hychology and Aging 1986, Vol. l,No. 2, 127-13: Copyright 1986 by the Ammo I Piychohviol Association, Inc. 0882-7974/86/$00.75 Adulthood Age Differences in Causal Attributions for Cognitive, Physical, and Social Performance Margie E. Lachman and Leslie Z. McArthur Brandeis University Past research has revealed an unflattering pattern of attributions for the performance of the elderly. More specifically, poor performance by the elderly is attributed to internal and stable factors such as inability, whereas poor performance by the young is attributed to external and unstable factors such as bad luck. In the present study, 42 young (M age = 19.18 years) and 39 elderly (M age = 74.90 years) men and women made causal attributions for their own or for another person's hypothetical performance in the cognitive, physical, and social domains. When attributions for the same performance by young and elderly adults were compared, the results presented an unflattering view of the elderly, similar to the pattern in previous research. In contrast, when attributions for good versus poor performance by the elderly were compared, a more favorable picture emerged: The elderly were more likely to be given credit for their good performance than to be blamed for their poor performance. These findings give reason to question the pervasiveness of the negative view of the elderly that has been presented in previous studies. In past research comparing performance attributions for young and elderly adults, a clear pattern of age differences has emerged. The failures of the elderly are more often attributed to internal and stable causes, such as age or inability, than are those of the young (Banziger&Drevenstedt, 1982; Blank, 1982; Reno, 1979; Rodin & Langer, 1980). Although age differences in attributions are more consistently found for failures than for successes, there is some evidence that the successes of the young are more often attributed to internal and stable causes than are those of the elderly (Banziger & Drevenstedt, 1982;Gekoski&Knox, 1983). This unflattering attributional pattern reflects a stereotypic view of the elderly as being less competent than the young. Whether these views are held by the young or by the elderly themselves, they can have an impact on behavior by and toward the elderly. In studies with young adults, this attributional pattern has been associated with learned helplessness (Abramson, Seligman, & Teasdale. 1978). Thus, it is important to establish the conditions under which this attributional pattern exists for elderly adults, so that potentially debilitating consequences such as depression, low motivation, lack of effort, or poor subsequent performance can be either prevented or modified. but the generalizability of the findings is uncertain because these investigations also have been limited in a number of ways. First, with few exceptions (e.g., Blank, 1982), the studies have used an observer paradigm, in which subjects make attributions for the performance of another person rather than for then- own performance. It is known that attributions about others (observer attributions) differ from self- (actor) attributions in a variety of ways (Watson, 1982). Second, those making the attributions for the elderly have typically been young adults rather than the elderly themselves. Although it is important to consider how the young perceive the elderly, it is also important to consider how the elderly perceive themselves (Blank, 1984). Third, the behavioral events and performances for which attributions are made have been restricted to the cognitive or physical domains. Both of these domains are associated with negative stereotypes about agingrelated decrements in functioning. One question addressed in the present study was whether the attributional results from the cognitive and physical domains would be generalizable to functioning in the social domain, an area in which the elderly are likely to be seen in a more favorable light. To begin to address some of the limitations of past research, in the present study we examined the causal attributions made by both young and elderly adults for performance in the cognitive, physical, and social domains. Subjects made attributions about hypothetical situations involving themselves, another person of the same age, or another person of a different age. Based on previous research, we predicted that more internal stable (inability) attributions would be made for the elderly than for the young for poor performance. It was also predicted that ability attributions would be less likely for the elderly than for the young for good performance. Whereas previous research has focused almost exclusively on between-age-group