Group Dynamics: Theory, Research, and Practice2000, Vol. 4. No. 1,27-43

Copyright 2000 by the Educational Publishing Foundation1089-2699/00/$5.00 DOI: 10.1037//1089-2699.4.1.27

Leadership Research and Theory: A Functional Integration

Martin M. ChemersUniversity of California, Santa Cruz

This historical overview of leadership theory and research with an eye for commonali-ties provides an opportunity for integration. Early unproductive research focused onpersonality traits and behaviors. A recognition of the more complex nature of thephenomenon resulted in the development of contingency theories that examined leadercharacteristics and behavior in the context of situational parameters. The 1970s broughtan awareness that perceptions of leaders by followers and others, and perceptions offollowers by leaders, were influenced by cognitive biases arising from prior expecta-tions and information-processing schema. Ironically, attention was belatedly drawn tothe study of female leaders, who were often the victim of cognitive biases and negativeassumptions. Recent research has reflected on the role of cultural differences inleadership processes and has been drawn again into the search for outstanding leaderswith universally effective characteristics. The article concludes with an integration ofcurrent knowledge in leadership effectiveness.

For much of its history, leadership theory andthe empirical supporting research have beenregarded as a fractured and confusing set ofcontradictory findings and assertions withoutcoherence or interpretability. In this article Targue that a considerable commonality ofwell-accepted findings points the way toward asuccessful and useful integration of currentknowledge. That integration, based on the keyfunctions performed by effective leaders, beginsto answer the question of how good leadersbehave and raises intriguing questions about thepersonal characteristics of leaders that facilitatethose behaviors.

This article takes an historical perspective,and the analysis is divided into four periods: (a)the period prior to the presentation of Fiedler's(1964) contingency model; (b) the period from1965 to 1975, focusing on the development andelaboration of contingency theories; (c) theperiod from 1975 to 1985, when cognitivetheories and concerns about gender differencesarose; and (d) the period since 1985, which hasmost extensively focused on transformationaltheories and cultural influences. The historicalanalysis is followed by a presentation of an

Correspondence concerning this article should be ad-dressed to Martin M. Chemers, Division of Social Sciences,117 Social Sciences I, University of California, 1156 HighStreet, Santa Cruz, California 95064. Electronic mail may besent to mchemers@zzyx.ucsc.edu.

integrative framework and a suggested directionfor future research.

In this analysis leadership is defined as "aprocess of social influence in which one personis able to enlist the aid and support of others inthe accomplishment of a common task." Thisdefinition places the subject distinctly within thepurview of social psychology, and the analysesthat follow reveal how fully the leadershipliterature is indebted to the field of socialpsychology for its dominant paradigms andcentral variables.

Before Contingency Theory:Lost in the Wilderness

Social philosophers have had a long-standinginterest in both organizational and politicalleadership. Western European philosophers,embedded in a strongly individualistic culturalmilieu, looked primarily to the characteristics ofleaders for explanatory premises, For example,Carlyle (1841/1907) proposed the great mantheory of leadership, which argued that success-ful leaders possessed traits of personality andcharacter that set them apart from ordinaryfollowers. The interest in individual characteris-tics of leaders was spurred by the emergence ofintelligence tests in the early 20th century.Empirical psychology turned toward the studyof traits, and the nascent leadership fieldfollowed suit.

27

28 CHEMERS

Traits

A sort of naive popularist psychology guidedthe choice of traits considered for leadershipimpact. Traits that were stereotypically associ-ated with leadership, such as dominance,assertiveness, intelligence, physical stature, so-cial sensitivity, and many others, became causalcandidates. The typical research format for theseearly studies was to identify a group withleaders and followers and test for differences onthe selected trait measures. Stogdill (1948)provided an extensive review of 30 years of thetrait studies. He reported that a few traits (mostnotably intelligence) were sometimes associatedwith reliable differences between leaders andfollowers (i.e., about 35% of the time), but therewas no single variable or even cluster ofvariables that was related to leadership across avariety of situations. Stogdill concluded thatalthough individual differences were certainlyimportant in identifying emergent or effectiveleaders, the great diversity of situations in whichleaders functioned made it unlikely that any onetrait would be a universal predictor. Although itwas not immediately recognized, StogduTsanalysis set the stage for theories of leadershipthat were predicated on an interaction betweenleader traits and situational contingencies.

Behaviors and Styles

Daunted by the failure of traits to predictleadership, but unwilling to abandon individual-istic explanations, researchers turned to thestudy of leader behavior. Observations of theeffects of leadership style (i.e., autocratic vs.democratic) on the atmosphere of small groups(Lewin, Lippitt, & White, 1939), processanalyses of interactions in laboratory discussiontasks (Bales & Slater, 1955), and reports ofindustrial workers on the behavioral styles oftheir supervisors (Kahn, 1951) sought to iden-tify patterns of leader behavior associated withhigh productivity or good morale.

The most extensive of these research pro-grams and the one with the most enduringimpact on the field of leadership was the set ofstudies surrounding the development of theLeader Behavior Description Questionnaire(LBDQ) at Ohio State University (Hemphill,1950). This 150-item behavioral inventory wasused to collect ratings of military and industrial

leaders by supervisors, subordinates, and observ-ers. Subsequent factor analyses revealed that amajor portion of the variability in leaderbehavior could be explained by two majorclusters (Halpin & Winer, 1957). The mostprominent factor, labeled Consideration, in-cluded behaviors such as showing concern forthe feelings of subordinates, making sure thatminority viewpoints were considered in deci-sion making, and attempting to reduce conflictin the work environment. These behaviorsseemed to reflect leader intentions to supportpositive group morale and follower satisfaction.A strong second factor, labeled Initiation ofStructure, included items measuring the leader'suse of standard operating procedures, criticismof poor work, and emphasis on high levels ofperformance. These behaviors appeared to berelated to a leader's focus on building a structurefor task accomplishment.

Although the LBDQ factors were foundreliably in ratings of leader behavior across awide range of settings, they were less thancompletely successful at predicting the impor-tant outcomes associated with leadership effec-tiveness, that is, follower satisfaction and groupperformance (Fleishmann & Harris, 1962; Kor-man, 1966). Considerate leadership was oftenrelated to follower satisfaction or morale, andConsideration and Initiation of Structure weresometimes but not always predictive of groupperformance. The failure of this carefullyconstructed and comprehensively researchedbehavioral measure to predict leadership ledmany researchers to throw up their hands infrustration and seemed to be yet anotherinstance of leadership research leading to nocoherent conclusions.

Legitimacy

One of the brightest spots in the earlyempirical work on leadership was the series ofstudies conducted by Hollander, which illumi-nated some of the facets of leadership statusaccrual and legitimacy. In both laboratory andfield research venues, Hollander (1964; Hol-lander & Julian, 1970) found that individuals ingroups gain status through the demonstration oftask-related competence and loyalty to groupvalues. Status acquisition is associated with theaccrual of so-called "idiosyncrasy credits,"which can be thought of as units of group

SPECIAL ISSUE: LEADERSHIP RESEARCH AND THEORY 29

acceptance that can be "spent" to influenceothers and provide leeway from group norms toallow for innovation in group processes andviews.

This pioneering work on how individualsdecide to follow those in leadership positionshas retained its currency to the present day, inpart because it embodies both cognitive andbehavioral elements in its approach. Contempo-rary information-processing theories of leader-ship posit leadership "prototypes" that arecharacterized by the elements of competencyand trustworthiness that are the bases for theaccrual of idiosyncrasy credits. Recent work byHogg and his associates (Hogg, Hains, &Mason, 1998) applying social identity theory toleadership perception indicates that, despite atendency for followers to value leaders whoembody group values (the basis for perceptionsof trustworthiness in Hollander's model), theyalso heavily weight task relevant competence inleadership evaluation. Those basic determinantsof leadership status turn up in more recentapproaches to understanding leadership judg-ments. A fuller discussion of the role ofperception in leadership process appears later.

The Mid-1960s to the Mid-1970s:The Contingency Era

The Contingency Model

The study of leadership took a dramaticchange of direction with the publication ofFiedler's first articles (1964) and subsequentbook (1967), which presented a new approach tounderstanding leadership effectiveness. Thecontingency model of leadership effectivenessemerged as an answer to StogduTs (1948) callfor an approach based on the interaction ofleader traits with situational parameters, but itdid not start out that way. Early work (Cleven &Fiedler, 1956; Fiedler, 1955, 1958) tested thepredictive validity of a leadership trait measureon the basis of the leader's views of coworkers.The measure, which eventually came to beknown as the Least Preferred Coworker (LPC)scale, differentiated leaders who viewed poorlyperforming coworkers in very negative terms(thought to reveal a very strong concern witheffective task performance) from those whoviewed poorly performing coworkers in less

negative terms (hypothesized to reflect a greateremphasis on interpersonal relationships).

