Journal of Personality and Social Psychology1989, Vol. 56, No. 3.407-415

Copyright 1989 by the American Psychological Association, Inc.0022-3514/89/S00.75

Impact of Conceptions of Ability on Self-Regulatory Mechanisms andComplex Decision Making

Robert WoodUniversity of New South Wales

Kensington, New South Wales, Australia

Albert BanduraStanford University

Tested the hypothesis that induced conceptions of ability as a stable entity or as an acquirable skillwould affect self-regulatory mechanisms governing performance in a simulated organization. Ssserved as managerial decision makers in which they had to match employees to subfunctions and todiscover and apply managerial rules to achieve a difficult level of organizational performance. Thosewho performed the challenging managerial task under an entity conception of ability suffered a lossin perceived self-efficacy, lowered their organizational goals, and became less efficient in their analyticstrategies. Ss who managed the organization under an acquirable skill conception of ability sustainedtheir perceived self-efficacy, set challenging organizational goals, and used analytic strategies effec-tively. These divergences in self-regulatory factors were accompanied by substantial differences inorganizational performance. Path analysis revealed that perceived self-efficacy had both a directeffect on organizational performance and an indirect effect through its influence on analytic strate-gies. Personal goals also affected organizational performance through the mediation of analytic strat-egies. The relation of prior organizational performance to subsequent performance was mediatedentirely by the combined influence of the self-regulatory factors.

Complex decision making is, by its nature, a motivated, cog-nitive process. This is especially true in dynamic organizationalenvironments. Decision making in such environments is an on-going process that requires complex integration of multiplesources of information to produce distal, socially mediated out-comes (Mintzberg, 1973; Stewart, 1967). Because organiza-tional outcomes must be achieved through the concerted effortsof others, some of the most important managerial decisions areconcerned with how to use human talent and how to guide andmotivate human effort. In executing this role, managers have tocope with numerous obstacles, failures, and setbacks that oftencarry perturbing self-evaluative implications as well as socialconsequences. These affective factors can undermine self-con-ceptions and motivation in ways that impair good use of deci-sion-making skills. Effective decision making thus involvesmore than applying a set of cognitive operators to existingknowledge for desired solutions. Self-regulatory influences haveconsiderable impact on how well cognitive-processing systemswork (A. Bandura, 1986; Wood, Bandura, & Bailey, in press).

The mechanisms and outcomes of managerial decision mak-ing do not lend themselves readily to experimental analysis in

This research was supported by a University of New South WalesResearch Board Grant and by Public Health Research Grant MH-5162-25 from the National Institute of Mental Health. We are gratefulto Mary-Anne Rogers for her able assistance with this research and toDavid Jemison and Tom Lenz for their help in arranging the subjectpool.

Correspondence concerning this article should be addressed to Rob-ert Wood, AGSM, University of New South Wales, P.O. Box 1, Kensing-ton, New South Wales, Australia 2033, or to Albert Bandura, Depart-ment of Psychology, Building 420, Jordan Hall, Stanford University,Stanford, California 94305.

actual organizational settings. The governing processes are usu-ally influenced by a multiplicity of interacting factors that aredifficult to identify and over which it is even more difficult toexercise experimental control. Advances in this complex fieldcan be achieved by experimental analyses of decision makingin simulated organizational environments. A simulated envi-ronment permits systematic variation of theoretically relevantfactors and precise assessment of their impact on organizationalperformance and the psychological mechanisms through whichthey achieve their effects. Much of the research on human deci-sion making involves single trial judgments in static environ-ments (Beach, Barnes, & Christensen-Szalanski, 1986; Ho-garth, 1981). Judgments under such conditions may not providea sufficient basis for developing either descriptive or normativemodels of decision making in dynamic naturalistic environ-ments that involve sequential judgments governed by learningand motivational mechanisms. By incorporating multiple trialsin the simulated environment it is possible to examine temporalinterdependencies and cumulative effects in decision-makingprocesses (Wood & Bailey, 1985).

The conception of ability with which people approach com-plex activities is likely to have a significant impact on the self-regulatory influences that govern ongoing motivation and per-sonal accomplishments in complex decision-making environ-ments. Recent research has identified two major conceptions ofability to which people subscribe (M. Bandura & Dweck, 1987;Dweck & Elliott, 1983; Nicholls, 1984), In one perspective, theyconstrue ability as an incremental skill that can be continuallyenhanced by acquiring knowledge and perfecting one's compe-tencies. People with this conception adopt a learning goal. Theyseek challenging tasks that provide opportunities to expandtheir knowledge and competencies. Errors are regarded as a nat-ural, instructive part of an acquisition process. They judge ca-

407

408 ROBERT WOOD AND ALBERT BANDURA

pabilities more in terms of personal improvement than by com-parison against the achievement of others.

In the contrasting perspective, ability is construed as a moreor less fixed entity. This type of conception of ability heightensevaluative concerns about personal competence that can havediverse effects on cognitive functioning (Nicholls, 1984). Be-cause performance level is regarded as diagnostic of intellectualcapacity, errors and deficient performances carry personal andsocial evaluative threats. Therefore, people adopting the entityview tend to pursue performance goals of demonstrating theircompetence. They prefer tasks that minimize errors and permitready display of intellectual proficiency at the expense of ex-panding their knowledge and learning new skills. High effort,which is often required to develop competencies in complex ac-tivities, also poses evaluative threats because high effort is takenas indicative of low ability. An entity conception of ability is lessconducive to effective management of failure than is the view ofability as an incremental skill (Elliott & Dweck, 1988).