comparisons, in the present study we also compared attributions for good versus poor performance by the elderly. Although attributions for performance by the elderly may be unflattering relative to those for the young, it is also important As we have summarized here, past studies on performance attributions in the elderly have yielded fairly consistent results, This research was supported in part by National Institutes of Health Biomedical Research Support Grant SO7 RR07044. Portions of this article were presented at the 37th annual meeting of the Gerontological Society of America, San Antonio, Texas, November 1984. We express our appreciation to Bart Moss, Shira Sanders, and Naomi Yadin for their assistance in developing the attribution questionnaire; to Barry Paul and Margaret Sail for their assistance with data collection; to Susan Trotter for her assistance with data analysis; and to two anonymous reviewers for their constructive comments on the manuscript. Correspondence concerning this article should be addressed to Margie E. Lachman, Department of Psychology, Brandeis University, Waltham, Massachusetts 02254. 127 128 MARGIE E. LACHMAN AND LESLIE Z. McARTHUR to compare attributions for success and failure by the elderly. Are the elderly more likely to receive blame for failure than to receive credit for success? In keeping with the unflattering patterns found in past age-comparative work, it was predicted that internal attributions would be more likely for the elderly for poor performance than for good performance. The predicted outcomes were expected for attributions for the self and for attributions for others in all domains, although the most favorable views of the elderly were expected for self-attributions and in the social domain. Table 1 Method 3. Subjects Ninety-six young and elderly adults were randomly assigned to one of Internal Stable (Ability) Attributions for Hypothetical Events in Three Domains Event no. Cognitive domain 1. 2. 4. 5. three yoked attribution conditions: attributions for the self, attributions for another person of the same age, and attributions for another person of a different age. The yoked subjects were grouped into 32 triads, each 6. composed of the three attribution conditions. The person in each triad who made self-attributions served as the target for the two members of the triad who made other attributions. The three subjects in each triad were always the same sex. Five triads (15 subjects) were dropped from Event and attribution 7. 8. the analyses because of five unusable questionnaires from elderly subjects. The final sample included 81 subjects with the following age and sex John wins a prize in a radio station contest because he is good at remembering numbers. (Good memory) John was asked to give out name tags at a reunion because he is good at remembering names. (Good memory) John's absentmindedness led him to purchase two subscriptions to a magazine. (Poor memory) John's forgetfulness led him to leave soup cooking too long. (Poor memory) John's solution to a brainteaser puzzle was printed in a newspaper because he is good at solving puzzles. (Good problem solving) John voted against a fund-raising activity because his good mathematical skills revealed that it would not be profitable. (Good problem solving) John's bank account interest was more than expected because John is poor at calculating interest. (Poor problem solving) There's an error in John's tax return and he owes money, because John is bad at figuring income taxes. (Poor problem solving) distribution in 27 triads: 8 young male triads, 7 young female triads, 6 Physical domain elderly male triads, and 6 elderly female triads. TheyoungaduIts,whohadameanageofl9.18years(SD = 1.5 years), were students in an introductory psychology class and were given research credit for their participation. The elderly adults, who had a mean age of 9. 10. 74.89 years (SD = 6.42 years) and a mean educational level of 12.83 years (SD = 2.53 years), were community-residing members of senior 11. citi2en organizations. A nominal donation was made to the senior organization for each member who participated in the study. 12. Measures The attribution questionnaire, which was modeled after Duval and 13. 14. Wicklund's (1973) questionnaire, included 24 hypothetical events, equally distributed across the cognitive, physical, and social domains. Within each domain there were two behavioral categories, each represented by 4 events. In the cognitive domain, behavioral categories were memory !5. 