Apparently promising early work found thattask-oriented leaders were more effective, butsubsequent studies yielded results showing thatrelationship-oriented leaders had more effectiveteams. Confused but not deterred by theseanomalous findings, Fiedler reanalyzed a largenumber of studies—this time classifying thegroup settings in terms of the degree of supportand cooperation offered by followers, the clarityand structure of the group's task, and theleader's formal authority to direct and rewardfollowers. These three variables were combinedinto a dimension of "situational favorableness"(Fiedler, 1967) or "situational control" (Fiedler,Chemers, & Maher, 1976), thought to reflect thedegree to which the overall situation gave theleader a feeling of certainty, predictability, andcontrol over group processes.

When the leader's orientation (i.e., LPCscore) was correlated with group performanceacross the dimension of situational favorable-ness, a reliable relationship was found. Specifi-cally, groups led by task-oriented leadersperformed best in situations of high control andpredictability or very low control and predictabil-ity, and groups led by relationship-orientedleaders performed best in the situations ofmoderate control or predictability. The explana-tory rationale for these findings was that therelatively more directive, task-focused leader-ship style is most appropriate when an orderlysituation provides the leader with the clarity togive directions and the follower support to besure of his or her performance, and the highlyvolatile and unpredictable environment of thevery low control situation also requires thesteadying influence of clear directions andstructuring leader behavior. However, the moreinterpersonally oriented, participative style ofleadership was thought to function most effec-tively when the complexities of a moderate-control situation required greater delicacy tonavigate a poorly understood task or to avoid thedangers associated with uncertain followersupport.

The inductive method by which the contin-gency model was constructed and the highlycomplex nature of its predictions led to manycriticisms of the model during the 1970s(Ashour, 1973; Graen, Alvarez, Orris, & Mar-tella, 1970). However, subsequent research and

30 CHEMERS

extensive meta-analyses (Peters, Hartke, &Pohlmann, 1983; Strube & Garcia, 1981)provided strong support for the basic principlesof the model. (For a more complete discussionof the development of the contingency modeland the controversy surrounding its validity, seeChemers, 1997.) Another criticism of thecontingency model was its apparent assumptionthat a leader could not choose to be both taskand relationship oriented when the situationdemanded it. Some other contingency ap-proaches did not make that same assumption.

Normative Decision Theory

Energized by the potential of a contingencyapproach to explain leadership performance, butworking from a more deductive theoretical base,Vroom and Yetton (1973) offered a model ofdecision-making effectiveness that integratedleaders' decision strategy with situational fac-tors. Leaders were conceived as having a rangeof decision-making strategies available to themthat varied in degree of follower involvement inthe process—ranging from autocratic styles(leader makes the decision with minimal fol-lower input) to consultative styles (leader makesthe decision after getting follower opinion andadvice) to group or participative styles (leaderand group make decision together, with equalweight).

The situational parameters included in themodel were represented as a series of questionsarranged to yield a decision tree. Leadersseeking the most effective decision strategywere asked to analyze situational factors thatincluded the clarity and structure of the task andsurrounding information, the degree of supportfor the leader and the organization among thefollowers, the degree of conflict among subordi-nates, and the time urgency for a decision to bemade. The model specifies that when the task isclear and the followers supportive, the leadershould use the more time-efficient autocraticstyles. If the task or information is unclear, usingthe consultative strategies increases the informa-tion yield and likelihood of a higher qualitydecision. When the leader lacks followersupport, the participative strategy helps toensure follower commitment to the decision andits implementation. Empirical research on thenormative decision model is not extensive but is

generally supportive of its basic premises (Field& House, 1990).

The contingency model and normative deci-sion theory have many features in common.They are both focused on the leader as thecentral actor in the group's efforts to interfacewith the task environment. Both theories regardthe leader's task as to gain the group's support insolving problems and implementing solutionseffectively. Also, they both hypothesize thatmore directive approaches will be most effectivewhen a clear task and a supportive group givethe leader the certainty to take charge but thatmore participative strategies will work betterwhen a less clear and orderly environmentargues against bold action and autocratic direc-tion. The two theories part company in thesituation of very low control with the contin-gency model more focused on immediate groupperformance through leader direct action butnormative decision theory suggesting moreparticipative strategies to build a more support-ive environment over the long run.

Path-Goal Theory

Contingency theories held the promise ofcorrecting the weaknesses of earlier approachesto leadership effectiveness prediction. For ex-ample, one promising but disappointing ap-proach had been the attempt to relate leaderbehavior (e.g., the LBDQ score) to organiza-tional outcomes. R. J. House and his associates(R. J. House, 1971; R. J. House & Dessler, 1974;R. J. House & Mitchell, 1974) picked up thatgauntlet and attempted to merge traditionalbehavioral approaches with emerging develop-ments in the study of worker motivation tounderstand the impact of the leader on themotivation and performance of followers.

Path-goal theory argues that the leader'smain purpose is to motivate subordinates byhelping mem to see how their task-relatedperformance could help them to achieve theirpersonal goals. Research within the path-goalframework attempted to understand how aleader's directiveness (i.e., Initiation of Struc-ture) or supportiveness (i.e., Consideration)behaviors might affect subordinate motivationand performance. Proceeding logically, thetheory predicted that a leader's structuringbehavior would be motivating to a subordinatewhen the subordinate's task environment lacked

SPECIAL ISSUE: LEADERSHIP RESEARCH AND THEORY 31

structure because of insufficient training orexperience or a highly complex task. However,when a subordinate had sufficient structure,leader directiveness would be regarded as overlyclose monitoring or "pushing" and would havenegative effects. Consideration behavior wasseen to have its most positive effects when thesubordinate needed psychological or emotionalsupport to deal with an aversive work environ-ment (made so by a boring or unpleasant task).Consideration was viewed as superfluous insituations that were engaging and intrinsicallyinteresting to the subordinate.

The typical research paradigm for path-goaltheory studies was to divide a group ofsubordinates into situations of low clarity(presumably interesting, but potentially frustrat-ing because of lack of structure) and of veryhigh clarity-predictability (presumably boringand uninvolving). Leader structuring behaviorwas predicted to have positive effects onsubordinate motivation and performance in theformer situation but not the latter, whereas thereverse was true for leader considerate behavior.Path-goal theory generated a considerable bodyof empirical research support for the basicpropositions. Considerate behavior, for ex-ample, was usually related to positive subordi-nate attitudes under boring or aversive tasksituations but often had similarly positive effectsacross all situations. Results regarding structur-ing behavior were even less consistent.

A study by Griffin (1981) that includedmeasures of subordinates * ' * growth needstrength" (Hackman & Oldham, 1976) indicatedthat the needs and expectations of subordinatesplayed a role in determining when subordinatesneeded or wanted different types of leaderbehavior. Growth-oriented, challenge-seekingsubordinates were quite comfortable with un-structured and challenging tasks and weretherefore less receptive to directive leaderbehavior under unstructured or structured condi-tions. However, these growth-oriented subordi-nates were very responsive to supportivebehavior when the task was boring. The morechange-averse, low-growth-need subordinateswere comfortable with leader structure across allsituations but needed less support when a taskwas ostensibly boring. A reasonable conclusionto be drawn from this literature is that leaderbehavior that is seen as supportive by subordi-nates is likely to lead to positive reactions and

higher motivation and that both characteristicsof the task and of the subordinate will contributeto that receptiveness.

In an interesting extension of path-goaltheory, Kerr and Jermier (1978) argued that ifthe leader's purpose is to supply missingelements in the subordinate's job environment(e.g., structure or support), then other sources ofthose missing elements might make the leader'sbehavior redundant and unnecessary. Their"substitutes for leadership" theory predicted,for example, that if a job provided plenty oftask-relevant feedback, leader structuring behav-ior would be unnecessary, or if a compatible andcohesive work group provided emotional sup-port, leader consideration would be redundant.Under such conditions, leader behaviors werehypothesized to show minimal or even negativerelationships with subordinate motivation, satis-faction, or performance. However, a review of anumber of studies of substitutes-for-leadershiphypotheses indicated very little support for thetheory's predictions (Podsakoff, Niehoff, Mac-Kenzie, & Williams, 1993) and revealed that aleader's behavior remains very important tosubordinates regardless of varying situationalconditions.