According to social cognitive theory (A. Bandura, 1986,1988), self-regulation of motivation and performance attain-ments is governed by several self-regulatory mechanisms oper-ating in concert. They include affective self-evaluation, per-ceived self-efficacy for goal attainment, and personal goal set-ting. Perceived self-efficacy refers to beliefs in one's capabilitiesto mobilize the motivation, cognitive resources, and courses ofaction needed to meet given situational demands. Self-beliefs ofefficacy affect the challenges that are undertaken, the amountof effort expended in an endeavor, the level of perseverance inthe face of difficulties, whether thinking patterns take self-aid-ing or self-impeding forms, and vulnerability to stress and de-pression.

Among the different modes of altering self-beliefs of efficacy,performance experiences are especially influential. Substan-dard performances are likely to carry markedly different diag-nostic implications depending on whether ability is construedas an acquirable skill or as a relatively stable entity. When per-formances are viewed as skill acquisition in which one learnsfrom mistakes, perceived self-efficacy is unlikely to be adverselyaffected by substandard performances. This is because errorsbecome a normative part of any acquisition process rather thanserving as indicators of basic personal deficiencies. In contrast,when performances are construed as being diagnostic of under-lying cognitive capability, frequent experience of substandardperformances can take a heavy toll on perceived self-efficacy.

Perceived managerial self-efficacy can influence organiza-tional performance both directly and indirectly through itseffects on personal goal setting and use of analytic strategies.The stronger the perceived self-efficacy, the higher the goals peo-ple set for themselves (A. Bandura & Cervone, 1986; Taylor,Locke, Lee, & Gist, 1984; Wood et al., in press) and the strongerthe commitment to them (Locke, Frederick, Lee, & Bobko,1984). The findings of diverse lines of research show that chal-lenging goals enhance motivation and performance attainments(Latham & Lee, 1986; Locke, Shaw, Saari, & Latham, 1981;Mento, Steel, & Karren, 1987) especially if combined with per-formance feedback (A. Bandura & Cervone, 1983).

The multifaceted nature of managerial activities and theirmazelike linkage to organizational performance introducescomplexities in the relation between personal goals and group

attainment (Wood et al., in press). In virtually all previous re-search, self-set goals are applied to personal performances overwhich individuals can exercise direct control by regulating theirlevel of effort. In organizational environments, managerialgoals must be socially mediated through the coordinated effortsof others. Sheer managerial effort alone does not ensure attain-ment of a collective goal. To complicate matters further, effortsto enhance the level of organizational functioning often requireconstituent changes in particular aspects of the social structureand the way in which social resources are allocated. Systematicpursuit of such operational subgoals contributes to eventualsuccess but does not necessarily produce sizable gains in organi-zational performance in the short run (Wood et al., in press).So prediction regarding the performance-enhancing effects ofgoal challenges at the group level must be tempered by consider-ations of these complexities.

In complex decision-making environments, appropriate de-cision rules are discovered through systematic application ofanalytic strategies (Bourne, 1965; Bruner, Goodnow, & Austin,1956; Wood et al., in press). Initially, performers must drawon their existing state of knowledge in constructing tentativecomposite rules for how various motivational factors may affectoutcomes. The optimal value of these different factors mustthen be tested by varying them one at a time and assessing howthey affect performance outcomes. Less skilled decision makersformulate vague composite rules, tend to alter several factorsconcurrently (making it difficult to assess the source of multiplyproduced effects), and make less effective use of informativeoutcome feedback.

Self-referent factors can affect how well analytic subfunctionsfor rule learning, such as hypothesis construction and testingand inductive reasoning, are executed. This is especially true inthe acquisition of complex functional rules, which place heavydemands on efficient cognitive processing of multidimensionalinformation (Wood et al., in press). In the formal characteristicsof the present managerial task used in this study, subjects hadto learn the form of the functions relating several motivationalfactors to aggregate outcomes. Some of the factors involvednonlinear and compound rules that are more difficult to learnthan are linear ones (Brehmer, Hagafors, & Johansson, 1980).Moreover, they had to figure out the best way to integrate theset of rules and to apply them discernibly to each member ofthe group. To achieve all of this, they had to generate hypothesesabout functional relations for different motivational factors, totest their judgments against outcome information, and to re-member which of the hypotheses they had tested and how wellthey had worked.

It requires a strong sense of efficacy to deploy one's cognitiveresources optimally and to remain task oriented in the face ofrepeated difficulties and failures. Those who judge themselvesto be inefficacious in coping with environmental demands tendto become more self-diagnostic than task-diagnostic (M. Ban-dura & Dweck, 1987). They dwell on their personal deficienciesand cognize environmental demands as being more formidablethan they really are. Such self-referent intrusive thinking cre-ates stress and undermines effective use of capabilities by divert-ing attention away from how best to proceed to concerns overpersonal deficiencies and possible adverse outcomes (Lazarus &Launier, 1978; Meichenbaum, 1977; Sarason, 1975).

ABILITY CONCEPTIONS AND DECISION MAKING 409

We designed this experiment to test hypotheses that concep-tions of ability will influence achieved levels of organizationalperformance through their effects on the mediating self-regula-tory mechanisms just described. Subjects engaged in manage-rial decision making in a simulated organization in which theyhad to match a set of employees to job requirements and to usegoals, instructive feedback, and rewards in appropriate ways toachieve substantial gains in productivity that were difficult tofulfill. They performed the managerial task over a series of trialsunder instated conceptions of ability either as an acquirableskill that is improvable with practice or as a basic intellectualentity reflecting underlying cognitive capacities. At three pointsin the managerial simulation, we assessed subjects' perceivedmanagerial self-efficacy and personal goals. We also measuredthe adequacy of their analytic strategies for discovering manage-rial rules and the level of organizational performance theyachieved.

