The Explanatory and Predictive Scope of Self-Efficacy Theory

Journal of Social and Clinical Psychology, Vol 4, No i l')Ht>, pp SS9-573 THE EXPLANATORY AND PREDICTIVE SCOPE OF SELF-EFFICACY THEORY ALBERT BANDURA Stanford University Convergent evidence from the diverse lines of research reported in the present special issue of this journal attests to the explanatory and predictive generality of self-efficacy theory. This commentary addresses itself to conceptual and empirical issues concerning the nature and function of self-percepts of efficacy. Progress in understanding human functioning is best achieved by theo ries that have a broad range of applicability. To appeal to separate con ceptual schemes for each aspect of functioning does not produce much of an advance in the field. The diverse articles comprising the present special issue of this journal attest to the integrative scope of self-efficacy theory and its value in stimulating research that clarifies the contribu tion of self-referent thought to human motivation, affect, and action. CONVERGENT EVIDENCE FROM DIVERGENT TESTS The explanatory and predictive generality of a theory are strengthened by evidence that divergent procedures produce convergent results. The research reported in this issue presents a rich diversity of operations. Perceived self-efficacy is altered in varied ways by direct mastery ex periences; by social-comparative information conveyed through vicari ous modes of influence; and by social persuasion in the form of bogus feedback, attributional evaluations, and proffered incentives. The do mains of psychological functioning to which the theory has been ap plied are also remarkably diverse. The studies encompass, among other My own research as reported in this paper was supported by Public Health Research Grant No MH 5162-24 from the National Institute of Mental Health. Requests for reprints should be sent to Albert Bandura, Department of Psychology, Building 420, Jordan Hall, Stanford University, Stanford, CA 94305. 359

360 BANDURA things, motivational changes, affective reactions of stress and depres sion, psychosocial dysfunction, development of cognitive skills, achieve ment strivings and accomplishments, athletic feats, career choice and pursuits, and self-regulation of motivation and refractory behavior. The subject populations are of highly varied characteristics. Moreover, varied investigatory methodologies are employed, including induced changes in perceived self-efficacy under laboratory conditions; controlled field studies in which perceived self-efficacy was differentially altered by theoretically relevant influences; and multivariate correlational analyses of naturally occurring covariations. The research embodies another valuable source of diversity. When investigators rely entirely on a single measure of a postulated cognitive mediator, concerns arise as to whether the accumulated knowledge partly reflects the idiosyncratic contents of that specific set of items. In the methodology of the research under discussion, measures of per ceived self-efficacy are tailored to the domains of functioning being analyzed. The use of diverse indices of the self-efficacy mediator in creases the generality of findings. The evidence from this extensive multiform research is consistent in corroborating the dual causal linkage: Postulated determinants alter self-percepts of efficacy: and self-percepts of efficacy, in turn, affect motivation and action. Mediational analyses lend support to the causal contribution of self-percepts of efficacy (Bandura, 1986). The general ized replicability of the causal dependencies amply documents the com monality of the self-efficacy mechanism in human motivation and ac tion. MULTIDETERMINATION OF MOTIVATION AND ACTION The fact that perceived self-efficacy operates as a common mechanism in psychosocial functioning does not mean that other mechanisms do not also come into play. One must distinguish between theorizing and experimentation aimed at elucidating a particular mechanism governing behavior, and studies aimed at maximizing the amount of variance explained in behavior by combining a variety of factors that contribute to it. In social cognitive theory (Bandura, 1986), human behavior is governed by multiple determinants operating through varied mecha nisms. Thus, for example, cognitive self-motivation based on achieve ment standards is mediated by three different types of self-reactive influences operating in concert affective self-evaluation, perceived self- efficacy, and personal goal setting (Bandura & Cervone, in press). The