16. John completed a long distance bicycle trip because he has a lot of strength. (Strength) A pane of glass fell out when John shut the door because he shuts doors very forcefully. (Strength) John scored higher than other players at shuffleboard because John is too weak to overshoot the mark (Weakness) John had internal bleeding after an operation because he was in poor physical shape. (Weakness) John is not hit when a baseball comes out of nowhere because he is able to move quickly. (Speed) A car ran into John when he stepped on the brakes to avoid hitting a child chasing a ball, because John's braking reflexes are very fast. (Speed) John typed accurately because he types slowly. (Slowness) John runs into another car because he has slow reflexes. (Slowness) and problem-solving; in the physical domain, strength and speed; and in the social domain, independence and nurturance. In each behavioral cat- Social domain egory there were two items for each performance outcome (poor, good). The performance led to negative consequences for one item in each outcome pair, and to positive consequences for the second item. Thus, performance outcome was crossed with consequences. This design feature was included to examine whether the elderly would be consistently blamed 17. 18: when things go wrong (negative consequences) and would not be given credit when things go well (positive consequences), regardless of whether 19. their performance was poor or good. The internal stable attributions for the 24 items are summarized and presented by domain, behavioral cat- 20. egory, performance outcome, and consequences, in Table 1. All of the subjects completed a demographic questionnaire about their 21. education (years in college for young adults, or number of years of schooling and degrees received for the elderly), occupation (job aspirations for the young, or current or previous job for the elderly), marital status, 22. 23. age, race, and sex. Although the only variables used for analyses in the present study were age and sex, the other information was included so these variables would not be made salient to the subjects. 24. John had a good time at a social gathering where he knew no one, because he is very self-reliant in making new friends. (Independence) John had a terrible time on a vacation group tour because he is so independent that he dislikes group tours. (Independence) A relative does nice things for John because John needs help and companionship. (Dependence) A relative is abrupt with John because John is too emotionally dependent. (Dependence) A child listens to John's stories for hours because John is good at amusing children. (Nurturance) Whenever John goes out his puppy cries, because John is too nurturant and spoiled the puppy. (Nurturance) John refuses to visit a depressed friend because he has little sympathy for depressed people, and the friend recovered anyway. (Non-nurturance) A child gets hurt when John is babysitting because John is bad at caring for children. (Non-nurturance) Procedure Subjects were tested in groups of from 5 to 10 persons. All of the subjects completed the demographic questionnaire before the attribution Note. Odd-numbered items have positive consequences, and even-numbered items have negative consequences. This version was for men who made attributions for others. 129 AGE DIFFERENCES IN ATTRIBUTIONS questionnaire. Those making self-attributions were instructed to imagine that the 24 hypothetical events described in the attribution questionnaire had happened to them, and to rate each cause for their own performance. Those making attributions for others were instructed to imagine that the event had happened to the person described on a biographical sheet that was attached to the questionnaire, and to rate each cause for that person. The biographical sheet contained accurate information (race, age, sex, marital status, education, and occupation), taken from the demographic questionnaire, about the target in the same triad. Those making attributions for another person were told that the information pertained to a real person who had participated in the same study, but that a fictitious name was being used to conceal the person's identity. For those making self-attributions, the questionnaire used the pronoun you; for men making attributions for others, it referred to John; and for women making attributions for others, it referred to Jane. For each event, subjects rated three causes—internal stable (ability), internal unstable (effort), and external— on a 7-point scale, very unlikely cause (1) to very likely cause (7). An additional forced-choice measure was created by asking subjects to circle what they considered to be the best explanation or the most important of the three causes for each event. The analyses focused on the ability attributions because these have the most debilitating consequences if they are made for