For reasons given earlier, we predicted that perceived mana-gerial self-efficacy would be sustained under the conception ofability as an acquirable skill but impaired under the entity con-ception. We further predicted that perceived self-efficacy wouldenhance organizational performance both directly and indi-rectly by its effects on personal goal setting and on use of ana-lytic strategies. The stronger the perceived self-efficacy, the morechallenging the organizational goals subjects would set forthemselves and the more systematically they would use strate-gies to discover the managerial rules. High self-set goals andsystematic strategies would, in turn, enhance the level of organi-zational performance.

Method

Subjects

The subjects were 20 men and 4 women from a graduate program inbusiness studies. Their average age was 26 years (SD = 3.16 years). Fif-teen subjects had prior managerial experience. We randomly assignedsubjects, balanced for sex, to the experimental conditions.

Simulated Organization

The study was presented to the subjects as a project in managerialdecision making in which they would manage a simulated organization.The introductory information described the simulation as one in whichmanagers receive weekly orders for the production of furniture items,along with a roster of available employees. The manufacture of the itemsin each of the weekly orders required five different production subfunc-tions, such as milling the timber, assembling the parts, staining and glaz-ing the assembled frame, upholstering the furniture, and preparing theproducts for shipment. Subjects managed the organizational unit for atotal of 18 production orders, with each order representing a perfor-mance trial in the simulation.

The subject's managerial task was to allocate workers from a five-member roster to the different production subfunctions so as to com-plete the work assignment within an optimal period. By correctlymatching employees to job requirements, subjects could attain a higherlevel of organizational performance than if employees were poorlymatched to jobs. To assist them in this decision task, subjects receiveddescriptions of the effort and skill required for each of the productionsubfunctions and the characteristics of each employee. This informa-tion described their particular skills, experience, motivational level,preference for routine or challenging work assignments, and standards

of work quality. Both the production subfunctions and the employeeattributes were selected on the basis of extensive observation of the ac-tual manufacturing process. The employee profile descriptions wereprovided at the beginning of the simulation, but subjects could refer tothem at any time during their organizational decision making.

In addition to allocating employees to jobs, subjects had to make sev-eral decisions about the use of various motivational factors to optimizethe group's performance. They had to decide how to use goals, instruc-tive feedback, and social rewards to enhance the job accomplishmentsof each employee. For each of these motivational factors, subjects hada set of options representing the types of actions that managers couldtake in an actual organization. We used a mathematical model to calcu-late the hours taken to complete a production order on the basis of theadequacy of subjects' allocation of employees to jobs and their use ofthe three motivational factors. The group performance for each trialwas reported to subjects as a percentage of a preset standard number ofhours to complete each manufacturing order. The performance stan-dard, which was based on information from a pretest of performanceattainments on this task, was set at a level that was difficult to fulfill. Wenext describe the logic of the simulation model and the decision optionsavailable to subjects. A more detailed description of the mathematicsand logic of the model can be found in Wood and Bailey (1985).

In performing the managerial role, subjects had to allocate the em-ployees to the various subfunctions for each manufacturing order. Theycould reassign employees if they felt that a particular employee wouldbe better suited for a different job assignment. After employees had beenallocated to jobs for a given trial, subjects could then assign each em-ployee a production goal from a set of options that included urging theemployees to do their best or assigning them one of three specific goalsset at, above, or below the established standard. A fifth option allowedsubjects to set no production goal for an employee, if they judged thatit would have a negative motivational effect. Goal assignments for em-ployees, which were made at the beginning of each trial, influenced anemployee's performance according to the calculations of the simulationmodel in the manner predicted by goal theory (Locke et al., 1981).Goals that present a moderate challenge lead to higher performancethan no goals or instructions to do one's best. However, repeated impo-sition of goals that exceed an employee's prior performance at a levelthat renders them unattainable has a negative effect on performanceafter two trials. Continued imposition of unattainable goals would even-tually lead to their rejection and diminished motivation. To enhancethe performance of their organizational unit, subjects had to learn thedecision rule for setting the optimal level of challenge for each employee.

Instructive feedback and social rewards were given after the produc-tion order for each trial had been completed. The feedback and rewarddecisions, which influenced performance on the subsequent trial, mod-eled the temporal effects of such actions in actual organizational envi-ronments. For the feedback decision, subjects could give employees nofeedback or select one of three options that varied in the amount ofdirection given regarding methods of workmanship and analysis ofdifficulties. Instructive feedback had a positive effect on performancefor employees who were performing below the established standard.When an employee performed above standard, the continued use ofhigh directive feedback on three or more trials was regarded as oversu-pervision that would have a negative effect on performance. Effectiveuse of the feedback options to improve organizational performance re-quired subjects to learn decision rules for optimal adjustment of levelof instructive feedback to performance attainments.

For decisions regarding social rewards, the effects of the three optionsvaried with the type of reward given (e.g., compliment, social recogni-tion, note of commendation) and with the degree to which rewards werecontingent on employees' performance attainments. Subjects also hadthe option of not making any laudatory comments regarding their em-ployees' work. Social rewards had a positive effect on performance.

410 ROBERT WOOD AND ALBERT BANDURA

However, in an organizational setting the impact of rewards on perfor-mance is affected by social comparison processes as well. Therefore, themagnitude of the incentive effect for a given employee depended on theratio of rewards to attainments for that employee compared with theequivalent ratio for other employees. Subjects, therefore, had to learn acompound decision rule combining incentive and equity factors on howbest to use social rewards to increase organizational performance.