SCOPE OF SELF-EFFICACY THEORY 361 relative contribution of these self-reactive influences to motivation varies predictably, depending on the level and direction of the discrepancy between performance attainments and the comparative standard. Wur- tele's (this issue) mandate that perceived self-efficacy must exceed all other predictors of performance accords neither with self-efficacy theory nor with what is known about the dynamic operation of causal deter minants. ON RECHRISTENING AND CONVERSION TO REINFORCEMENT THEORY Kirsch (this issue) would like to rechristen \"perceived self-efficacy\" as \"expectancy for success in achievement situations\" and recast it within the framework of reinforcement theory. Neither the cumbrous relabel ing nor the conceptual baptism has much to recommend itself. In the standard lexicon, the term \"success\" is defined as \"a favorable or sat isfactory outcome or result.\" Unless terms are being used idiosyn- cratically, the label \"expectancy for success\" designates an outcome expectancy. Such a label is a misnomer when applied to judgments of operative capabilities. It is ironic that the ponderous appellation \"ex pectancy for success in achievement situations\" should be prescribed in the name of accuracy and succinctness. Serious confusions have been introduced into the expectancy litera ture by misconstruing the specifying criteria of a performance level as its outcomes. A \"performance\" is conventionally defined as \"an ac complishment\" or \"something done,\" and an \"outcome\" as \"some thing that follows as a result or consequence of an activity.\" Thus, if a given performance level is defined as a 6-foot jump, then a 6-foot leap is the realization of the performance, not the outcome that flows from it. Similarly, in assessments of academic performance, A, B, C, D, and F are the specifying criteria of performance level, not the outcomes. Remove the letter indicants of performance level, and one is left with an indefinite or undescribable performance. The social reactions, per sonal benefits, costs, and self-satisfactions anticipated for an A-level performance, or an F-level performance, constitute the outcome expect ancies. To conceptualize a performance level as the outcome of itself is to destroy the conventional meanings of performance and outcome. Kirsch prefers to view perceived self-efficacy from the perspective of reinforcement theory. As interest in the doctrine of reinforcement has waned, this type of theorizing has fallen on hard times in recent years. According to this approach, efficacy expectancies are reinforced, extinguished, and generalized in the likeness of responses. In contrast,

362 BANDURA social cognitive theory (Bandura, 1986), conceptualizes perceived self- efficacy in terms of self-referent judgments arrived at through cognitive processing of diverse sources of efficacy information. Because self-per cepts of efficacy are formed through acts of self-appraisal based on mul tidimensional information, perceived self-efficacy is more closely allied to the field of human judgment than to the subject of expectancy, which refers to an anticipation that something is likely to happen. The deter minants and processes of self-efficacy judgment are concerned with how information conveyed enactively, vicariously, persuasively, or physiologically is selected, weighted, and integrated into self-efficacy judgments. As regards causal determination, social cognitive theory seeks to clarify how self-efficacy judgment affects human action, thought, and affect, rather than treating perceived self-efficacy as a trait-like enti ty. Self-efficacy judgments influence human functioning through their impact on choice behavior, on effort expenditure and perserverence, on self-hindering or self-aiding thought patterns, and on affective and neurophysiological reactions to environmental demands. Kirsch expresses concern that the appearance of self-efficacy theory has \"obscured the considerable body of existing knowledge\" regarding expectancy for success and beliefs about environmental contingencies. Because a new line of theorizing and experimentation casts no veil, it can make existing literature neither more nor less visible. Kirsch sum marizes much of this early research on expectancy for success. Some of the studies touch on issues of interest to self-efficacy theory. How ever, because of the types of tasks and paradigms employed, much of this research is of limited generality at best, and is misleading at worst. In the prototypical expectancy research, people perform a simple task over a series of trials with varied feedback, and at periodic intervals they are asked to indicate what score they expect to get on the next trial. They are not told specifically what they will have to do next. Indeed, ambiguous tasks are usually employed so that attainment feedback can be varied independently of actual performance. Given the vagueness of prospective tasks and the unrelatedness of actual performance to feedback, performers' judgments are more likely to reflect extrapolative guesses from prior feedback than authentic self-appraisals of personal capabilities. To show that performance expectancies are highly sensitive to even small variations in feedback rates says more about the limitations of the task than about self-appraisal of capabilities from enactive ex perience. In the prototypical self-efficacy paradigm, people judge their ef ficacy in advance over a wide range of task demands within a mean ingful domain of functioning. This assessment procedure is designed to identify the pattern, strength, and upper limits of perceived self-