failure, or not made for success. The effort and external variables were included for the forced-choice analysis, to determine if ability was the most frequent variable chosen for the elderly's failure and the least likely to be chosen for their success. Results A 2 X 2 X 2 X 2 (Age Group [young, elderly] X Sex [male, female] X Performance Outcome [poor, good] X Consequences [positive, negative]) analysis of variance was performed for each behavioral category, to obtain error terms for subsequent comparisons of means. All a priori predictions were examined with two-tailed t tests. Analyses were done separately for each attribution condition and for positive and negative consequences, collapsed across sex.1 As we mentioned earlier, the analyses were performed on the ability attributions and on the forced-choice variable. Between-Subjects Analyses Berween-subjects analyses compared the ability and forcedchoice attributions made for each event when the target was young versus elderly. Of the 24 events for each attribution condition, significant age differences were found in ability attributions for the self for three events, in ability attributions for different-age others for four events, and in ability attributions for same-age others for four events. All but one of these outcomes reflect the predicted negative view of the elderly. These results are summarized in Table 2. Attributions for the self. The results for self-attributions supported the predictions that the elderly would be more likely than the young to rate inability as a cause for poor memory (Event 4: M young = 3.60, M elderly = 5.33), t(25) = 2.47, p < .001; poor problem solving (Event 7: M young = 2.87, M elderly = 4.67), f(25) = 2.31, p < .05; and weakness (Event 12: M young = 2.20, M elderly = 4.92), f(25) = 3.49, p < .01. Attributions for different-age others. Also consistent with predictions, those making attributions for a target of a different age were less likely to rate ability as a cause for the good memory of elderly than for the young targets' good memory (Event 1: M Table 2 Age Differences in Ability Attributions That Reflect a Flattering (+) or Unflattering (-) View of the Elderly Attribution condition Domain Self Different-age other Same-age other Cognitive Good memory Poor memory Good problem solving Poor problem solving Physical Strength Weakness Speed Slowness Social Independence Dependence , Nurturance Non-nurturance Note. An empty cell indicates no significant age difference. Entries are collapsed across positive and negative consequences. • Indicates an unflattering view of the elderly for both positive and negative consequences. young = 5.53, M elderly = 3.58), t(25) = 2,79, p < .01; strength (Event 9: M young = 5.93, M elderly = 4.25), t(25) = 2A5,p< .05; and speed (Event 14: M young = 5.80, M elderly = 2.75), t(25) = 3.86, p < .001. These effects are consistent with the predictions that ability attributions would be less likely for the elderly than for the young for good performance, and they reflect a tendency to make more unflattering attributions for the elderly. One final significant finding for those making attributions for different-age others was that they were more likely to rate inability as a cause for non-nurturance by the elderly than by the young, *(25) = 2.54, p < .05 (Event 23: M young = 2.20, M elderly = 4.08), which is also consistent with the general tendency to make unflattering attributions for the elderly. Attributions for same-age others. As predicted, those who were the same age as the target were more likely to rate inability as a cause for the elderly than for the young for weakness (Event 12: M young = 3.20, M elderly = 4.83), t(25) = 2.08, p < .05, and for the two (positive and negative) non-nurturance events (Event 23: M young = 3.00, M elderly = 5.17), t(25) = 2.93, p < .01, and (Event 24: M young = 2.33, M elderly = 4.75), t(25) = 3.21, p < .01, respectively. Those making attributions for same-age others also were more likely to rate ability as a cause for speediness by the elderly than by the young, t(25) = 2.34, p < .05, which was an unexpected outcome (Event 14: M young = 3.07, M elderly = 4.92). In contrast to all of the other findings presented thus far, this latter attribution represents a positive view of the elderly. 1 Only one significant sex difference was found: The elderly women were seen as more independent than the elderly men. To increase cell size and for ease of reporting the results, sex was ignored in the remaining analyses. 