In sum, to optimize performance of the organization the subjectswere managing they had to match employee attributes to job require-ments and to master a complex set of decision rules on how best to guideand motivate their supervisees. To discover the rules they had to testoptions, cognitively process the outcome feedback information of theirdecisional actions, and continue to apply analytic strategies in ways thatwould reveal the governing rules. To complicate matters further, the mo-tivational factors involved both nonlinear and compound rules, whichare especially difficult to learn. Knowing rules does not ensure optimalimplementation of them. Subjects also had to gain proficiency in tailor-ing the applications of the rules to individual employees and to applythem in concert to achieve desired results.

Induction of Conceptions of Ability

We instated the entity and acquirable skill conceptions of abilitythrough the introductory instructions that subjects read before begin-ning the managerial simulation task. In the acquirable skill condition,subjects were told that decision-making skills are developed throughpractice. In acquiring a new skill, people do not begin with faultlessperformance. However, the more they practice making decisions themore capable they become. They were further told that the simulationtask provided a vehicle for cultivating cognitive decision-making capa-bilities.

In the entity condition, subjects were told that decision making re-flects the basic cognitive capabilities that people possess. The highertheir underlying cognitive-processing capacities, the better is their deci-sion making. Subjects were additionally told that the simulation pro-vided a vehicle for gauging the underlying cognitive capacities. Thesedifferential conceptions of ability were embedded in instructions for thesimulation task that otherwise were identical in every respect. Construalof decisional ability as a basic capacity that is revealable by level ofperformance is more likely to heighten concerns about personal compe-tence than is construal of ability as an acquirable skill.

After subjects read the introductory information, the characteriza-tion of ability as a stable personal quality or as an acquirable skill, andthe descriptive profiles of the employees and subfunctions, they per-formed the simulation at a computer terminal. They entered all of theirdecisions on the keyboard of a personal computer. After subjects dem-onstrated that they understood how to use the computer keyboard, theexperimenter left the room. Subjects received information about theweekly production orders, the roster of available employees, and feed-back on the organization's level of productivity on the computer screen.During the experiment, a timer emitted an audible beep every 5 min.Subjects were informed that the timer was provided to help them pacetheir progress rather than to time their decision making.

After the final trial, the experimenter gave subjects a full explanationof the nature and purpose of the study. They learned that they had per-formed the organizational simulation against a difficult performancestandard.

All data were collected in the context of the simulation, which in-cluded a total of 18 trials. The scales for the different self-regulatorymeasures were presented on the monitor following Trials 6, 12, and 18.Subjects recorded their responses on the keyboard. The first assessmentwas conducted after the sixth trial so that subjects would have someexperience with the simulation before being asked to judge their per-ceived efficacy and to set goals for themselves.

Mediating Determinants

Perceived self-efficacy was recorded on a multi-item efficacy scale thatdescribed nine levels of production attainments, ranging from 30% bet-ter to 40% worse than standard production time. Subjects rated thestrength of their perceived self-efficacy that they could get the groupthey were managing to perform at each of the levels of productivitydescribed. The ratings were made in terms of a 10-point scale rangingfrom no confidence at all (I) to total confidence (10). The strength ofperceived self-efficacy was the sum of the confidence scores for the ninelevels of organizational performance.

In assessing self-set goals subjects recorded the level of organizationalperformance they were personally aiming for in the succeeding trials.They selected their personal goal from nine levels of possible organiza-tional attainments ranging from 40% below to 30% above the estab-lished standard and from a tenth option of no particular goal.

We derived the adequacy of subjects' analytic strategies from theirdecisions regarding job assignments and how they varied the motiva-tional factors to discern the managerial rules across each block of trials.The number of systematic tests that subjects carried out to determinehow job allocations and motivational arrangements affected the perfor-mance of individual employees provided the measure of analytic strat-egy. The strategy score was the sum of the decisions across a block oftrials in which subjects changed only a single factor (i.e., job allocation,goal level, instructive feedback, or social reward) for individual employ-ees. Changing more than one factor concurrently for a given employeeis a deficient analytic strategy for testing hypotheses regarding the im-pact of motivational factors on performance because it confounds thecontribution of factors to outcomes. Systematic analytic strategies re-quire changing one factor at a time. Five systematic tests, one for eachemployee, could be made in each trial. Therefore, a subject's analyticstrategy score across a block of six trials could range from 0 to 30.

Another aspect of decision making is the subject's sheer level of deci-sion activity, represented by the total number of factors changed for allemployees in each trial without consideration of confounding varia-tions. This is a quantity measure of decision-making activity, whereassystematic analytic strategy is a quality measure.

Organizational Performance

We measured organizational performance in terms of the total num-ber of hours taken by the group of employees to complete each weeklyorder. The simulation model automatically calculated the number ofproduction hours for each trial on the basis of subject's job allocationsand selections of motivational factors (Wood & Bailey, 1985). The fewerthe production hours, the better the subject's managerial decision mak-ing. Levels of organizational performance attained by subjects are re-ported as percentages of the standard, with a higher score indicatingbetter performance. Organizational performance scores were averagedacross three blocks of six trials each.

Results

Effects of Ability Conceptions on Self-Regulatory Factors

We analyzed the effects of the differential ability conceptionson self-regulatory factors using 2 x 3 analyses of variance (AN-OVAS), with conception of ability as a between-subjects variableand assessment phase as a repeated measures variable. Figure1 shows the mean strength of perceived managerial self-efficacyafter each of the trial blocks. Subjects in the stable entity condi-tion displayed a much less resilient sense of self-efficacy in theface of substandard organizational performances than thosewho regarded decision making as a skill that can be developed

ABILITY CONCEPTIONS AND DECISION MAKING 411

55

g 50LLLUtL

LUC/3

q 45

LUOCCLUO.LL 40

LUEE 3555

\\\\\

• • ACOUIRABLE SKILL

O--O STABLE ENTITY

TRIAL BLOCKS

Figure 1. Changes in strength of perceived managerial self-efficacyacross trial blocks under acquirable skill and entity conceptions ofability.

incrementally. This is shown in the significant interaction be-tween ability conception and phases of assessment, F(2, 44) =3.64, p < .05. Initially, subjects in both conditions expressed amoderately strong sense of managerial efficacy. However, as theycontinued to try to fulfill the difficult production standard,•those in the stable entity condition displayed a progressive de-cline in perceived self-efficacy, whereas those in the acquirableskill condition maintained their sense of managerial efficacy.