The research literature on the contingencytheories suggests that actions by a group's leadercan have strong effects on the motivational andemotional states of followers and on thesuccessful accomplishment of the group's task.The relationship of the specific leader actions tothose outcomes depends on the interaction ofthose actions with relevant features of theinterpersonal and task environment.

The Mid-1970s to the Mid-1980s:Cognitive Models and Gender Concerns

The growing influence of cognitive theoriesin social psychology led to a similar interestamong leadership researchers. Two broad classesof investigation were concerned with percep-tions of leaders by others (i.e., followers,superiors, and observers) and leaders' percep-tions and evaluations of subordinates.

Leadership Perceptions

In the mid-1970s, studies involving ratings ofleader behavior began to reveal certain anoma-lous findings. Eden and Leviatan (1975) re-

32 CHEMERS

ported that when research participants wereasked to make ratings of leader behavior bysimply imagining a leader, the resultant datashowed factor structures similar to those derivedfrom ratings of actual leaders. Staw (1975)showed two sets of observers the same video-tape of a group interaction but told the observersthat the group had either been very successful orvery unsuccessful on task performance. Ob-server ratings of the "successful" leader werehigher on measures of both directive andsupportive leadership than were the ratings ofthe "unsuccessful" leader.

That ratings of leaders might be stronglybiased created a problem on both theoretical andmethodological grounds. Leader legitimacy, acentral construct in understanding the bases ofleader influence, was based on follower percep-tions. Furthermore, almost every research para-digm in the leadership field depended on ratingsof leader behavior (Rush, Thomas, & Lord,1977).

Attribution theory (Jones & Davis, 1965;Kelley, 1967) provided a theoretical frameworkfor the investigation of leadership biases. Takinga very strong position, Calder (1977) argued thatthe very concept of leadership is rooted inpopular language and poorly articulated as ascientific construct. He argued that with no wayof measuring leadership apart from socialperceptions, leadership exists primarily as anattribution rather than a testable construct andshould, therefore, be abandoned as a subject ofscientific inquiry. Few researchers were willing"to throw the baby out with the bathwater" andbegan instead to make a systematic study ofleadership perceptions and the processes thatgave rise to them.

A useful model was provided by the researchon implicit personality theories, which Hastorf,Schneider, and Polefka (1970) defined as astructure of association about what traits orcharacteristics are related that guides andorganizes perceptions, thoughts, and memoriesabout a phenomenon. Implicit theories ofleadership, then, would define the assumptionsthat people held about what behaviors leadersdisplayed and how those behaviors were associ-ated with group and organizational outcomes.

An extensive research program by Lord andhis associates (Lord, 1985; Lord, Binning, Rush,& Thomas, 1978; Lord & Maher, 1991) revealedthat leadership attributions were based on two

processes. Recognition processes determinedwhen an individual's behavior would result inthe perception of that person as a leader.Observers were found to hold highly articulatedprototypes (Cantor & Mischel, 1979; Rosch,1978) of leadership. When an actor's behaviorshowed sufficient overlap with the prototypi-cally driven expectations of observers, a leader-ship attribution was made. Once an individualwas seen as a leader, selective attention andmemory reinforced that judgment.

Leadership judgments were also found to beinfluenced by inferential processes. Becauseimplicit theories of leadership associate teamsuccess with effective leadership, observers arelikely to infer the presence of good leadershipfrom evidence of group success (Phillips &Lord, 1981). Thus, once a person is seen as aleader, observer inferences are likely to rein-force and enhance that perception. Of course, ifa person's characteristics, such as gender orrace, are inconsistent with observers' prototypicexpectations, then such a person is less likely tobe perceived as a legitimate or effective leaderdespite any objective achievements.

The strength of common beliefs in theimportance of leadership for group outcomesled Meindl (1990) to develop the "romanceof leadership" concept. In an ingenious seriesof experiments and naturalistic observations,Meindl showed that any remarkable group ororganizational outcome, whether highly positiveor highly negative, is likely to be attributed toleadership effects, while other reasonable causesare largely ignored.

Although the strong susceptibility to percep-tual biases in the observations of leadershipmight have constituted a problem for researchmethodology, it opened a fascinating area fortheoretical development. If leadership is aprocess of social influence, then factors thataffect the legitimacy, credibility, and influenceof leaders become a central aspect of leadershipfunction.

Another important component of the leader-ship process involves the perception of follow-ers by leaders. Almost every theory of leader-ship posits that a central function of leadershipinvolves the direction of subordinates. Thefollower-oriented contingency theories, such aspath-goal theory, maintain that it is the leader'sresponsibility to provide the subordinate withtask-directed guidance or emotional support to

SPECIAL ISSUE: LEADERSHIP RESEARCH AND THEORY 33

help them to be effective and satisfied. Implicitin these premises is the expectation that leadersare able to judge what kinds of behaviors ontheir part are likely to have positive effects onsubordinates. In other words, leaders mustobserve the actions and reactions of subordi-nates to judge what is needed. This clearlyplaces attributionai processes at the center of therelationship between leader and follower.

Mitchell and his associates (Green & Mitchell,1979; Mitchell Larson, & Green, 1977; Mitchell& Wood, 1980) applied Kelley's (1967) attribu-tion model to leader evaluations of subordinatesand the effects that those evaluations have onsubsequent leader actions. That research re-vealed that processes affecting attributions offollowers by leaders are consistent with earlierattribution research. For example, head nurseswho were asked to make judgments about thecauses of poor performance by a floor nurseintegrated available information about the con-sistency of the poor performance over time andsetting and how the performance compared withthat of other nurses. Also consistent with earlierwork was the finding that these judgments tendto be susceptible to the fundamental attributionerror (Jones & Nisbett, 1971; Ross, 1978), inwhich performance is more likely to be ascribedto internal, personal causes (such as motivationor ability) over equally plausible external causes(such as poor training or support). Furthermore,the more extreme the consequences of the poorperformance (e.g., a patient injury), the greaterthe tendency to ascribe cause to the individual.Mitchell and Wood (1980) demonstrated that thesort of attributions made by the supervisor (i.e.,internal vs. external [to the subordinate]) had asignificant impact on the kinds of leadershipactions (e.g., training, punishment, termination)that were likely to be used by the leader toaddress the situation.

Brown (1984) made some insightful observa-tions about leader attributions in real-worldwork groups. Most attribution studies do notinvolve any real or long-term involvementbetween the observer and the actor. However, inreal work groups, leaders and followers arebound together in a relationship of mutualdependency; that is, when followers performpoorly, leaders are usually held to account.Furthermore, leader and follower are in arelationship of reciprocal causality in that it isthe leader's responsibility to direct and support

the work of the subordinate. Subordinate failuremight be evidence of leadership failure. Thesefactors strengthen the tendency of leaders tomake ego-defensive attributions, blaming subor-dinates for poor performance and possiblytaking personal credit for group success. Inaccu-rate judgments that arise from these blind spotscan easily erode the working relationship andbases of influence between the leader andfollower.

Gender Effects

Few careful, scientific studies of differencesbetween men and women in leadership effectswere done prior to the 1970s. Despite the lack ofscientific evidence on this issue, popular viewswere widespread and strong. Bowman, Worthy,and Greyser (1965) reported that surveys ofmanagers and business school students revealedthe strong belief that women were unsuited formanagerial roles and would make poor leaders.Popular writers, such as Hennig and Jardim(1977), offered quasi-theoretical justification forsuch beliefs by proposing that women lacked theskills and traits necessary for managerialsuccess. It is interesting that the 1980s brought arash of popular books, also with little empiricalbasis, that proposed that feminine traits, such aswarmth, nurturance, and flexibility, made womenbetter leaders and managers than power-oriented, controlling male leaders.

The questions that present themselves in thisarea concern whether men and women actuallyare different in their leadership orientations andbehaviors and whether such differences have aneffect on follower reactions and group ororganizational performance. Three theoreticalexplanations exist for potential differencesbetween male and female leaders: (a) womenand men are biologically different (e.g., hor-mones, temperament, etc.), (b) men and womenare culturally different (i.e., differentially social-ized for gender roles), and (c) observeddifferences between men and women andreactions to those differences are structurallydetermined (i.e., by differences between men'sand women's relative standing in organizationalstructures).

