAN, T. S. , & PITTMAN, N. L. (1980). Deprivation of control and the attribution process. Journaf of Personality and Social Psychology, 39, 377-389.

RAPS, C. S., PETERSON, C. REINHARD, K. E., ABRAMSON, L. Y., & SELIGMAN, M. E. P. (1982). Attributional style among depressed patients. Journal of Abnormal Psychology, 91, 102-108.

SARGENT, J . , & LAMBERT, W. E. (1978). “Learned helplessness” or “learned incompetence. ” (ERIC Docu- ment Reproduction Service No. ED 152 105)

SELIGMAN, M. E. P. (1975). Helplessness: On depression, development and death. San Francisco: Freeman. SELIGWN, M. E. P., ABRAMSON, L. Y., SEMMEL, A., & VON BAEYER, C. (1979). Depressive attributional

SIMON, J. G. , Causal attributions for success and failure at university exarnina-

SPSS user’s guide. (1983).

style. Journal of Abnormal Psychology, 88, 242-247.

tions. Journal of Educational Psychology, 64, 46-56. FEATHER, N. T. (1973).

New York: McGraw Hill.

Documents

The effect of noncontingent feedback on attributional style