Self-set goals followed a similar pattern of change under thetwo conceptions of ability. These changes are plotted in Figure2. Subjects in the acquirable skill condition set for themselvesmore challenging organizational goals than did those in the sta-ble entity condition, F(2,22) = 6.50, p< .01. Subjects' personalgoal setting became progressively more divergent as they con-tinued to manage the simulated organization, as shown in thesignificant interaction effect, F(2,44) = 5.44, p < .01.

Analytic Strategies

Conceptions of ability also had significant effects on subjects'use of analytic strategies (see Figure 2). Subjects who managedthe simulated organization under an acquirable skill conceptionof ability were more systematic in their use of analytic strategiesin testing the effects of job allocations and motivational factorsthan were their counterparts in the entity condition, F( 1,22) =4.94, p < .025. As subjects continued to perform the simulationtask, those who regarded errors as a natural part of developingdecision-making skills became increasingly more systematic intheir testing of managerial options, whereas those for whom er-rors implied basic cognitive deficiencies became less efficient,F(2,44) = 4.35/><.05.

As they continued to perform the task, subjects in the entity

condition changed more motivational factors in their efforts todiscover optimal motivators but did so in a confounding fashionthat undermined opportunities to learn from the outcome feed-back. Those in the acquirable skill condition made fewer butmore systematic changes in motivational factors. The more pro-duction assignments the entity subjects had to manage, themore erratic they became in their strategic thinking. The inter-action effect for sheer amount of changes was F(2, 44) = 3.83,p<.05.

Organizational Performance

In Figure 3, the mean organizational performance that sub-jects achieved is plotted as a function of conceptions of abilityand blocks of production assignments. Subjects in both condi-tions performed below the preset standard of productivity. Bothgroups attained similar levels of performance in the early trialsof organizational management and did not differ in this respect.However, as predicted from the postulated influence of self-reg-ulatory dynamics, organizational performance was markedlyimpaired under the entity conception of ability, showing asteady decline across progressive blocks of trials. In contrast,those in the acquirable skill condition sustained a high level oforganizational performance. The growing divergence in organi-zational performance produced a highly significant interactionbetween conceptions of ability and trial blocks, F(2,44) = 9.69,p < .001. We turn now to the mediating effects of the variousself-regulatory mechanisms and their contributions to organi-zational performance.

Path Analysis

We conducted path analyses to test the causal ordering ofvariables. The direction of causality in the path model is estab-lished by theoretical considerations supported by prior researchand temporal sequencing of variables. In this causal model,prior performance influences perceived self-efficacy, personalgoal setting, and subsequent performance. We included priorperformance as the first variable in the analyses as a proxy fora host of possible determinants other than the self-regulatoryinfluences examined in this study.

Performance attainments and self-regulatory factors involvebidirectionality of influence (A. Bandura, 1986). That is, self-influences are changed by performance experiences and subse-quent level of performance is, in turn, altered by motivationalself-influences, which represent nonability factors. As a result,using unadjusted performance scores when controlling for pastperformance will also remove effects that are attributable to themotivational effects of self-influences. When self-influences areautocorrelated, unadjusted control for past performance willalso remove effects that are attributable to self-influences in thecurrent performance. To avoid this overcorrection for past per-formance, the effects of perceived self-efficacy and self-goals atPhase 1 and strategy in Block 2 were removed from Block 2performance before it was introduced into the analyses forBlock 3 performance. For the analyses of Block 2 performance,there was no prior measure of the self-regulatory influences thatcould be used to remove their effects from performance inBlock 1. Therefore, the measure of past performance included

412 ROBERT WOOD AND ALBERT BANDURA

110

^J 105

OO

LUo> 100LL

LLJ

^° 95

90

1

_^ — ' '

—

o--»,̂

- "^^^O-\\

^^o

_

,

C.C.

LU

DC

te 18

OH

rf 16Z<LL0 14

LLJwI—z 12LU

OLL

LU 10

1

1 /'

/^

X*

•

O- —— — o —

"̂"~ «-̂"̂ - o

—

• • ACQUIRABLE SKILL

O O STABLE ENTITY

1 1 1

1 . 2 3 1 2TRIAL BLOCKS

Figure 2. The left panel shows the changes in self-set goals across trial blocks under acquirable skill andentity conceptions of ability; the right panel shows the changes in effective analytic strategies across trialblocks under the different conceptions of ability.

in analysis of Block 2 performance had only the prior effects forstrategy removed.

Perceived self-efficacy was entered into the equation as a sec-ond predictor because beliefs about one's capabilities influencethe goals people set for themselves and how proficiently theyuse analytic strategies. Perceived self-efficacy also contributesindependently to performance. We expected personal goals toaffect subsequent performance directly and indirectly throughtheir influence on analytic strategies. The full set of structuralequations representing the hypothesized causal relations wereanalyzed separately for Trial Blocks 2 and 3.

The standardized path coefficients that were significant at orbeyond the .05 level are shown in Figure 4. In the second blockof trials, the relation of prior organizational performance tosubsequent performance was mediated entirely by perceivedself-efficacy, personal goals, and analytic strategies. No signifi-cant relation was found between prior and subsequent organiza-tional performance when the combined influence of the self-regulatory factors was controlled.