130 MARGIE E. LACHMAN AND LESLIE Z. McARTHUR Choice. Subjects were asked to choose the most salient of the three (ability, effort, external) causes for each event. The proportion of subjects who selected the ability attribution as the most likely cause was computed for each event. An arcsine transformation was performed on these proportions prior to conducting the (-test analyses. For those making self-attributions, there were four significant differences. A greater percentage of elderly (42%) than young (07%) chose the ability attribution for good memory (Event 2), t(ao) = 2.24, p < .05, which reflects a positive view of the elderly. In contrast, a negative view of the elderly was observed for three events. A greater percentage of elderly than young chose ability as the cause for poor problem solving (Event 7: young = 07%; elderly = 40%), f(oo) = 2.03, p < .05, and strength (Event 12: young = 07%; elderly = 50%), /(oo) - 2.53, p < .05. A greater percentage of the young (80%) than the elderly (36%) chose ability for nurturance (Event 22), I(oo) = 2.16, p < .05. For those making attributions for another of the same age, there was one significant difference that reflected a positive view of the elderly. A greater percentage chose the ability attribution for independent behavior (Event 17), t(ao) = 2.65, p < .01, for the elderly (50%) than for the young (07%). For those making attributions for another of a different age, there were two significant effects. A greater percentage chose ability as the cause of nurturance (Event 21), /(oo) = 2.72, p < .01, for the elderly (92%) than for the young (47%), which reflects a positive view of the elderly. Also, a greater percentage chose ability as the cause of slowness (Event 16), f(oo) = 3.09, p < .01, for the elderly (50%) than for the young (0%), which reflects a negative view of the elderly. Within-Subjects Analysis Within-subjects analyses compared attributions for poor performance versus good performance for event pairs in the same behavioral category with the same consequences, using the dependent t test. Only the results found with the elderly are reported here.3 Based on results of previous studies examining between-subjects differences, it was predicted that ability attributions would be more likely for the poor performance of elderly targets than for their good performance. Overall, 3 out of the 17 significant effects were in the predicted direction, and 14 were contrary to predictions, reflecting a positive view of the elderly. These results are summarized in Table 3. Attributions for the self. There were significant performance outcome effects for self-attributions on 6 of the possible 12 pairs. One of these revealed the predicted tendency to make greater ability attributions for poor performance, whereas 5 revealed a reversal of the prediction. Contrary to predictions, the elderly made more internal stable attributions for their own nurturance than for non-nurturance (Event 21, M = 5.58, vs. Event 23, M = 2.00), /(!!) = 8.95, p < .001, and (Event 22, M = 4.50, vs. Event 24, M = 3.25), r ( l l ) = 3.13, p < .01; for speed than for slowness (Event 13, M= 5.83 vs. Event 15, M = 4.83), J ( l l ) = 2.27, p< .05, and (Event 14, M = 5.00, vs. Event 16, M = 3.50), t(l 1) = 3.41, p < .01; and for strength (Event 9, M = 4.08) than for weakness (Event 11, M = 3.00), 1(11) = 2.57, p < .05. Consistent with predictions, the elderly were more likely to make Table 3 Ability Attributions for the Elderly That Reflect a Flattering (+) or Unflattering (-) View of the Elderly Attribution condition Domain Self Different-age other Same-age other Cognitive Good vs. poor memory Good vs. poor problem solving Physical Strength vs. weakness Speed vs. slowness Social Independence vs. dependence Nurturance vs. nonnurturance Note. An empty cell indicates no significant difference in attributions for good versus poor performance. Entries are collapsed across positive and negative consequences. ' Indicates a flattering view of the elderly for positive consequences and an unflattering view of the elderly for negative consequences. ability attributions for dependence (Event 20, M = 3.92) than for independence (Event 18, M = 2.75), «11) = 2.54, p < .05. Attributions for different-age others. For attributions made for different-age others, 7 out of 12 effects were significant. Contrary to predictions, 6 of these revealed greater ability attributions when they were flattering to the elderly (i.e., for good performance) than when they were unflattering. More specifically, ability was more likely to be selected as a cause for nurturance than for non-nurturance (Event 21, M = 6.0, vs. Event 23, M = 4.08), t( 11) = 4.8, p < .001, and (Event 22, M = 4.75, vs. Event 24, M = 2.83), 1(11) = 4.8, p < .001; for strength (Event 9, M = 4.25) than for weakness (Event 11, M = 3.08), «(!!) = 2.79, p < .05; for good problem solving (Event 5, M = 5.25) than