SCOPE OF SELF-EFFICACY THEORY 363 efficacy. Stringent performance tests are later conducted with varied tasks measuring generalized use of capabilities, rather than having peo ple repeatedly do essentially the same thing in the same situation under conditions shrouded in ambiguity except for the performance feedback. In the self-efficacy paradigm, people have to appraise the limits of their operative capabilities, whereas in the usual expectancy paradigm they typically set their expectancy slightly above what they have been told. Expectancy research that renders people heavily dependent on feed back as the indicant of capability will find that performance expectan cies closely mimic performance feedback. Feedback is not \"reinforcing\" anything; it simply tells subjects where to set their next estimate. Ex pectancy research using veridical feedback poses similar problems of generalizability. When people are asked to do the same thing repeated ly under unvarying circumstances, performance feedback will account for most of the variance in expectancy for a subsequent try. In everyday life, people must deal with changeable circumstances embodying uncertain and unpredictable elements. Appraisal of self- efficacy from performance feedback is complicated by the reality that many factors can affect how well one's capabilities are put to use. There fore, there is no simple equivalence of past performance to perceived self-efficacy. Appraisal of self-efficacy is an inferential process in which the relative contribution of ability and nonability factors to performance must be weighted. The extent to which people will alter their perceived self-efficacy on the basis of performance feedback will depend on such factors as the difficulty of the task, the amount of effort they expend, the amount of external aid they receive, the situational circumstances under which they perform, and their mood and physical state at the time. Moreover, performances are not compressed into a massed block of discrete trials in a barren setting stripped of contextual and personal factors except for instant performance feedback. Under usual circum stances, performances are separated in time, with some performances surpassing, others matching, and still others falling below one's typical attainments. Such temporal variability allows leeway for cognitive biases to distort how one's attainments are processed and coded for memory representation. To complicate the self-appraisal process further, the weight given to new experiences depends on the nature and strength of pre-existing self-efficacy into which they must be integrated. Self-percepts of efficacy are not simply reflective imprints of past action or performance feed back. This is shown in fine-grained analyses of performance attainments and changes in perceived self-efficacy at each step in the change process (Bandura, Reese, & Adams, 1982). Self-percepts of efficacy often ex ceed, only occasionally match, and sometimes remain below past per-

364 BANDURA formance attainments, depending on how performances are cognitively appraised. When people are fully assured of their capabilities, they remain unshaken in their perceived self-efficacy and persevere doggedly, even though they fail repeatedly (Brown & Inouye, 1978). Analyzing the multiple sources of efficacy information and how it is cognitively processed in forming self-efficacy judgments holds greater promise of furthering understanding the nature and function of self-referent thought than does reinforcing expectancies. The level-of-aspiration research of yesteryear, which Kirsch also reviews, provides another illustration of how trivial tasks have yielded some misleading generalizations. Most of these studies used simple tasks, such as tossing darts or rings, that call for little in the way of effort, mobilization of resources, and perseverance. In such situations, people generally set their expectancies and goals slightly above their immediately preceding performance level. When accomplishments require arduous effort over an extended period, as they usually do in everyday life, people do not judge that they can surpass each past accomplishment in an ever-rising series of triumphs. Lofty accomplishments achieved through sustained extra ordinary effort are not easily repeated or excelled. Feedback of having surpassed a demanding standard through laborious effort does not automatically strengthen perceived self-efficacy and raise aspiration (Bandura & Cervone, in press). It is true that some performers respond to their notable attainment by affirming a strong sense of self-efficacy and setting themselves even more challenging goals to accomplish. However, a sizable number are left with self-doubts that they can muster the same level of laborious effort again, and they set their sights on simply trying to match the standard they previously surpassed. Having driven themselves to success, some performers judge them selves inefficacious to repeat the demanding feat and lower their aspira tions. Clearly, much work remains to be done in this area in order to gain generalizable knowledge about the impact of human attainments on perceived self-efficacy and aspiration. EXPECTED FEAR AND AVOIDANT ACTION AS COEFFECTS OF PERCEIVED SELF-INEFFICACY Social cognitive theory posits an interactive, though asymmetric, rela tion between perceived self-efficacy and fear arousal, with coping ef ficacy exercising the greater sway (Bandura, 1982, 1986). Perceived self- inefficacy leads people to approach intimidating situations anxiously, and experience of disruptive levels of arousal may further lower their