Perceived self-efficacy enhanced the level of organizationalperformance both directly and indirectly through its effects onanalytic strategies. Prior organizational performance affectedthe goals subjects set for themselves. Personal goals, in turn,influenced performance by their impact on analytic strategies,but they did not contribute independently to organizationalperformance. Effective use of analytic strategies enhanced or-ganizational performance after controlling for all prior determi-nants.

The structure of the significant causal relations was repli-cated in the third block of trials with the exception that personalgoal setting dropped out as a mediating influence. The com-bined set of explanatory variables in the conceptual model ac-counted for a major share of the variance in organizational at-tainments in both the second trial block (R2 = .75, p < .001)and the third block (R2 = .78, p < .001).

Discussion

Our findings provide supporting evidence that the conceptionof ability with which people approach complex decision makinghas substantial impact on self-regulatory mechanisms that gov-ern performance attainments. Construing ability as an acquir-able skill fostered a highly resilient sense of personal efficacy.Although the challenging preset level of organizational perfor-mance eluded the subjects, they remained steadfast in their per-ceived managerial self-efficacy, continued to set for themselveschallenging organizational goals, and used analytic strategies inways that aid discovery of effective managerial decision rules.Such a self-efficacious orientation, which is well suited for han-dling adversity, paid off in uniformly high organizational attain-ments.

These findings are in accord with a growing body of evidencethat human attainments and positive well-being require astrong sense of personal efficacy (Alloy & Abramson, 1979; A.Bandura, 1986; M. M. Bandura, 1987; Glasgow & Arkowitz,1975; Lewinsohn, Mischel, Chaplin, & Barton, 1980). This is

ABILITY CONCEPTIONS AND DECISION MAKING 413

100

t 90

|

QO

o_ 80

O

70

OccO

60

\\

\\

• • ACQUIRABLE SKILL

O—-O STABLE ENTITY

I I

TRIAL BLOCKS

Figure 3. Level of organizational productivity achieved across trialblocks by subjects who managed the simulated organization under aninduced acquirable skill conception of decision-making ability or underan induced entity conception of ability.

because ordinary social realities are usually fraught withdifficulties. People must, therefore, have a robust sense of per-sonal efficacy to sustain the productive attentional focus andperseverance effort needed to succeed.

Construing ability as reflecting an underlying personal capac-ity greatly increased vulnerability to the adverse effects of fail-ure. Viewed from this cognitive perspective, substandard per-formances become diagnostic of personal deficiencies ratherthan a natural, instructive element in the acquisition of compe-tencies. The longer the subjects with the entity orientation man-aged the complex task, the more they were beset with doubtsabout their managerial efficacy. They kept lowering their organi-zational aspirations and achieving progressively less with theorganization they were managing.

In this study we raised evaluative concerns about personalcompetence by the way in which complex decision making wassocially construed. Other forms of social influence that focusattention on self-evaluation rather than on task mastery—suchas valuational feedback, normative grading, competitive struc-tures, and coercive incentive systems—can similarly have ad-verse effects on the level of interest, motivation, performanceaccomplishments, and creativity (Amabile, 1983; Butler, 1987;Nicholls, 1984; Ryan, 1982). The findings of this study furtherreveal the substantial impact exerted by evaluatively orientedconceptions on self-percepts of efficacy, aspirations, and ana-lytic thinking and the causal structure through which these me-diating regulatory processes affect performance accomplish-ments.

Evidence regarding the decisional activities of subjects in theentity condition indicates that their declining organizational at-tainments reflect impairment in the use of decisional skillsrather than simply slackening of effort or diminished involve-ment in the managerial activity. Subjects who performed themanagerial task under the entity conception surpassed thosewho approached the task as an acquirable skill in sheer amountof decisional activity. However, the attempt to hit on optimalmotivators by varying many factors concurrently impeded dis-covery of the managerial rules. The more their perceived mana-gerial efficacy declined, the more erratic they became in theirdecisional activities (r = .56, p < .05, and r = .76, p < .01, inthe second and third trial blocks, respectively). Their counter-parts in the acquirable skill condition continued to think in astrategically effective manner.

Subjects in the entity and acquirable skill conditions did notdiffer initially in either their perceived self-efficacy or organiza-tional attainments. Their subsequent growing divergence in per-ceived self-efficacy and inferential activity suggests that differ-ential conceptions of ability biased how initially similar sub-standard performances were being cognitively processed.Construal of substandard attainments as indicants of personaldeficiencies would gradually create an inefficacious self-schemain the particular domain of functioning, whereas construal ofsubstandard attainments as instructive guides for enhancingpersonal competencies would foster an efficacious self-schema.Such evolving self-beliefs can further bias cognitive processingof outcome information and promote actions that create con-

PASTPERFORMANCE

.53 (.53)SELF- AJ l |̂ ANALYTIC

EFFICACY "*" STRATEGIES

.39 (.69)PERFORMANCE

I

JO (.67)

.87 <.87>SELF- № ( !̂> ANALYTIC

EFFICACY STRATEGIES

.38 (.81)PERFORMANCE

I

.49 (.85)

Figure 4. Path analysis for the second and third trial blocks. (The initial numbers on the paths of influenceare the significant standardized path coefficients [ps <; .05]; the numbers in parentheses are the first-ordercorrelations. The network of relations on the left half of the figure are for the second trial block, and thoseon the right half are for the third trial block.)

414 ROBERT WOOD AND ALBERT BANDURA

firmatory behavioral evidence for them. Indeed, the data revealan exacerbation cycle of motivational and performance effects.These divergent changes are all the more significant becausemost of the subjects were seasoned managers and all were ofhigh aptitude.