for poor problem solving (Event 7, M= 3.33), /(!!) = 4.17, p<.0l; for independence (Event 17,Af = 5.25) than for dependence (Event 19, M = 4.17), t( 11) = 2.35, p < .05; and for speed (Event 13, M = 4.75) than for slowness (Event 15, M = 3.42), f(l 1) = 3.02, p < .05. The latter effect held true for positive consequences only. The one unflattering attribution was that ability was seen as a more likely cause for slowness (Event 16, M = 5.17) than for speed (Event 14, M = 2.75), when the event had negative consequences, t(\ 1) = 5.5, p < .001. Attributions for same-age others. For attributions for sameage others, there were four significant effects, all of which reflected greater ability attributions when they were flattering to the elderly than when they were unflattering. Ability was seen as a more 1 The primary focus of the within-subjects analyses was on attributions for the elderly, because this had not been examined in past research. For comparison purposes, the same analyses were conducted for the young adults. A similar pattern was found: The young were more likely to take or to be given credit for good performance than to be blamed for poor performance. For ability attributions, there were 24 significant effects, and 22 of these reflected a flattering view. For the forced-choice variable, there were 16 significant effects, and 14 of these showed a flattering view of the young. AGE DIFFERENCES IN ATTRIBUTIONS likely cause for nurturance (Event 22, M = 5.75) than for nonnurturance (Event 24,M = 4.75), t(\ 1) = 2.5,p< .05; for strength (Event 9, M = 5.42) than for weakness (Event 11, M = 3.25), ((11) = 5.17, p < .001; and for a good memory (Event 1, M = 5.50) than for a bad memory (Event 3, M = 3.42), ((11) = 3.92, p < .01. The latter effect held true for positive consequences, but was reversed for negative consequences (Event 4, M poor memory = 6.25) and (Event 2, M good memory = 4.83), J ( l l ) = 2.68,p<.05. Choice. The choice variable again was analyzed with t tests after arcsine transformations were performed on the proportions of those selecting the ability attribution. For attributions for the self there were no significant effects. There was one significant effect for attributions for others of the same age. A greater percentage chose ability as the cause of good problem solving (Event 5 = 64%) than of poor problem solving (Event 7 = 18%), ((oo) = 2.28, p < .05, which reflects a positive view of the elderly. Six effects were significant for attributions for different-age others. Two effects represent negative views of the elderly: A greater percentage chose ability as the cause of poor memory (Event 4 = 50%) than of good memory (Event 2 = 08%), r(oo) = 2.43, p < .05, and as the cause of slowness (Event 16 = 50%) than of speed (Event 14 = 0%), /(oo) = 2.82, p < .01. In contrast, a positive view of the elderly was reflected in the greater tendency by those making attributions for another of a different age to choose ability as the cause of good problem solving (Event 5 = 75%) than of poor problem solving (Event 7 = 17%), t(oo) = 3.04, p < .01; as the cause of nurturance than of non-nurturance (Event 21 = 92% vs. Event 23 = 42%), ((oo) = 2.82, p < .01 (Event 22 = 42% vs. Event 24 = 0%), f(oo) = 2.43, p < .05; and as the cause of independence (Event 17 = 75%) than of dependence (Event 19 = 17%), f(oo) = 3.04, p < .01. Discussion The findings were consistent with earlier research, which had revealed more unflattering attributions for performance by the elderly when compared to the young (Banziger & Drevenstedt, 1982; Blank, 1982), whereas there was also evidence for adaptive attributions for performance by the elderly. As in earlier research, poor cognitive or physical performance by the elderly was attributed more to inability than was identical performance by the young, whereas good cognitive or physical performance by the elderly was attributed less to ability than it was for the young. Although the attributions in all three conditions were equally unflattering to the elderly, an interesting pattern emerged when the three groups were compared. Those making self-attributions and attributions for another of the same age were more likely to rate inability as the cause for poor performance by the elderly than by the young. In contrast, the negativity of those making attributions for another of a different age was manifested in their attributions for good performance. They were less likely to rate ability as a cause for good performance by the elderly than by the young. In short, the young discounted good performance in the elderly, whereas the elderly confirmed expectancies for poor performance in themselves and in other elderly. In several ways