One thing that is very clear is that theleadership stereotypes held by the general publicabout males and females are quite different. In1971 Bass, Krusell, and Alexander reported an

34 CHEMERS

analysis of male managers' responses to asurvey of attitudes toward women at workindicating that men felt that women lackedcareer orientation, leadership potential, andwere undependable and emotionally unstable—all of which made women unsuitable candidatesfor management. Schein (1973,1975) found thatstereotypes of women, held by both men andwomen, were very different from stereotypes ofmen, with the latter being much closer than theformer to stereotypic perceptions about thecharacteristics of a manager. As late as 1989Heilman, Block, Martell, and Simon replicatedSchein's (1973) study and found little change inthese stereotypes. Clearly, then, the commonview was and may still be that women and menare very different in their leadership style andperformance. How good is the evidence?

In a classic treatise on the subject of genderdifferences, Deaux (1984) effectively dismissedany biological bases for gender differences insocial behavior. The evidence simply does notsupport such differences. But what about thepossibility that differences in socialization togender roles carry over to behavior in theworkplace, so-called "gender role spillover"(Nieva&Gutek, 1981)?

In the second edition of the Handbook ofLeadership, Bass (1981) reported that theempirical evidence available at that time showedno consistent pattern of differences betweenmen and women in supervisory style. However,definitive analysis on this topic waited untilEagly and her associates conducted a series ofmeta-analyses on male-female differences inleadership style (Eagly & Johnson, 1990),leadership emergence (Eagly & Karau, 1991),and leadership evaluation (Eagly, Makhijani, &Klonsky, 1992). A careful reading of theseanalyses suggests that to the extent that theobservations of leadership were taken in organi-zational settings; using standard behavioralmeasures; by observers, superiors, or subordi-nates, the differences found between men andwomen are so small as to be of little practicalsignificance. Women tend to emerge as leadersabout as often as men, and they tend to beevaluated similarly to men when all othervariables are equal.

How about when other variables are notequal? Women tend to emerge less frequentlyand are evaluated less positively in situationswhere followers are hostile to women in

leadership or when organizational settings arenot congenial to female leadership. In otherwords, women show few differences from menin actual leadership behavior but are stillsusceptible to the impediments created bynegative stereotypes about female leadership.This conclusion is quite compatible with Deaux's(1984) view that gender is more important as asocial category than as a biological or culturalcharacteristic. Negative views of women lead tonegative expectations that bias women's oppor-tunities for achieving leadership roles and beingfairly evaluated in those roles. Research on thestructural approach supports this view.

J. House (1981) argued that actors in a socialstructure are often strongly influenced by theirplace in that structure. In a series of studies onwomen and power in organizations, Ragins(1989,1991; Ragins & Sundstrom, 1989) foundthat women face a number of barriers along thepath to acquisition of status and power inorganizations. However, when male and femalemanagers are matched for level (i.e., power andauthority) within the organization almost nodifferences are found in leader behavior, perfor-mance, or acceptance by subordinates.

The conclusion that may be drawn from thisliterature is that although few real differences inleadership behavior or style exist between menand women, false but persistent stereotypesimpede equal access and fair evaluation forwomen in organizational leadership.

The Mid-1980s to the Mid-1990s:Transformational Leadership and Cultural

Awareness Transformational Theories

A major shift of interest in leadership researchwas sparked by the work of a political historian.Burns's (1978) book on great leaders differenti-ated transactional leaders, whose relationship tofollowers was based on mutually beneficialtransactions, from transformational leaders,who influence followers to transcend personalinterests and transform themselves into agentsof collective achievement. This was an excitingperspective for a field locked in molecularanalyses of trait-situation interactions andperceptual biases.

Anticipating this development by a year, R. J.House (1977) published a theoretical analysis ofcharismatic leadership in which he analyzed thecharacteristics of historical leaders who elicited

SPECIAL ISSUE; LEADERSHIP RESEARCH AND THEORY 35

extraordinary levels of devotion and commit-ment from followers—for example, Gandhi,Martin Luther King, Jr., and others. Houseidentified three sets of characteristics thattypified charismatic leaders. Personal character-istics included a strong belief in the moralrighteousness of one's beliefs, high levels of selfconfidence, and a strong need to influence anddominate others. Behaviors included dramaticgoal articulation, role modeling of desiredattitudes and behaviors, image building, exhibit-ing high expectations of and confidence infollowers, and arousing follower motives thatwere consistent with desired behavior (i.e.,aggressive or altruistic motives). Finally, situ-ational influences might include high levels ofenvironmental stress (e.g., economic crises,social upheaval) or an opportunity to expressgroup goals in moralistic or spiritual terms.

The most careful, empirical analyses oftransformational leadership have been con-ducted by Bass and his associates (Bass, 1985;Bass & Avolio, 1990a, 1990b, 1993). Bassstarted by interviewing managers about transfor-mational leaders they had known. On the basisof the interviews Bass built and validated aquestionnaire designed to measure transforma-tional leadership: the Multi-Factor LeadershipQuestionnaire (MLQ). Factor analysis of theMLQ yielded seven factors, including three"transactional" factors (Contingent Reward,Management By Exception, and Laissez-FaireLeadership) that were associated with moderateto poor leadership effects and four transforma-tional factors that were associated with highlevels of subordinate motivation and group ororganizational success. The transformationalfactors included (a) Idealized Influence (cha-risma), reflecting extremely high levels of leadercompetency, trustworthiness, or both; (b) Inspi-rational Motivation, involving the articulation ofthe group's goals in emotional, moral, orvisionary terms; (c) Intellectual Stimulation,entailing the encouragement of followers tothink independently and creatively and to moveaway from past ideas or limitations; and (d)Individualized Consideration, relating to theleader's capacity to understand each follower'spersonal needs and goals. Bass (1998) reporteddata from many organizations in countriesaround the world that indicate that leaders whoare rated highly on transformational leadershipcharacteristics by superiors, peers, or subordi-

nates are associated with high-performing teamsand organizations.

House and Shamir (1993) returned to thestudy of transformational leadership, emphasiz-ing the psychological processes of followers thatmediated the effects of charismatic or transfor-mational leader actions. Weaving together path-goal theory (with its emphasis on expectancymotivation) with theories of intrinsic motivationand self-concept, they argued that transforma-tional leaders have several significant psycho-logical effects on followers. By placing thegroup's mission into moral or spiritual contexts,such leaders raise the salience of collectivegoals over personal or selfish interests of thefollowers. Second, tying the follower's self-concept to the group mission makes self-esteemcontingent on group success and fosters self-motivation and self-regulation by followers.

Both the theoretical explication and perfor-mance outcomes associated with transforma-tional leadership make the construct quitecompelling but also leave the leadership scholarwith another conundrum. Transformational theo-ries arc stated in terms of "universally"effective leadership behavior—that is, for allleaders in all situations. It is difficult to squarethat idea with the equally compelling evidencesupporting various contingency theories thatshow that effective leadership is the result of theappropriateness or fit between particular behav-iors and particular situations. Some recentdevelopments applying the concept of self-efficacy (Bandura, 1982) to leadership offers apotential resolution of this contradiction.

Leadership Efficacy

A number of contingency model studiesshowing that in-match leaders felt more confi-dent and in control (Chemers, Ayman, Sorod, &Akimoto, 1991; Chemers, Hays, Rhodewalt, &Wysocki, 1985) led me and my associates toconduct a series of studies designed to assess therole of leadership confidence or efficacy inperformance. Chemers, Watson, and May (inpress) reported concurrent, predictive, anddiscriminant validity for a measure of leadershipefficacy in a longitudinal study of ReserveOfficer Training Corps cadets. Cadets filled outa measure of self-perceived leadership abilityand were rated for military leadership potentialby their military science instructors. Several

36 CHEMERS

months later, the same cadets were followed upthrough a 6-week U.S. Army-sponsored summerleadership camp in which cadets rotated througheveryday leadership roles, participated in train-ing, and were tested in realistic militarysimulations. Ratings by course instructors,training camp superiors, peer cadets, andsimulation exercise grading staff all revealeddramatic effects indicating superior perfor-mance of cadets who expressed greater confi-dence in their leadership capability. By way ofdiscriminant validity, these performance ratingswere not predicted by general self-esteem, andleadership efficacy did not predict nonleader-ship performance (e.g., marksmanship, landnavigation).

Watson, Chemers, and Preiser (1996) re-ported the results of leadership efficacy andteam collective efficacy on the success of smallcollege basketball teams (both men's andwomen's). Before the start of the basketballseason, players were administered question-naires assessing leadership efficacy and collec-tive team efficacy. Teams were followed throughthe season, and the team's win-loss recordprovided the measure of performance. Pathanalyses revealed clear and significant supportfor the role of efficacy in team performance. Thestrongest predictor of team success was teamcollective efficacy, and team efficacy was, inturn, most strongly predicted by leadershipefficacy self-ratings of the identified team leader(usually the captain). Other factors potentiallyrelated to team success (e.g., previous seasonrecord, number of returning players, starters,etc.) were controlled for in the analyses and didnot prove as predictive as the efficacy measures.(No differences were found between men's andwomen's teams.)