SCOPE OF SELF-EFFICACY THEORY 365- sense that they will be able to perform well. However, people are much more likely to act on self-percepts of efficacy inferred from mastery experiences and social comparison of capabilities than to rely heavily on the indefinite stirrings of their viscera. This is because mastery ex periences and comparative appraisals are more reliable diagnostic in dicants of capability than affective arousal, which bears no uniform rela tionship to performance accomplishments. Moreover, whether or not perceived self-efficacy is affected by emotional arousal depends on how such information is cognitively processed. Many factors, including ap praisal of the source of arousal, the level of activation, the circumstances under which arousal is elicited, and personal experiences on how arous al affects one's performance, influence the efficacy meaning given to arousal. Thus, for example, accomplished actors interpret their anticipa tory stage fright as a normative situational reaction rather than as an indicant of inefficacy, and are not at all dissuaded by their agitated viscera from going on stage and performing what they assuredly know they theycan do once commence. Kirsch reconstrues measures of perceived self-efficacy in phobic domains of functioning as indicants of fear. His analysis rests on ques tionable premises and on an outmoded conception of \"skill.\" Before addressing the latter issues, let us examine the proposition that fear causes avoidance. Kirsch argues that people behave avoidantly because of expected fear. The notion that anticipatory fear controls avoidance behavior has been investigated extensively with diverse lines of research (Bandura, 1986; Bolles, 1975; Herrnstein, 1969; Schwartz, 1978). In some studies, feedback of autonomic arousal, which presumably constitutes the aver sive motivator, is removed surgically or blocked pharmacologically. In other studies, the occurrence of avoidance behavior is measured after fear arousal to threats has been thoroughly eliminated. In still other studies, changes in fear arousal are related to changes in avoidance behavior. The evidence is highly consistent in showing that avoidance behavior is not controlled by anticipatory fear. Kirsch never addresses the issue of how the paler expectation of fear can control behavior when the anticipatory actual experience of fear does not. To attribute avoidance behavior to expected fear simply begs the question, because the source of fearful expectations itself needs ex plaining. Fearful expectations are not sourceless. Kirsch does not in dicate where they come from. Within the conceptual framework of social cognitive theory (Bandura, 1986), it is mainly perceived inefficacy in coping with potential aversive events that gives rise to both fearful expectations and avoidance behavior. People who judge themselves as efficacious in managing potential threats neither fear nor shun them.

366 BANDURA In contrast, if people judge themselves as inefficacious in exercising control over potential threats, they view then anxiously, conjure up possible calamities, and are unwilling to have any commerce with them (Bandura, 1983). In short, behind expected fears and calamities, and the unwillingness to try coping tasks, lie judgments of personal inef ficacy to exercise control over risky situations. Fearful expectations and avoidance behavior are thus largely coeffects of perceived coping inef ficacy. Consider now Kirsch's correlational argument for reconstruing self- efficacy measures in transactions involving threats. Kirsch reasons that because perceived self-inefficacy in coping with phobic objects and an ticipated fear correlate positively with each other and with avoidance behavior, they must be measuring the same thing namely, fear. Nei ther the conclusion of equivalence nor the conferral of priority on fear necessarily follows from such correlations. For example, in a school population, age and weight are highly correlated with each other and both may correlate with a third factor, but one would hardly conclude from such evidence that indices of age and weight measure the same construct. Williams and his colleagues (Williams, Dooseman, & Kleifield, 1984; Williams, Turner, & Peer, 1985) have analyzed by partial correlation several data sets from experiments in which perceived self-efficacy, expected fear, and phobic behavior were measured. Perceived self- efficacy retains its predictiveness of phobic behavior when expected fear is partialed out, whereas the relationship between expected fear and phobic behavior essentially disappears when perceived self-efficacy is partialed out. Contrary to what Kirsch claims, perceived self-efficacy and expected fear are not measuring the same thing. The variance con tribution of perceived self-efficacy may be reduced in pretreatment assessments if the analyses are confined only to the severest cases, which markedly curtails the range of self-efficacy scores. Williams and Watson (1985) similarly demonstrate the predictive superiority of per ceived self-efficacy over perceived dangerous consequences and level of fear arousal associated with performance of threatening activities. When perceived dangers and performance fear are controlled, per ceived self-efficacy accounts for a substantial amount of variance in phobic behavior, whereas perceived dangers and performance fear have no predictive value when perceived self-efficacy is partialed out. It is interesting to speculate on why the belief that anticipatory fear triggers avoidance behavior remains firmly entrenched in psychological thinking, despite substantial evidence to the contrary. A possible answer may lie in the force of confirmatory biases in judgments of causality (Nisbett & Ross, 1980). Confirming instances in which fear and avoid-