the present results suggest a more favorable view of the elderly than has been found in previous research. First, only 11 out of the 72 possible differences in attributions 131 for the performance of elderly versus young adults were significant. Although this is beyond what would be expected by chance, the number is still small. Second, the negative view of the elderly occurred predominantly in the cognitive and physical domains; there were fewer negative outcomes in the social domain. An examination of attributions for good versus poor performance by the elderly also revealed a more favorable view of the elderly than had been expected on the basis of past research focusing on between-age-group comparisons. Here 14 out of the 17 significant effects showed a positive view of the elderly. Ability was seen as a more likely cause for good performance than inability was for poor performance. Good performance for the elderly was more likely to be attributed to ability, perhaps because it was generally unexpected and, therefore, distinctive.' These findings are similar to the results of another study by Lachman and Jelallian (1984), in which the elderly took credit for their successes but did not accept the blame for their failures. Rather than using a self-deprecating or helpless attributional style, the elderly, like the young, used a more adaptive, egotistical attributional style. Although the present study involved imagining one's successes and failures, the results were remarkably similar to the results of previous studies in which the elderly made attributions for their own actual performance (Lachman & Jelallian, 1984; Rutman & Freedman, 1984). Although valence of consequences did not appear to have an effect on the results, there was one noteworthy pattern. For the within-subjects comparisons, when the consequences of the events were positive the causal attributions for the performance of the elderly were always flattering. Thus, the elderly were always given credit for positive outcomes. In contrast, when the outcomes were negative, both unflattering (3 event pairs) and flattering (4 event pairs) attributions were made. What might account for the more favorable views of those making attributions about the elderly in the present study? One possible explanation is that the attribution targets were presented in a more realistic way. In the literature on attitudes toward the elderly, more favorable views are found when more realistic and detailed descriptions of the stimulus persons are provided (Lutsky, 1980). In the present study, more information about the targets was given than in previous research, and age was not made as salient. Moreover, the subjects were told that the targets were real people, whereas in past research the subjects were given a vague description of a fictitious person. In conjunction with the positive views of the elderly, there were still findings consistent with the negative view of the elderly found in past research. Although the number of significant effects was small, all but one was in the same direction as previous research: Compared to the young, internal attributions were more likely for the failures of the elderly and less likely for their successes. One possible explanation for this negative view is that in most cases poor performance is expected for the elderly. Some 3 A pilot study with 24 subjects examined whether the events were seen as more likely to occur (a) for a young person, (b) for an old person, or (c) equally likely for both young and old. For all items, 75% or more of the subjects rated poor performance as more likely to occur for the elderly and good performance as more likely to occur for the young. The only exceptions were for the nurturance items, for which the opposite pattern was found. 132 MARGIE E. LACHMAN AND LESLIE Z. McARTHUR research has suggested that expected outcomes tend to be attributed internally, and unexpected events are attributed externally (Deaux, 1976; Gekoski & Knox, 1983; Green, 1984; Rutman & Freedman, 1984). If this is so it could explain both past and present results, because when performance was good it was also unexpected for the elderly (e.g., solving a difficult problem, moving quickly). In contrast, when performance was poor it was also expected for the elderly (e.g., forgetting about a pot on the stove, driving slowly). In the present study all of the events depicting poor performances were expected for the elderly with the exception of nonnurturance (see Footnote 3). Even though non-nurturance was unexpected for the elderly it was still attributed internally, however, suggesting a negative bias rather than an expectancy effect. One challenge for future studies is to find