These findings on leadership efficacy providea possible resolution of the contradictionbetween contingency theories, which makesituation-specific predictions of leadership suc-cess and transformational theories, which makeuniversal predictions. The fit between theleader's personal characteristics and situationalparameters is an important determinant of aleader's confident and efficacious behavior—behavior that is the basis for the criticalfunctional elements of leadership. That behav-ior, in turn, gives rise to the effective groupprocesses and positive perceptions by observersthat constitute transformational leadership.

Transformational leadership measures leader-ship at the outcome (i.e., dependent variable)level, whereas the contingency theories tend toplace more focus on the leader characteristics(i.e., independent variable) level. Leadershipefficacy may be the psychological link betweencontingent fit and transformational behavior.Later in the present article a functional integra-tion of contemporary leadership theory willelaborate these critical functions.

Cultural Differences

Two streams of thought on cultural differencehave had an influence on leadership theorizing.One stream involved the work of socialpsychologists who were interested in the effectsof culture on social processes but not necessarilyinterested in leadership (e.g., Fiske, 1991;Markus & Kitayama, 1991; Triandis, 1990,1993). These theorists focused attention on thedramatic value differences (i.e., individualismvs. collectivism) among various national groups.Briefly stated, individualistic cultures (such asthose of western Europe and English-speakingcountries) place a high value on personalexpression and achievement, whereas morecollectivist cultures (including most of the restof the world) are more concerned with groupharmony and collective success.

In a value-based approach more closelyfocused on leadership and motivation, Hofstede(1980, 1983) presented an analysis of fourdimensions of national values with profoundeffects on organizational functioning. Powerdistance refers to people's comfort with andacceptance of large differences in power,influence, and wealth among groups or classeswithin the society. Uncertainty avoidance re-flects the extent to which individuals in a societyresist risk and unexpected events by emphasiz-ing rules, norms, and expertise. Hofstede's thirddimension was individualism-collectivism, andthe fourth was masculinity-femininity, whichdifferentiates cultures in which members valuestereotypically masculine pursuits such asstrength, competitiveness, and material achieve-ment from those in which members are moreconcerned with quality of life and concern forothers.

The basic thrust of the value theories is thatorganizational processes in different cultureswill reflect what is considered appropriate and

SPECIAL ISSUE: LEADERSHIP RESEARCH AND THEORY 37

important. For example, leaders should be seenas aloof and powerful in high power distancecultures; as expert, confident, and orderly inhigh uncertainty avoidance cultures; as paternal-istic and caring in collectivist cultures, and asmacho and competitive in masculine cultures.Likewise, the needs of followers and the natureof the leader-follower relationship—for ex-ample, as manifested in a desire for structure(uncertainty avoidance) or personal achieve-ment (individualism)—would affect the waysthat leaders and followers interact in thedetermination of follower satisfaction, morale,and motivation.

A second, more focused and more empiricalstream of research involved testing variousleadership theories across different nationalgroups. In Japan, Misumi (1984; Misumi &Peterson, 1985) conducted an extensive pro-gram of laboratory and field research, influencedby work in the United States on the LBDQfactors Initiation of Structure and Consideration.Misumi adapated the LBDQ to Japanese cultureby identifying two broad classes of leaderbehavior: (a) behavior associated with workaccomplishment through direction and produc-tivity emphasis (called Performance) and (b)behavior intended to maintain high groupmorale (called Maintenance). Misumi and Peter-son (1985) reported that the most productivework groups in Japanese organizations were ledby supervisors who were high on both Perfor-mance and Maintenance behavior

Ayman and Chemers (1983) reported similarresults for Iranian managers. They factoranalyzed a Persian translation of the LBDQ towhich they added some probe items related tothe tendency for the worker to identify thesupervisor in fatherly terms (i.e., "My supervi-sor is like a kind father to me"). Ayman andChemers found that structuring, consideration,and the new items collapsed into a single factor,which they labeled Benevolent Paternalism andwhich was strongly associated with subordinatesatisfaction and performance ratings by superi-ors. Ayman and Chemers concluded that subor-dinates in highly collectivistic and power-oriented cultures derive satisfaction from aleader who is both directive and nurturant,whereas subordinates in individualistic, low-power cultures such as the United States aremore satisfied with a leader who providesfollowers with autonomy and opportunities for

personal achievement. Here again we see thatthe successful leader is the one who providessubordinates with an atmosphere conducive tothe fulfillment of the followers' personal needsand goals but that the nature of those needs andgoals is influenced by culturally socializedvalues.

The 80 or 90 years of leadership researchbriefly described in the preceding pages cover alot of territory. From contingency theories totransformational leadership, and cognitive, gen-der, and cultural factors, a complex pattern oftheoretical and empirical material has beengenerated. The question remaining is whether acoherent integration of these seemingly dispar-ate findings is possible.

A Functional Integration

The apparent complexity of research findingsand theoretical perspectives in the field ofleadership might be reduced if one examinedthis literature by focusing on the major functionsthat leaders need to fulfill to be successful. Ibelieve that there are three such functions. Aleader must build credibility in the legitimacy ofhis or her authority by projecting an image thatarouses feelings of trust in followers (imagemanagement). A leader must develop relation-ships with subordinates that enable thosesubordinates to move toward individual andcollective goal attainment (relationship develop-ment). Finally, leaders must effectively use lieknowledge, skills, and material resources pre-sent within their group to accomplish thegroup's mission {resource deployment).

Image Management

The definition of leadership provided earlierstressed that social influence is at the core of theleadership function. Influence depends on cred-ibility. For followers to abdicate personalautonomy and allow themselves to be led, theymust believe that the leader's authority islegitimate. The information-processing modelsof leadership make clear the central role of theperceptions on which legitimacy is based.Individuals who are seen as behaving in waysthat are consistent with observer-held leadershipprototypes are afforded authority, and subse-quent perception, attention, and memory aremore likely to reaffirm the leader's legitimacy.

CHEMERS

We also know that the specific traits orbehaviors associated with leadership credibilityvary somewhat by leadership domain (e.g.,political leadership vs. business leadership) andacross cultures (Lord, Foti, & De Vader, 1984;Lord, Foti, & Phillips, 1982). Nonetheless, acertain commonality exists across leadershipprototypes. As Hollander (1958, 1964) demon-strated in his early work on leadership statusaccrual, leaders must be seen as competent intask-relevant abilities and as honest, trustwor-thy, and loyal to group norms and values. Thismakes a great deal of sense. Leaders needtask-relevant competencies to move the grouptoward a goal, and they must be trustworthy toensure that the goal pursued is in the collectiveinterest.

The literature on charismatic and transforma-tional leadership is consistent with this perspec-tive. Bass's (1985) concept of idealized influ-ence suggests that such leaders are seen ashaving exceptional abilities. R. J. House's(1977) discussion also addresses the extensiveefforts of charismatic leaders in demonstratingtheir loyalty to the group cause—frequently bytaking great risks (e.g., Martin Luther King, Jr.)or incurring significant hardship to follow thepath of the cause (e.g., Gandhi, Joan of Arc).

Relationship Development

Many leadership theories focus on the lead-er's responsibility in motivating and guidingfollowers to enable them to achieve task goals.The work of Graen and his associates (Graen,1976; Graen & Cashman, 1975; Graen, Cash-man, Ginsburgh, & Schiemann, 1978; Graen &Scandura, 1987) places the quality of theleader-follower relationship at the center ofeffective leadership.

The literature suggests that effective relation-ships are based on the provision of levels ofcoaching and guidance that are appropriate tothe follower's individual situation. The situationis determined by the follower's task-relevantskills and knowledge and personal values,needs, and goals. Patterns of cultural socializa-tion, personal histories, and contemporary fac-tors might all contribute to the follower'sreadiness for one or another type of treatment.Effective leadership, then, depends on theleader's ability to overcome potential egocentricand defensive biases to make accurate judg-

ments about where the subordinate is and wherethe subordinate needs to go next. At the highestlevels of leadership performance, these capabili-ties are captured by Bass's (1985) concepts ofintellectual stimulation (i.e., delicately targetedcoaching and guidance that arouse intrinsicmotivation) and individualized consideration(i.e., the subtle but comprehensive awareness ofthe follower's situation).