SCOPE OF SELF-EFFICACY THEORY 367 ance occur jointly are likely to remain highly salient, whereas noncon- firming instances in which fear and approach behavior occur jointly, or avoidance occurs without fear, command less attention. It is not that the nonconfirming instances are any less prevalent. Quite the contrary is true: People commonly perform activities at low strengths of self- judged efficacy, despite high fear arousal. Thus, for example, actors strut on stage and students take intimidating examinations, although they may be plagued by aversive anticipatory arousal. Similarly, people regularly take self-protective action without having to wait for expec tations of fear to impel them to action. They disconnect electrical ap pliances before repairing them without having to conjure up fears. These different types of disconfirming occurrences tend to be ignored in judging the relation between fear and avoidance. Kirsch's conclusions regarding the implications of social induce ments on self-judged efficacy rest on a mistaken premise. He seems to assume that if one can strengthen people's beliefs in their efficacy by the prospect of escalated benefits, such evidence challenges the status of self-efficacy measures in phobic domains of functioning, as if self- efficacy judgments are authentic only if unalterable by social influences. In point of fact, judgment of self-efficacy is not an immutable entity reflecting a fixed faculty of the organism. People alter judgments of their efficacy on the basis of direct mastery experiences; social comparisons through vicarious influences; inferences from bodily states; and varied forms of social persuasion, including bogus feedback of attainment, arbitrary attributional interpretations, and monetary lures (Bandura, 1986). Kirsch cites his study in which hypothetical escalated benefits per suaded students who said they feared snakes that they could handle a snake or toss a wad of paper into a distant wastepaper basket. As the inducements were raised to $1 million, saving another's life, or sparing one's own life, eventually all the students persuaded themselves that they could handle a snake, and many of them similarly persuaded themselves that they could marshal sufficient dexterity to hit a waste- paper basket at some distance (46%) or at 50 feet (24%). Kirsch concluded from these responses to the pretended activity that self-efficacy judg ments regarding shifty snakes reflect expected fear rather than skill. The results of this hypothetical exercise have little bearing on the nature of perceived self-efficacy. Self-efficacy scales do not measure skill; they measure what people believe they can do under varied cir cumstances whatever skills they possess or the particular skills required by the task. How well people perform on a task partly depends on their beliefs about how well they will be able to orchestrate the subskills and cognitive resources they possess, and about how much effort they will

368 BANDURA be able to mount and sustain in a given endeavor. Thus, they perform poorly if persuaded by bogus feedback that they are physically inef ficacious, but they perform at a high level if persuaded that they are physically efficacious (Weinberg, Gould, & Jackson, 1979). The illusorily instated self-efficacy even overrides pre-existing sex differences in phys ical stamina. Students persevere doggedly when arbitrarily high ref erence points positively bias their self-efficacy judgments as problem solvers, but they abort their efforts quickly when arbitrarily low reference points bias their self-efficacy judgments negatively (Cervone & Peake, 1986). Even training in a skill that can impair functioning produces beneficial results when presented through bogus feedback in ways that enhance perceived self-efficacy (Holroyd et al, 1984). As these and many other experiments reveal, the same capability can spawn performances that are subpar, ordinary, or extraordinary, depending on the influen tial contribution of self-judged efficacy. Judgments of operative self- efficacy are concerned not with the skills one has, but with beliefs about what one can do with the subskills one possesses in dealing with con tinuously changing realities, most of which contain ambiguous, unpre dictable, and stressful elements. If Kirsch wishes to conceptualize a self- efficacy judgment as an unalterable reflection of a fixed skill, then that should be regarded as his conception to defend, rather than burdening self-efficacy theory with such a notion. Although perceived self-efficacy is concerned with perceived opera tive capability, not inherent skill, one might examine in passing Kirsch's rationale for the choice of tasks in the preceding study. Kirsch (1982) is of the view that coping with a shifty reptile involves no skill, whereas tossing a wad of paper into a wastepaper basket does. He rests his argument on an elementalistic conception of \"skill\": Since the con stituent motor responses of locomotion, grasping, and the like exist within the behavioral repertoires of snake phobics, presumably no skill is involved in coping with reptiles. This is analogous to arguing that, because people possess a large stock of words in their verbal repertoire, neither verbal discourse nor written composition involves skill. Human skill is a generative capability requiring variable utilization of multiple subskills, not a fixed entity. Neither theorizing nor experimentation on how perceived self- efficacy contributes to human endeavors and accomplishments is well served by false dichotomies that coping with threats involves no skill, but tossing a wad of paper does. Isn't it time to retire the conception of \"skill\" as a fixed faculty and the trite arguments that flow from it? Herpetologists come to deal routinely with deadly reptiles not by ex tinguishing their \"expected fear\" or infusing them with \"willingness,\" but by building their sense of coping self-efficacy through modeling and mastering sets of strategies for exercising control over reptiles, whatever