examples of good performance other than nurturance that are expected for the elderly. This will be required in order to conduct a systematic test of whether attributions are based on expectancies for performance versus negative biases or stereotypes. Consistent with other studies (Banziger & Drevenstedt, 1984), the attributions about the elderly in the present study seemed to vary across tasks. What are the relevant task features that determine the nature of the attributions for performance by the elderly? The results shed some light on this question, but there is a need for further research aimed at examining task characteristics in more detail. In the present study, tasks were varied by domain, and as expected, the attributions were more favorable for the elderly in the social domain than in the cognitive or physical domains. In concert, the findings give reason to question the largely negative view of the elderly that has been presented in previous attribution studies. The present results suggest that there is indeed an age difference in attributional patterns that favors the young, but this seems to be consistent with expectancies about agingrelated performance decrements in the cognitive and physical domains. Despite this unflattering view, when the elderly are compared to the young, distinctions are made in attributions for various behaviors by the elderly, and they are more likely to receive credit for successes than they are to be blamed for failures. This distinction may be important for future examination of the affective and behavioral implications of attributions for young and elderly adults. Previous research has shown that providing the elderly with adaptive attributions through experimental manipulations can have a positive impact on performance (Prohaska, Parham, & Teitelman, 1984; Rodin & Langer, 1980). This suggests that attributions may be an important antecedent of performance changes in later life. Further work is needed to determine the antecedents of attributions and to examine whether performance changes have an impact on attributions by and about the elderly. References Abramson, L. Y., Setigman, M. E. P., & Teasdale, J. D. (1978). Learned helplessness: Critique and reformulation. Journal of Abnormal Psychology, 87, 49-74. Banziger, G., & Drevenstedt, J. (1982). Achievement attributions by young and old judges as a function of perceived age of stimulus person. Journal of Gerontology, 37, 468-474. Banziger, G., & Drevenstedt, J. (1984). Age as a variable in achievement attributions: The Weiner model and beyond. Basic and Applied Social Psychology, 5, 97-104. Blank, T. O. (1982). A social psychology of developing adults. New \brk: Wiley. Blank, T. O. (1984). Meaning and motivation in adult perceptions of causality. Basic and Applied Social Psychology, 5, 111 -120. Deaux, K.. (1976). Sex: A perspective on the attribution process. In J. H. Harvey, W. J. Ickes, & R. F. Kidd (Eds.), New directions in attribution research (Vol. 1, pp. 335-352). Hillsdale, NJ: Eribaum. Duval, S., & Wicklund, R. A. (1973). Effects of objective self-awareness on attribution of causality. Journal of Experimental Social Psychology, 9, 17-31. Gekoski, W. L., & Knox, V. J. (November, 1983). Identifying age stereotypes in adolescents: An attributional approach. Paper presented at meeting of the Gerontological Society of America, San Francisco. Green, S. K. (1984). Senility vs. wisdom: The meaning of old age as a cause for behavior. Basic and Applied Social Psychology, S, 105-110. Lachman, M. E., & Jelallian, E. (1984). Self-efficacy and attributions for intellectual performance in young and elderly adults. Journal of Gerontology, 39, 577-582. Lutsky, N. S. (1980). Attitudes toward old age and elderly persons. In C. Eisdorfer (Ed.), Annual review of gerontology and geriatrics (Vol. 1, pp. 287-336). New York: Springer. Prohaska, T. R., Parham, I. A., & Teitelman, J. (1984). Age differences in attributions to causality: Implications for intellectual assessment. Experimental Aging Research, 10, 111-117. Reno, R. (1979). Attribution for success and failure as a function of perceived age. Journal of Gerontology, 34, 709-715. Rodin, }., & Langer, E. (1980). Aging labels: The decline of control and the fall of self-esteem. Journal of Social Issues, 36, 12-29. Rutman, D. L., & Freedman, J. L. (1984, August). Age differences in causal attributions for success and failure experiences. Poster presented at the meeting of the American Psychological Association, Toronto, Ontario, Canada. Watson, D. (1982). The actor and the observer: How are their perceptions of causality divergent? Psychological Bulletin, 92, 682-700. Received May 31, 1985 Revision received September 25, 1985