There is some evidence to indicate that in organizational mi-lieus, low perceived self-efficacy to produce desired outcomesfosters attributions of blame and deficiencies to others. Thus,for example, teachers who have a low sense of instructionalefficacy tend to regard difficult students as lacking ability andunteachable, whereas those who have a strong belief in theirinstructional efficacy view student problems as surmountablethrough extra effort and variation of educational approach(Ashton & Webb, 1986). The subjects in this experiment didnot perform the managerial task dispassionately. They gotdeeply involved in the activity and personalized the employees.Those who suffered a loss in perceived managerial efficacy wereuncharitable in their views of their employees. They regardedsome of them as "unmotivatable," unworthy of supervision,and deserving of dismissal.

Although the sample size was relatively small, the temporalordering of changes in self-regulatory factors antecedently toperformance, the replication of the paths of influence in thepath analyses across two series of trials, and the high degree ofstability between the estimates for the two sets of parametersadds to the significance of the findings.

In accord with the proposed causal model, prior performanceaffected perceived self-efficacy, which, in turn, influenced sub-sequent level of organizational performance both directly andindirectly by its impact on analytic strategies. These findings arein accord with previous evidence that a high sense of personalefficacy fosters strategic thinking and raises organizational at-tainments under different levels of organizational complexityand goal assignments (Wood et al., in press). In these simula-tions, subjects had to regulate the actions of others. Consider-able research documents the influential role played by per-ceived self-efficacy in the regulation of one's own motivationand actions (A. Bandura, 1986,1988).

High personal goals also augmented organizational attain-ments by promoting effective problem-solving strategies, butgoals did not affect performance independently. The task of ex-ercising control over organizational outcomes differs in severalimportant respects from that of realizing individualized out-comes. At the organizational level subjects have to rely on theaggregate efforts of others to achieve desired outcomes, whereasat the individual level one need regulate only one's own efforts.Socially mediated regulation of a group is considerably morecomplex than direct self-regulation. Functional relations estab-lished at the individual level may thus require qualifications atthe group level.

Personal goals are readily translatable into performance at-tainments when people possess the knowledge and means to ex-ercise control. Goals can affect performance directly by chan-neling attention and mobilizing effort and sustaining it in theface of obstacles (Locke et al., 1981). In most of the researchdemonstrating performance-enhancement effects of goals, sub-jects already possess the means and need only to intensify theirefforts (Mento et al., 1987). Even on tasks that are directly con-trollable by effort alone, goal effects are weaker for more com-

plex activities (Wood, Mento, & Locke, 1987). In the organiza-tional task used in this study, the managers not only had to relyon the efforts of others to achieve desired outcomes, but theyhad to discover complex managerial rules and how best to applythem as they tried to conduct their managerial function. Untilthe rules were identified, goals could produce more effortful anddiscerning cognitive processing of outcome information, butnot necessarily immediate improvements in organizational per-formance. Indeed, goals exerted their impact on performanceby promoting effective managerial rule-learning strategies.However, after productive analytic strategies were adopted, per-sonal goals no longer served this mediating function.

The multifaceted nature of the managerial activities and theircomplex linkages to organizational outcomes may also partlyexplain why personal goal setting did not have a direct impacton group performance. In the postexperiment inquiry, manysubjects indicated that they were attempting to realize their or-ganizational goal by concentrating their efforts on improvingparticular operational aspects of the organization. These opera-tional subgoals involved reassignment of certain productionsubfunctions or a concerted effort to improve the performanceof a particular employee. If grounded in sound judgment, suchfractional changes would eventually raise organizational attain-ments without necessarily producing appreciable immediategains.

Perceived self-efficacy did not account for variance in per-sonal goal setting. This finding is discrepant from evidence thatperceived self-efficacy influences the goals people set for them-selves and the strength of their commitment to them (Locke etal., 1984; Taylor et al., 1984). Variation in the temporal proper-ties of the goals may explain this discrepancy. In previous stud-ies subjects set themselves proximal goals, whereas in this studythey selected more distal goals for an entire series of perfor-mances. The instated ability conceptions would make the chal-lenging preset production standard a highly salient distal goal.Subjects in the entity condition would strive for it as the eviden-tial standard of competence, whereas more modest distal ac-complishments would prove disappointing. Subjects in the ac-quirable skill condition would also aim for a high eventual ac-complishment. The substantial negative discrepancy betweenprior performance and self-set goals indicates that the presetstandard was indeed exercising considerable influence on thedistal goal setting. Although prior performance had some effecton personal goals, the ones subjects were seeking to fulfill werefar in excess of their attainments. This disparity remained evenfor subjects in the entity condition who began to lower theirsights under the cumulative impact of failure.

The factors encompassed by prior performance level had noindependent influence on subsequent organizational attain-ments. This finding underscores the considerable extent towhich nonability self-regulatory factors may contribute to per-formance attainments. Among highly talented individuals,much of the variation in performance may be attributable tohow well they use the skills they possess. An ability is only asgood as its execution. Our research indicates that the concep-tions of ability with which people approach complex tasks canaffect self-regulatory factors in ways that are self-enhancing orself-impeding.

ABILITY CONCEPTIONS AND DECISION MAKING 415

References

Alloy, L. B., & Abramson, L. Y. (1979). Judgment of contingency indepressed and nondepressed students: Sadder but wiser? Journal ofExperimental Psychology, 108, 441-487.

Amabile, T. M. (1983). The social psychology of creativity. New \brk:Springer-Verlag.

Ashton, P. T, & Webb, R. B. (1986). Making a difference: Teachers'sense of efficacy and student achievement. White Plains, NY: Long-man.