SCOPE OF SELF-EFFICACY THEORY 369 situation might arise. When severe snake phobics verbalize aloud their thoughts while attempting to cope with a snake at the beginning of treatment, judgments of their coping capabilities determine what they are willing to risk (Bandura, 1983). They shun activities when they believe their inept coping efforts will provoke adverse reactions. They promptly abort interactions they have initiated when they do not know- how to deal with unpredictable situations that ensue. They are reluctant to attempt tasks they might be able to handle, because they believe the encounter will escalate to a level that will exceed their coping capabili ties. The most profound level of self-inefficacv involves perceived vul nerability to total loss of personal control, which they believe will leave them defenseless. An efficacy-oriented treatment that restores a strong sense of self-efficacy by conveying effective coping strategies rapidly eliminates phobic thinking, expected fear, and phobic avoidance. Pho bics would be ill served by the view propounded by Kirsch that their problems stem from a lack of \"willingness\" and that no coping tech niques need be offered them because their \"behavioral repertoire\" already contains the necessary \"motor responses.\" Kirsch makes much of the fact that the students were more per suadable regarding coping with a reptile than tossing a wad of paper into a small remote target. There is little of conceptual import here. All this shows is that it is difficult to persuade people that thev can execute a performance with little margin of error on a task set at or near a physically unattainable limit. As regards self-efficacy for pretended snake handling, the findings unsurprisingly show that escalated bene fits can raise self-efficacy judgments among subjects selected by a self- report criterion known to be deficient for identifying the types of severe phobics used in self-efficacy research. Among people who report thev dread snakes and seek help for the problem, almost 40% perform boldly when administered an actual behavioral test (Bandura, Blanchard, & Ritter, 1969; Bandura, Jeffery, & Wright, 1974). Thus, selecting subjects on the basis of reported fear yields about a 35% false-positive rate. In Kirsch's student sample, chosen according to self-reported fear, slightly over 40% did not require strong inducement to raise their self-efficacy. However, the self-efficacv judgments of the remaining students were unusually resistant to change; they yielded only to extraordinary in ducements, considering that the students were involved only in a pre tended situation. Kirsch cites the study by Valins and Ray (1967), which was similarly flawed by self-report selection, as showing that monetary inducements weaken avoidance behavior. When phobics are chosen by a behavioral test, they display a refractory distrust of their coping efficacy that even real money fails to dislodge in situations requiring transactions with real snakes (Rimm & Mahoney, 1969). It will be recalled from the earlier discussion that self-efficacy theory

370 BANDURA regards social persuasion as one means of raising people's beliefs con cerning their operative capabilities. To show that the prospect of large benefits can lead students to persuade themselves that they might be able to mount an extraordinary coping effort, which they need not even perform, corroborates predictions from self-efficacy theory that judg ments of self-efficacy are influenceable by social persuasion. Consider a familiar example. In their life pursuits, countless young athletes per suade themselves, goaded by the prospect of fame and fortune, into believing they possess the capabilities to make it into the professional ranks. For the vast majority, the self-belief is ill founded, but it sustains long hours of grueling practice under miserable conditions for years on end. Except for attainments that clearly exceed human capacity, there are numerous difficult things people can persuade themselves they could conceivably do for prized benefits. Kirsch seriously distorts the conception of self-efficacy when he claims that it refers solely to perceived capability independent of task demands. Self-efficacy theory has never tendered the preposterous no tion that perceived self-efficacy is an autistic entity that is identifiable independently of the nature and complexity of performance tasks. This misrepresentation is especially ironic, because self-efficacy theory has explicitly argued against omnibus trait conceptions of self-efficacy dis joined from the nature and complexity of the activity in question. In point of fact, the standard assessment methodology measures self-per cepts of efficacy in relation to task demands ordered in difficulty or complexity. Level of perceived self-efficacy is defined in terms of whether a person's self-judged efficacy is limited to simple tasks, whether it extends to moderately difficult ones, or whether it includes even the most taxing performances within a particular domain of functioning. Because perceived self-efficacy is measured for different levels of task demands spanning a wide range, the assessment procedure reveals variations in perceived self-efficacy for the same set of task difficulties. Locke, Motowidlo, and Bobko (this issue) show how self-efficacy meas ures bring order to conflicting findings across studies arising from ex- pectancy-of-success measures that confound expectancy with task dif ficulty. In this same connection, self-efficacy theory regards fear and stress as products of a relational condition the fit between perceived coping capability and perceived task demands that carry negative con sequences if unfulfilled. Kirsch further misrepresents self-efficacy theory when he alleges that the theory postulates low perceived self-efficacy as the cause of fear, irrespective of the domain of activity. He points out that perceived inefficacy regarding innocuous laboratory tasks does not make subjects fearful. I cannot imagine why self-inefficacy on such tasks would pro-