Resource Deployment

Once a leader has established credibility andmobilized follower motivation, the resultantenergies, knowledge, skills, and material re-sources must be harnessed and directed toachieve success in the group's mission. Thesuccessful deployment of the group's resourceshas two facets: first, the empowerment of theindividuals in the group, and second, theeffective interface of group processes with taskand environmental demands. Both facets areinfluenced by contingency principles.

Individuals are not always able to make themost effective use of their skills and abilities.Contemporary theories of intelligence (e.g.,Sternbcrg, 1988) suggest that effective interac-tion with an individual's environment (i.e.,successful utilization of personal resources) isinfluenced by the fit between the actor's set ofskills and knowledge and the critical demands ofthe challenge. Contingency theory researchconducted by me and my associates (Chemers &Ayman, 1985; Chemers et al., 1985, 1991) hasfound that leaders whose motivational orienta-tion (LPC score) was in match with environmen-tal factors (situational control) not only outper-formed less well-matched leaders but alsoshowed higher levels of satisfaction, morepositive mood and confidence, and lower levelsof stress and stress-related illness. This patternof findings suggests that leaders who are in agood "fit" with their leadership situation aremore confident, and more of them perform athigh levels.

My later research on leadership efficacy(Watson et al., 1996) suggests that leadershipconfidence (resulting from a good person-situation match or as a dispositional characteris-tic) is associated with high levels of teamperformance and positive evaluations by follow-ers and observers. Feelings of efficacy may beone of the primary moderators of the effects of

SPECIAL ISSUE: LEADERSHIP RESEARCH AND THEORY 39

contingency effects on leadership performanceat the personal level.

Both Fiedler's (1967; Fiedler & Chemers,1974, 1984) contingency model and Vroom andYetton's (1973) normative decision model arebuilt around the notion that internal groupprocesses, such as decision-making processes,must match with external task demands toensure high levels of group performance. Forexample, overly centralized information-process-ing strategies and autocratic decision structuresare likely to be more effective in highly routineand predictable leadership situations than theyare in more ambiguous, less predictable situa-tions requiring creative solutions to novelproblems. Effective leadership depends onrecognizing the nature of the group's environ-ment and matching group process to externaldemand.

On reflection, it appears that self-confidenceor self-efficacy might play an important role inmany aspects of leadership effectiveness. If oneexamines the three functions just discussed, onesees a role for self-confidence in each. Effectiveimage management depends on projecting theappearance of competence. Confidence in one'sabilities provides a good marker for compe-tence. When we observe competent people it isnatural to infer that their confidence is based onsome actual competence. R. J. House's (1977)analysis of charismatic leadership emphasizedthat outstanding leaders do indeed exhibit highlevels of self-confidence. Staw and Barsade(1992) reported that MBA students with morepositive emotional dispositions were seen asmore appropriate for leadership roles by ob-servers in an assessment center managementsimulation.

Confidence may also play a role in relation-ship development. Effective coaching and guid-ance are dependent on accurate perceptions ofsubordinates and cogent attributions about thecauses of their behavior and performance. Themajor impediment to such perceptual accuracyare the ego-defensive motivations created by theleader's own concern for positive evaluations.When a group performs poorly, the leader mayblame subordinates and be less attentive toproblems caused by other factors, such as thesupport structure or his or her own shortcom-ings. Confidence in one's own abilities mightallow leaders to be less concerned about such

judgments and allow for empathic relationshipswith followers.

The deployment of personal and team re-sources should be especially affected by confi-dence. On the personal level, a voluminousliterature in social psychology tells us thatconfident and optimistic people are better able tocope with environmental demands (Scheier &Carver, 1985) and stressful life events (Taylor &Brown, 1988) and are more likely to take risks(Isen, Nygren, & Ashby, 1988) and solveproblems creatively (Isen, Daubman, & Nan-icki, 1987)—all of which are characteristics thatare related to effective leadership.

At the level of team deployment that requiresthe ability to read the environment accuratelyand respond flexibly, confidence should alsoplay a role. Staw and Barsade (1992) measuredmanagement students for positive or negativedispositional affect. More upbeat individualsperformed more effectively at a business deci-sion task, integrating more information andmaking better decisions. Experiments by Guzzo(1986) and by Zaccaro, Peterson, Blair, andGilbert (1990) also show that collective efficacy(i.e., the shared perception of group members ofthe capability and effectiveness of the group)has been positively related to group perfor-mance in both experimental and organizationalsettings.

We also know that when leaders feel that theyare in a congenial, accepting environment theyare more likely to act in a directive, "take-charge" fashion (Eagly & Johnson, 1990) andthat directive leaders are more likely to makeeffective use of their cognitive abilities (Fiedler& Garcia, 1987).

Summary and a Few Conclusions

This historical overview of leadership re-search reveals the extent to which this researcharea, like many others, is influenced by periodicfashions in research theory; for example, anemphasis on traits at one time; on cognition atanother time, and so on. When we take a longerview, we are able to find common findings andstreams of thought across theoretical perspec-tives. The functional integration offered in thisarticle is an attempt to take such a perspectivedriven especially by an emphasis on whatleaders must do to be effective, that is, toinfluence followers toward goal attainment.

40 CHEMERS

My analysis argues that leaders must firstestablish the legitimacy of their authority byappearing competent and trustworthy to theirfollowers. When leaders are extremely effectivein image management they are seen as possess-ing remarkable, charismatic levels of capabilityand trust. Next, leaders must coach, guide, andsupport their followers in a way that allows thefollowers to contribute to group goal attainmentwhile satisfying their own personal needs andgoals. To do this, leaders must understand theabilities, values, and personalities of theirsubordinates, so they can provide the type ofcoaching and support that will be most effective.Sometimes leaders are so effective at creating amotivational environment that followers mergetheir personal goals with collective group goalsand are transformed in the process. Finally,effective leaders must use the skills and abilitiespossessed by themselves and their followers toaccomplish the group's mission. The first step inutilizing these resources is creating a sense ofconfidence and personal empowerment thatencourages each group member to release his orher best efforts. The second step is focusing theresultant resources on the task environment in away that provides the best fit between groupprocess and environmental demand. Sensitiveinformation processing and intelligent decisionmaking are the keys to the group environmentalinterface.

In the final paragraphs of this article I raisethe hypothesis that leadership efficacy andgroup collective efficacy may be the mostimportant contributors to each of the functionalnecessities of leadership performance. Feelingsof efficacy in the leadership role are thought tolead to calm decision making, sensitive interper-sonal relations, ambitious goal setting, boldaction, and long-term perseverance that energizeand maintain the leader and the followers toeffective common effort.

References

Ashour, A. S. (1973). The contingency model ofleadership effectiveness: An evaluation. Organiza-tional Behavior and Human Performance, 9,339-355.

Ayman, R., & Chemers, M. M. (1983). Therelationship of supervisory behavior ratings towork group effectiveness and subordinate satisfac-tion among Iranian managers. Journal of AppliedPsychology, 68, 338-341.

Bales, R. E, & Slater, P. E. (1955). Role differentia-tion in small decision-making groups. In T. Parsons(Ed.), Family, socialization, and interaction pro-cesses. Glencoe, IL: Free Press.

Bandura, A. (1982). Self-efficacy mechanism inhuman agency. American Psychologist, 37, 122-147.

Bass, B. M. (1981). Stogdill's handbook of leadership(2nd ed.). New York: Free Press.

Bass, B. M. (1985). Leadership and performancebeyond expectations. New York: Free Press.

Bass, B. M. (1998). Transformational leadership;Industry, military, and educational impact. Mahwah,NJ:Erlbaum.

Bass, B. M., & Avolio, B. J. (1990a). The implicationsof transactional and transformational leadership forindividual, team, and organizational development.In R. W. Woodman & W. A. Passmore (Eds.),Research in organizational change and develop-ment. Greenwich, CT: JAJ Press.

Bass, B. M., & Avolio, B. J. (1990b). Manual for themultifactor leadership questionnaire. Palo Alto,CA: Consulting Psychologists Press.

Bass, B. M., & Avolio, B. J. (1993). Transformationalleadership: A response to critiques. In M. M.Chemers & R. Ayman (Eds.). Leadership theoryand research: Perspectives and directions (pp.49-80). San Diego, CA: Academic Press.

Bass, B. M., Krusell, J., & Alexander, R. A. (1971).Male managers' attitudes toward working women.American Behavioral Scientist, 15, 221-236.

Bowman, G. W., Worthy, N. B., & Greyser, S. A.(1965). Are women executives people? HarvardBusiness Review, 43, 14-28.

Brown, K. A. (1984). Explaining group poor perfor-mance: An attributional analysis. Academy ofManagement Review, 9, 54-63.