Bandura. A. (1986). Social foundations of thought and action: A socialcognitive theory. Englewood Cliffs, NJ: Prentice-Hall.

Bandura, A. (1988). Self-regulation of motivation and action throughgoal systems. In V. Hamilton, G. H. Bower & N. H. Frijda (Eds.),Cognitive perspectives on emotion and motivation (pp. 37-61). Dor-drecht, the Netherlands: Kluwer Academic Publishers.

Bandura, A., & Cervone, D. (1983). Self-evaluative and self-efficacymechanisms governing the motivational effects of goal systems. Jour-nal of Personality and Social Psychology, 45, 1017-1028.

Bandura, A., & Cervone, D. (1986). Differential engagement of self-reactive influences in cognitive motivation. Organizational Behaviorand Human Decision Processes, 38, 92-113.

Bandura, M. M., & Dweck, C. S. (1987). The relationships of concep-tions of intelligence and achievement goals to patterns of cognition,affect and behavior. Manuscript submitted for publication.

Beach, L. R., Barnes, V. E., & Christensen-Szalanski, J. J. J. (1986).Beyond heuristics and biases: A contingency model of judgmentalforecasting. Journal of Forecasting, 5, 143-157.

Bourne, L. E., Jr. (1965). Hypotheses and hypothesis shifts in classifica-tion learning. Journal of General Psychology, 72, 251-262.

Brehmer, B., Hagafors, R., & Johansson, R. (1980). Cognitive skills injudgment: Subject's ability to use information about weights, func-tion forms, and organizing principles. Organization and Human Per-formance, 26, 373-385.

Bruner, J. S., Goodnow, J., & Austin, G. A. (1956). A study of thinking.New York: Wiley.

Butler, R. (1987). Task-involving and ego-involving properties of evalu-ation: Effects of different feedback conditions on motivational per-ceptions, interest, and performance. Journal of Educational Psychol-ogy, 79, 474-482.

Dweck, C. S., & Elliott, E. S. (1983). Achievement motivation. In P. H.Mussen & E. M. Heatherington (Eds.), Handbook of child psychology:Socialization, personality and social development (4th ed., Vol. 4, pp.644-691). New York: Wiley.

Elliott, E. S., & Dweck, C. S. (1988). Goals: An approach to motivationand achievement. Journal of Personality and Social Psychology, 54,5-12.

Glasgow, R. E., & Arkowitz, H. (1975). The behavioral assessment ofmale and female social competence in dyadic heterosexual interac-tions. Behavior Therapy, 6,488-498.

Hogarth, R. (1981). Beyond discrete biases: Functional and dysfunc-

tional aspects of judgmental heuristics. Psychological Bulletin, 90,197-217.

Latham, G. P., & Lee, T. W. (1986). Goal setting. In E. A. Locke (Ed.),Generalizing from laboratory to field settings (pp. 101-117). Lexing-ton, MA: Heath.

Lazarus, R. S., & Launier, R. (1978). Stress-related transactions be-tween person and environment. In L. A. Pervin & M. Lewis (Eds.),Perspectives in interactional psychology (pp. 287-327). New \fork:Plenum Press.

Lewinsohn, P. M., Mischel, W., Chaplin, W, & Barton, R. (1980). So-cial competence and depression: The role of illusory self-perceptions.Journal of Abnormal Psychology, 89, 203-212.

Locke, E. A., Frederick, E., Lee, C., & Bobko, P. (1984). Effect of self-efficacy, goals, and task strategies on task performance. Journal ofApplied Psychology, 69, 241-251.

Locke, E. A., Shaw, K. N., Saari, L. M., & Latham, G. P. (1981). Goal-setting and task performance: 1969-1980. Psychological Bulletin, 90,125-152.

Meichenbaum, D. H. (1977). Cognitive behavior modification: An inte-grative approach. New York: Plenum Press.

Mento, A. J., Steel, R. P., & Karren, R. J. (1987). A meta-analytic studyof the effects of goal setting on task performance: 1966-1984. Organi-zational Behavior and Human Decision Processes, 39, 52-83.

Mintzberg, H. (1973). The nature of managerial work. EnglewoodCliffs, NJ: Prentice-Hall.

Nicholls, J. G. (1984). Achievement motivation: Conceptions of ability,subjective experience, task choice, and performance. PsychologicalReview, 91, 328-346.

Ryan, R. M. (1982). Control and information in the intrapersonalsphere: An extension of cognitive evaluation theory. Journal of Per-sonality and Social Psychology, 43, 450-461.

Sarason, I. G. (1975). Anxiety and self-preoccupation. In I. G. Sarason& C. D. Spielberger (Eds.), Stress and anxiety (Vol. 2, pp. 27-44).Washington, DC: Hemisphere.

Stewart, R. (1967). Managers and their jobs. London: Macmillan.Taylor, M. S., Locke, E. A., Lee, C., & Gist, M. E. (1984). Type A behav-

ior and faculty research productivity: What are the mechanisms? Or-ganizational Behavior and Human Performance, 34, 402-418.

Wood, R. E., & Bailey, T. (1985). Some unanswered questions aboutgoal effect: A recommended change in research methods. AustralianJournal of Management, 10, 61-73.

Wood, R. E., Bandura, A., & Bailey, T. (in press). Mechanisms govern-ing performance in complex decision-making environments. Organi-zational Behavior and Human Decision Processes.

Wood, R. E., Mento, A. J., & Locke, E. A. (1987). Task complexity as amoderator of goal effects: A meta-analysis. Journal of Applied Psy-chology, 72,416-425.

Received February 1,1988Revision received June 21,1988

Accepted July 11, 1988 •