SCOPE OF SELF-EFFICACY THEORY 371 voke fits of fear. The fact is that self-efficacy theory explicitly posits perceived self-inefficacy to exercise control over potentially aversive events as the source of fear and anxiety. This important domain qualifier is not something that can be easily overlooked, since it is italicized both in an early publication addressing this very issue (Bandura, 1978), and in a later publication (Bandura, 1982) explaining that perceived self- inefficacy to control aversive outcomes is central to anxiety, whereas perceived self-inefficacy to control valued outcomes is likely to give rise to despondency. PERCEIVED SELF-EFFICACY VERSUS EFFORT-PERFORMANCE EXPECTANCY Some expectancy-value theories include an expectancy that effort will beget requisite performances (Vroom, 1964). It should be noted, how ever, that perceived self-efficacy encompasses much more than effort determinants of performance. Effort is but one of many factors that govern the magnitude and quality of performance. People judge their capacity for challenging activities more in terms of their perceptions of the knowledge, skills, and strategies they have at their command than solely in terms of how much they will exert themselves. Performances that call for ingenuity, resourcefulness, and adaptability depend more on adroit use of skills and specialized knowledge than on effort. People who cope poorly with stress expect that marred performances in intim idating situations will be determined by their self-debilitating thought patterns rather than by how much effort they mount. Indeed, the harder they try, the more they may impair their execution of the activity. Expectancy theorists have probably singled out effort as the sole cause of performance accomplishments because the theory has usually been concerned with how hard people work at routine activities. Hence, the aspect of self-efficacy that is most germane to how much is ac complished is people's perceived perseverant capabilities that is, their assurance that they can exert themselves sufficiently to attain desig nated levels of productivity. MULTIDOMAIN SELF-EFFICACY VERSUS OMNIBUS GLOBAL SELF-EFFICACY Self-perceptions of efficacy vary across different activity domains, dif ferent levels of demands within activity domains, and different environ mental circumstances of performance. Therefore, the role of perceived

372 BANDURA self-efficacy in psychosocial functioning is best elucidated by self-efficacy measures tailored to particular domains of functioning, rather than as global disposition assessed by an omnibus test. Such omnibus tests include a fixed set of items, many of which may have little relevance to the domain of functioning being analyzed. It is unrealistic to expect such all-purpose tests to predict with appreciable accuracy how people will perform different activities under diverse circumstances. Indeed, domain self-efficacy scales predict better than global tests (Bandura, 1986; in press). Domain scales do not mean that self-efficacy items must be cast in minute particulars. Rather, the items are constructed at an intermediate level of generality representing a generic level of competence at each aspect of a domain. Thus, for example, individuals are asked to judge their perceived efficacy to cope with congested city traffic rather than with traffic on a specific street in a specific city. Nor do domain-linked scales mean that there is no generality to perceived capability. Multi- domain measures reveal the patterning and degree of generality of people's sense of personal efficacy. Some may judge themselves as highly efficacious across a wide range of domains and levels of task demands within each domain; others may judge themselves as inef ficacious in most domains of functioning; and many may judge them selves as relatively efficacious in domains in which they have cultivated their competencies, moderately efficacious in domains with which they are somewhat less conversant, and inefficacious in activity domains foreign to them. One can derive the degree of generality from multi- domain scales, but one cannot extract the patterning of perceived self- efficacy from conglomerate omnibus tests. The research reported in this special issue documents the promise of the microlevel approach for achieving broad understanding of how self-percepts of efficacy affect human functioning. REFERENCES Bandura, A (1978). Reflections on self-efficacv. In S. Rachman (Ed.), Advances in behaviour icscarch and therapy (Vol. 1, pp. 237-2r>Li) Oxford. Pergamon Press Bandura, A. (1482) Self-efficacy mechanism in human agency. American Psychologist, 37, 122-147 Bandura, A (1983). Selt-efficacy determinants ot anticipated fears and calamities fournal of Personality and Social Psychology, 45, 464-469. Bandura, A (1986). Social foundations of thought and action: A social cognitive theon/. Engle wood Cliffs, NJ: Prentice-Hall Bandura, A. (in press). Self-efficacy mechanism in physiological activation and health- promoting behavior. In J. Madden IV, S Mattvsse, & J. Barchas (Eds.) Adaptation. learning and affect New York: Raven Press.

