Casual attributions for success and failure in relation to expectancies,: sequential information, locus of control and stability

Non-independent sub-samples of a group of 224 introductory psychology volunteer subjects provided data for each of 5 simulation experiments. Although each experiment employed a design peculiar to that experiment, the overall approach was to study subjects' use of information from different sources (context information, skill versus chance orientation, performers' causal attributions, sequences of success and failure outcomes) in determining the nature of underlying causal processes. Experiment 1 pitted a "stable" causal factor (task difficulty or ability) against an "unstable" factor (luck or effort). Subjects read context paragraphs and printed sequences of success and failure outcomes which had been designed to suggest change in the process determining performers' outcomes. Subjects were required to indicate whether the "stable" or the "unstable" causal factor was more likely to have changed during the time period covered by the context paragraph and the success and failure sequence. Subjects' responses to four of the six items presented in Experiment 1 favored change in the "stable" factor, demonstrating that "stable" factors are viewed -as changeable under the conditions suggested by these items. A2X2X2X2 orthogonal design was employed in Experiment 2 to assess the effects of performer's Motivation (motivated or unmotivated), Outcome (success or failure), Locus of Control of causal factor (internal or external) and Stability of causal factor (stable or unstable) upon expectancies for success for the performer's next attempt at a skill task. Subjects read paragraphs in which they were told that a fictitous motivated (unmotivated) performer had just succeeded (failed), and that he attributed his success (failure) to luck (effort, task difficulty, ability). Subjects were then asked to indicate their expectancies for the performer's success on the next trial by making a mark through a line, the end points of which were anchored at 0% and 100%. Expectancies were taken as indirect statements of subjects' perceptions of the nature of underlying causal processes. Subjects' responses indicate that successful or motivated performers are thought to be more likely to succeed on the next trial than are unsuccessful or unmotivated performers, that a repetition of a success is thought to be more likely than is a repetition of a failure, that a change in outcome is thought to be more likely when a current outcome is attributed to an unstable as opposed to a stable causal factor, and that effort is viewed as unstable when it is invoked as an explanation for a failure but stable when it is used to explain a success. The findings for Experiment 2 must be qualified by the possibility that the experimental manipulations also produced differential prior probabilities for success among the different conditions. Experiment 3 required subjects to respond to printed sequences of success and failure outcomes by attributing cause for a designated outcome to luck, effort, task difficulty and ability. That subjects were sensitive to the arrangements of success and failure outcomes is demonstrated by the apparent systematic relationships found between sequence arrangements and attributions. Relatively high effort attributions occurred when there was consistency between the target outcome and the rest of the sequence, indicating that effort is treated as a more stable causal factor than is luck. Luck received high attributions only when the target outcome was inconsistent with the rest of the sequence. Experiment 4 attempted to demonstrate that outcomes attributed to luck are weighted differently depending upon the skill or chance nature of the task in which the outcome occurs. Subjects read paragraphs stating that a performer in a dart throwing (skill condition) or a spinner (chance condition) game would win $1.00 for each time his dart or spinner landed on a designated area. A five trial sample of the performer's outcomes was presented (0 $1 $1 $1 0), along with the information that the performer attributed the 0 on his last trial to bad luck. It was predicted that expectancies for winnings on the next.10 trials would be greater for the skill condition, since the outcome attributed to luck should carry less weight in the skill as opposed to the chance condition. Although average dollar predictions were higher for the skill condition, the results also suggested that a response bias of .5 probability of success perhaps operated in the chance but not in the skill condition. Further research is needed to determine whether luck attributions are weighted differently in skill and chance situations. The task for Experiment 5 required subjects to indicate whether printed sequences of success and failure outcomes were records of outcomes from skill or chance games. The prediction that "random" or stable probability sequences would be classified as chance derived while changing probability sequences would be classified as skill derived received support from the subjects1 responses to the different sequences presented in a within-subjects design. Since stable probability sequences were thought to be chance derived, and since luck is the major causal factor in chance situations, luck is apparently perceived as a fairly stable causal factor in the chance situation, in contrast to its unstable nature in the skill situation. The findings for the 5 experiments are discussed in relation to several theoretical perspectives. Points of agreement are noted with Heider's formulations (1958), Kelley's covariation principle (1973), the Bayesian approach to attribution of Ajzen and Fishbein (1975) , and the use of judgmental heuristics as proposed by Tversky and Kahneman (1974). However, the results reported here conflict sharply with the classification model for causal factors in achievement situations proposed by Weiner, Frieze, Kukla, Reed, Rest and Rosenbaum (1971). The stability dimension of the model appears to be invalid, since factors classified as stable factors were treated as changeable in Experiment 1, and since effort, classified by the model as an unstable factor, was treated as a relatively stable factor in Experiments 2 and 3. Finally, the findings from Experiment 5 refute the criticism by Weiner, et al. that a confound of locus of control with stability is responsible for results reported in the locus of control literature.Psychology, Department o