Burns, J. M. (1978). Leadership. New York: Harper &Row.

Calder, B. J. (1977). An attribution theory ofleadership. In B. M. Staw & G. R. Salancik (Eds.),New directions in organizational behavior. Chi-cago: St. Clair Press.

Cantor, N., & Mischel, W. (1979). Prototypes inperson perception. In L. Berkowitz (Ed.), Advancesin experimental social psychology (Vol. 12). NewYork: Academic Press.

Carlyle, T. (1907). Heroes and hero worship. Boston:Adams. (Original work published 1841)

Chemers, M. M. (1997). An integrative theory ofleadership. Mahwah, NJ: Erlbaum.

Chemers, M. M, & Ayman, R. (1985). Leadershiporientation as a moderator of the relationshipbetween performance and satisfaction of Mexicanmanagers. Personality and Social PsychologyBulletin, //, 359-367."

Chemers, M. M., Ayman, R., Sorod, B., & Akimoto,S. (1991, July). Self-monitoring as a moderator of

SPECIAL ISSUE; LEADERSHIP RESEARCH AND THEORY 41

leader-follower relationships. Paper presented atthe International Congress of Psychology, Brus-sels, Belgium.

Chemers, M. M , Hays, R., Rhodewalt, F., &Wysocki, J. (1985). A person-environment analy-sis of job stress: A contingency model explanation.Journal of Personality and Social Psychology, 49,628-635.

Chemers, M. M., Watson, C. B., & May, S. (in press).Dispositional affect and leadership performance: Acomparison of self-esteem, optimism, and efficacy.Personality and Social Psychology Bulletin.

Cleven, W. A., & Fiedler, F. E. (1956). Interpersonalperceptions of open-hearth foremen and steelproduction. Journal of Applied Psychology, 40,312-314.

Deaux, K. (1984). From individual differences tosocial categories. American Psychologist, 39,105-116.

Eagly, A. R , & Johnson, B. T. (1990). Gender andleadership style: A meta-analysis. PsychologicalBulletin, 108, 233-256.

Eagly, A. H., & KarauT S. J. (1991). Gender and theemergence of leaders: A meta-analysis. Journal ofPersonality and Social Psychology. 60, 685-710.

Eagly, A. R , Makhijani, M. G., & Klonsky, B. G.(1992). Gender and the evaluation of leaders: Ameta-analysis. Psychological Bulletin, 111, 3-22.

Eden, D., & Leviatan, U. (1975). Implicit leadershiptheory as a determinant of the factor structureunderlying supervisory behavior scales. Journal ofApplied Psychology, 60, 736-741.

Fiedler, F. E. (1955). The influence of leader-keymanrelations on combat crew effectiveness. Journal ofAbnormal and Social Psychology, 51, 227-235.

Fiedler, F. E. (1958). Leader attitudes and groupeffectiveness. New York: McGraw-Hill.

Fiedler, F. E. (1964). A contingency model ofleadership effectiveness. In L. Berkowitz (Ed.),Advances in experimental social psychology (Vol.1, pp. 149-190). New York: Academic Press.

Fiedler, F. E. (1967). A theory of leadershipeffectiveness. New York: McGraw-Hill.

Fiedler, F. E., & Chemers, M. M. (1974). Leadershipand effective management. Glenview, IL: Scott,Foresman.

Fiedler, F. E., & Chemers, M. M. (1984). Improvingleadership effectiveness: The Leader Match con-cept (2nd ed.). New York: Wiley.

Fiedler, F. E., Chemers, M. M., & Maher, L. (1976).Improving leadership effectiveness: The LeaderMatch concept. New York: Wiley.

Fiedler, F. E., & Garcia, J. E. (1987), New approachesto effective leadership: Cognitive resources andorganizational performance. New York: Wiley.

Field, R. H. G., & House, R. J. (1990). A test of theVroom-Yetton model using manager and subordi-

nate reports. Journal of Applied Psychology, 75,362-366.

Fiske, A. P. (1991). Structures of social life: The fourelementary forms of human relations. New York:Free Press.

Fleishmann, E. A., & Harris, E. F. (1962). Patterns ofleadership behavior related to employee greivancesand turnover. Personnel Psychology, 15, 43-54.

Graen, G. (1976). Role making processes withincomplex organizations. In M. D. Dunnette (Ed.),Handbook of industrial and organizational psychol-ogy. Chicago: Rand McNally.

Graen, G., Alvarez, K., Orris, J. B., & MarteUa, S. A.(1970). Contingency model of leadership effective-ness: Antecedent and evidential results. Psychologi-cal Bulletin, 74,285-296.

Gracn, G., & Cashman, J, (1975). A role-makingmodel of leadership in formal organizations: Adevelopmental approach. In J. G. Hunt & L. L.Larson (Eds.), Leadership frontiers. Kent, OH:Kent State University Press.

Graen, G., Cashman, J. E, Ginsburgh, S., &Schiemann, W. (1978). Effects of linking-pinquality on the quality of working life of lowerparticipants: A longitudinal investigation of themanagerial understructure. Administrative ScienceQuarterly, 22, 491-504.

Graen, G., & Scandura, T. A. (1987). Toward apsychology of dyadic organizing. Research inOrganizational Behavior, 9, 175-208.

Green, S. G., & Mitchell, T. R. (1979). Attribution^processes of leaders in leader-member interac-tions. Organizational Behavior and Human Perfor-mance, 23, 429-458.

Griffin, R. N. (1981). Relationships among indi-vidual, task design, and leader behavior variables.Academy of Management Journal, 23, 665-683.

Guzzo, R. A. (1986). Group decision making andgroup effectiveness in organizations. In P. Good-man & Associates (Eds.), Designing effective workgroups (pp. 34-71). San Francisco: Jossey-Bass.

Hackman, J. R., & Oldham, G. R. (1976). Motivationthrough the design of work: Test of a theory.Organizational Behavior and Human Perfor-mance, 16, 250-279.

Halpin, A. W., & Winer, B. J. (1957). A factorial studyof the leader behavior descriptions. In R. M.Stogdill & A. E. Coons (Eds.), Leader behavior: Itsdescription and measurement. Columbus: OhioState University Bureau of Business Research.

Hastorf, A. R , Schneider, D., & Polefka, J. (1970).Person perception. Reading, MA: Addison-Wesley.

Heilman, M. E., Block, C. J., Martell, R. F., & Simon,M. C. (1989). Has anything changed? Currentcharacterizations of men, women, and managers.Journal of Applied Psychology, 74, 935-942.

42 CHEMERS

Hemphill, J. K. (1950). Leader behavior description.Columbus: Ohio State University Personnel Re-search Board.

Hennig, M., & Jardim, A. (1977). The managerialwoman. New York: Anchor Press.

Hofstede, G. (1980). Culture's consequences; Interna-tional differences in work-related values. BeverlyHills, CA: Sage.

Hofstede, G. (1983). Motivation, leadership, andorganization: Do American theories apply abroad?Organizational Dynamics, 9, 42-63.

Hogg, M. A., Hains, S. C , & Mason, I. (1998).Identification and leadership in small groups:Salience, frame of reference, and leader stereotypi-cality effects on leadership evaluations. Journal ofPersonality and Social Psychology, 75, 1248-1263.

Hollander, E. P. (1958). Conformity, status, andidiosyncrasy credit. Psychological Review, 65,117-127.

Hollander, E. P. (1964). Leaders, groups, andinfluence. New York: Oxford University Press.

Hollander, E. P., & Julian, J. W. (1970). Studies inleader legitimacy, influence, and innovation. In L.Berkowitz (Ed.), Advances in experimental socialpsychology (Vol. 5). New York: Academic Press.

House, J. (1981). Social structure and personality. InM. Rosenberg & R. Turner (Eds.), Social psychol-ogy: Sociological perspectives (pp. 525-561). NewYork: Basic Books.

House, R. J. (1971). Apath-goal theory of leadership.Administrative Science Quarterly, 16, 321-338.

House, R. J. (1977). A 1976 theory of charismaticleadership. In J. G. Hunt & L. L. Larson (Eds.),Leadership: The cutting edge. Carbondale: South-ern Illinois University Press.

House, R. J., & Dessler, G. (1974). The path-goaltheory of leadership: Some post-hoc and a prioritests. In J. G. Hunt & L. L. Larson (Eds.),Contingency approaches to leadership. Carbon-dale: Southern Illinois University Press.

House, R. J., & Mitchell, T. R. (1974). Path-goaltheory of leadership. Journal of ContemporaryBusiness, 3, 81-98.