SCOPE OF SELF-EFFICACY THEORY 373 Bandura, A., Blanchard, E. B., & Ritter, B. (1W1) Relative efficacy of desensitization and modeling approaches lor inducing behavioral, affective, and attitudinal changes fournal of Personality and Social Psychology, 13, 173-199. Bandura, A., & Cervone, D. (in press). Differential engagement of self-reactive influences in cognitive motivation. Organizational Behavior and Human Decision Processes. Bandura, A., Jeffery, R. W , & Wright, C. L. (1974) Efficacy of participant modeling as a function ot response induction aids, lournal of Abnonnal Psychology, 83, 56-64. Bandura, A , Reese, L., & Adams, N. E (1982). Microanalysis ot action and fear arousal as a function of differential levels of perceived self-efficacv Journal of Pcisouality and Social Psychology, 43, 5-21. Bolles, R. C. (1975), Learning theory. New York: Holt, Rinehart & Winston. Brown, I Jr , & Inouye, D. K. (1978). Learned helplessness through modeling. The role of perceived similarity in competence, lournal of Personality and Social Psychology, 36, 900-908. Cervone, D., & Peake, P. K. (1986) Anchoring, efficacy, and action: The influence of judgmental heuristics on self-efficacy judgment and behavior. Journal of Personality and Social Psychology, 50, 492-501, Herrnstein, R J. (19h9) Method and theory in the study of avoidance. Psychological Revieiv, 76, 49-69. Holroyd, K A., Penzien, D. B., Hursey, K. G , Tobin, D. L., Rogers, L., Holm, | W., Marcille, P. J., Hall, J. R., & Chila, A. G, (1984) Change mechanisms in EMG biofeedback training. Cognitive changes underlying improvements in tension head ache lournal oj Consulting and Clinical Psychology, 52, 1039-1053. Kirsch, I. (le>82) Efficacy expectations as response predictors The meaning of efficacy ratings as a function of task characteristics. Jounial of Personality and Social Psychologi/, 42, 132-136 Nisbett, R., & Ross, L. (1980) Human inference: Strategics and shortcomings of social judgment. Englewood Cliffs, NJ' Prentice-Hall. Rimm, D. C, & Mahoney, M J. (1969). The application of reinforcement and participant modeling procedures in the treatment of snake-phobic behavior. Behaviour Research and Therapy, 7, 369-376. Schwartz, B. (1978). Psychology of learning and behavior. New York: Norton. Valins, S , & Rav, A. A. (1967). Effects of cognitive desensitization on avoidance behavior. Journal of Personality and Social Psychology, 7, 345-350. Vroom, V. H. (1964). Work and motivation. New York: Wiley Weinberg, R. S., Gould, D., & Jackson, A (1979). Expectations and performance: An empirical test of Bandura's self-efficacv theory. Journal of Sport Psychologi/, 1, 320-331. Williams, S L., Dooseman, G., & Kleifield, E. (1984). Comparative power of guided mastery and exposure treatments for intractable phobics journal of Consulting and Clinical Psychology, 52, 505-518. Williams, S. L., Turner, S. M., & Peer, D. F. (1985). Guided mastery and performance desensitization treatments for severe acrophobia. Journal of Consulting and Clinical Psychology, 53, 237-247 Williams, S. L., & Watson, N. (1985). Perceived danger and perceived self-efficacy as cognitive mediators of acrophobic behavior. Behavior Therapy, 16, 136-146