House, R. J.. & Shamir, B. (1993). Integratingtransformational, charismatic, and visionary theo-ries. In M. M. Chemers & R. Ayman (Eds.),Leadership theory and research: Perspectives anddirections (pp. 81-108). San Diego, CA: AcademicPress.

Isen, A. M., Daubman, K. A., & Nanicki, G. P. (1987).Positive affect facilitates creative problem solving.Journal of Personality and Social Psychology, 51,1122-1131.

Isen, A. M., Nygren, J. E., & Ashby, G. F. (1988). Theinfluence of positive affect on the subjective utilityof gains and losses: It's not worth the risk. Journal

of Personality and Social Psychology, 55, 710—717.

Jones, E. E., & Davis, K. E. (1965). From acts todispositions: The attribution process in personperception. In L. Berkowitz (Ed.), Advances inexperimental social psychology (Vol. 2). NewYork: Academic Press.

Jones, E. E., & Nisbett, R. E. (1971). The actor andthe observer: Divergent perceptions of the causesof behavior. Morristown, NJ: General LearningPress.

Kahn, R. L. (1951). An analysis of supervisorypractices and components of morale. In H.Guetzkow (Ed.), Groups, leadership, and men.Pittsburgh, PA: Carnegie Press.

Kelley, H. H. (1967). Attribution theory in socialpsychology. In D. Levine (Ed.), Nebraska sympo-sium on motivation. Lincoln: University of Ne-braska Press.

Kerr, S., & Jermier, J. M. (1978). Substitutes forleadership: Their meaning and measurement. Orga-nizational Behavior and Human Performance, 22,375-403.

Korman, A. K. (1966). "Consideration," "initiatingstructure," and organizational criteria. PersonnelPsychology, 18, 349-360.

Lewin, K., Lippitt, R., & White, R. K. (1939).Patterns of aggressive behavior in experimentallycreated social climates. Journal of Social Psychol-ogy, 10, 271-301.

Lord, R. G. (1985). An information processingapproach to social perceptions, leadership, andbehavioral measurement in organizations. In B. M.Staw & L. L. Cummings (Eds.), Research inorganizational behavior (Vol. 7). Greenwich, CT:JA1 Press.

Lord, R. G., Binning, J. F., Rush, M. C , & Thomas, J.C. (1978). The effect of performance cues andleader behavior on questionnaire ratings of leader-ship behavior. Organizational Behavior and Hu-man Performance, 21, 27-39.

Lord, R. G., Foti, R. J., & De Vader, C. (1984). A testof leadership categorization theory: Internal struc-ture, information processing, and leadership percep-tions. Organizational Behavior and Human Perfor-mance, 34, 343-378.

Lord, R. G., Foti, R. J., & Phillips, i. S. (1982). Atheory of leadership categorization. In J. G. Hunt,U. Sekaran, & C. Schriesheim (Eds.), Leadership:Beyond establishment views. Carbondale: SouthernIllinois University Press.

Lord, R. G., & Maher, K. J. (1991). Leadership andinformation processing: Linking perceptions andperformance. Boston: Unwin Hyman.

Markus, H. R., & Kitayama, S. (1991). Culture andthe self: Implications for cognition, emotion, andmotivation. Psychological Bulletin, 98, 224—253.

SPECIAL ISSUE: LEADERSHIP RESEARCH AND THEORY 43

Meindl, J. R. (1990). On leadership: An alternative tothe conventional wisdom. In B. A. Staw (Ed.),Research in organizational behavior (Vol. 12, pp.159-203). New York: JAI Press.

Misumi, J. (1984). The behavioral science ofleadership (2nd ed.). Tokyo: Yuhikaku.

Misumi, J., & Peterson, M. F. (1985). The perfor-mance-maintenance (PM) theory of leadership:Review of a Japanese research program. Adminis-trative Science Quarterly, 30, 198-223.

Mitchell, T. R., Larson, J. R., & Green, S. G. (1977).Leader behavior and situational moderators ingroup performance: An attributional analysis.Organizational Behavior and Human Perfor-mance, 18, 254-268.

Mitchell, T. R., & Wood, R. E. (1980). Supervisors'responses to subordinates' poor performance: Atest of an attribution model. Organizational Behav-ior and Human Performance, 25, 123-138.

Nieva, V. R, & Gutek, B. A. (1981). Women andwork: A psychological perspective. New York:Praeger.

Peters, L. H., Hartke, D. D., & Pohlmann, J. T.(1983). Fiedler's contingency theory of leadership:An application of the meta-analysis procedure ofSchmidt and Hunter. Psychological Bulletin, 97,274-285.

Phillips, J. S., & Lord, R. G. (1981). Causalattributions and perceptions of leadership. Organi-zational Behavior and Human Performance, 28,143-163.

Podsakoff, P. M., Niehoff, B. P., MacKenzie, S. B., &Williams, M. L. (1993). Do substitutes forleadership really substitute for leadership? Anempirical examination of Kerr and Jermier'ssituational leadership model. Organizational Behav-ior and Human Decision Processes, 54, 1—44.

Ragins, B. R. (1989). Power and gender congruencyeffects in evaluations of male and female manag-ers. Journal of Management, 15, 65-76.

Ragins, B. R. (1991). Gender effects in subordinateevaluations of leaders: Real or artifact? Journal ofOrganizational Behavior, 12, 258—268.

Ragins, B. R., & Sundstrom, E. (1989). Gender andpower in organizations: A longitudinal perspective.Psychological Bulletin, 105, 51-88.

Rosch, E. (1978). Principles of categorization. In E.Rosch & B. B. Lloyd (Eds.), Cognition andcategorization. Hillsdale, NJ: Erlbaum.

Ross, L. (1978). The intuitive psychologist and hisshortcomings. In L. Berkowitz (Ed.), Cognitivetheories in social psychology. New York: Aca-demic Press.

Rush, M. C , Thomas, J. C , & Lord, R. G. (1977).Implicit leadership theory: A potential threat to theinternal validity of leader behavior questionnaires.

Organizational Behavior and Human Perfor-mance, 20, 93-110.

Scheier, M. E, & Carver, C. S. (1985). Optimism,coping, and health: Assessment and implications ofgeneralized outcome expectancies. Health Psychol-ogy, 4, 219-247.

Schein, V. E. (1973). The relationship of sex rolestereotypes and requisite management characteris-tics. Journal of Applied Psychology, 57, 95-100.

Schein, V. E. (1975). The relationship of sex rolestereotypes and requisite management characteris-tics among female managers. Journal of AppliedPsychology, 60, 340-344.

Staw, B. M. (1975). Attribution of the "causes" ofperformance: A general alternative interpretation ofcross-sectional research on organizations. Organi-zational Behavior and Human Performance, 13,414-432.

Staw, B. M., & Barsade, S. G. (1992). Affect andmanagerial performance: A test of the sadder-but-wiser vs. happier-and-smarter hypothesis. Adminis-trative Science Quarterly, 38, 304-331.

Stemberg, R. J. (1988). The triarchic mind: A newtheory of human intelligence. New York: Viking.

Stogdill, R. M. (1948). Personal factors associatedwith leadership: A survey of the literature. Journalof Psychology, 25, 35-71.

Strube, M. J., & Garcia, J. E. (1981). A meta-analytical investigation of Fiedler's contingencymodel of leadership effectiveness. PsychologicalBulletin, 90, 307-321.

Taylor, S. E., & Brown, J. (1988). Illusion andwell-being: A psychological perspective on mentalhealth. Psychological Bulletin, 103, 193-210.

Triandis, H. C. (1990). Cross-cultural studies ofindividualism and collectivism. In J. Berman (Ed.),Nebraska symposium on motivation (pp. 41-133).Lincoln: University of Nebraska Press.

Triandis, H. C. (1993). The contingency model incross-cultural perspective. In M. M. Chemers & R.Ayman (Eds.), Leadership theory and research:Perspectives and directions (pp. 167-188). SanDiego, CA: Academic Press.

Vroom, V. H., & Yetton, P. W. (1973). Leadership anddecision-making. Pittsburgh, PA: University ofPittsburgh Press.

Watson, C. B., Chemers, M. M., & Preiser, N. (1996,June). Collective efficacy: A multi-level analysis.Paper presented at the annual meeting of theAmerican Psychological Society, San Francisco.

Zaccaro, S. J., Peterson, C , Blair, V., & Gilbert, J. A.(1990, August). Some antecedents and conse-quences of collective control. Paper presented atthe 98th Annual Convention of the AmericanPsychological Association, Boston.