In 1945, C. L. Shartle (1950b) launched the Ohio State Leadership Studies. He was influenced by his work on gathering occupational information to describe tasks, jobs, and occupational requirements in all echelons of industry, government, and the military. He focused on defining and classifying work done in each of the 30,000 occupations that had appeared originally in the 1939 Dictionary of Occupational Titles and in the U.S. Military Occupational Specialties. He and his colleagues now applied this background to the study of leaders. This was different from the emphasis in previous leadership research, which had sought to identify the traits of leaders. Reviews by Bird (1940), W. O. Jenkins (1947), and Stog-dill (1948) had concluded that the personality-traits approach had reached a dead-end, for several reasons: (1) attempts to select leaders in terms of traits had had little success; (2) numerous traits differentiated leaders from followers; (3) the traits demanded of a leader varied from one situation to another; and (4) the traits approach ignored the interaction between the leader and his or her group. Attention was shifted to what leaders did.
Needed were descriptions of individuals’ actions when they acted as leaders of groups or organizations. Hemp-hill (1949a) had already initiated such work at the University of Maryland. After joining the Ohio State Leadership Studies, Hemphill and his associates developed a list of approximately 1,800 statements that described different aspects of the behavior of leaders, such as “He insists on meeting deadlines.” Most statements were assigned to several subscales. However, staff members agreed on 150 statements that could each be assigned to only one subscale. These statements were used to develop the first form of the Leader Behavior Description Questionnaire, or LBDQ (Hemphill, 1950a; Hemp-hill & Coons, 1957). On the LBDQ, respondents rated a leader by using one of five alternatives to indicate the fre-quency or amount of a particular behavior that was descriptive of the leader being rated. Responses to items were simply scored and added in combinations to form subscales on the basis of the similarity of their content. These subscale totals were then intercorrelated and factor-analyzed (Fleishman, 1951, 1953c; Halpin & Winer, 1957). “Consideration” and “initiation of structure” were primary factors identified by Halpin and Winer (1957) with regard to Air Force officers, and by Fleishman (1951, 1953c, 1957) with regard to industrial supervisors. They emerged in successive factor studies (Fleishman, 1973) as the two most prominent factors that described leaders according to questionnaires completed by themselves and others.
This first factor describes the extent to which a leader exhibits concern for the welfare of the other members of the group. The considerate leader expresses appreciation for good work, stresses the importance of job satisfaction, maintains and strengthens the self-esteem of subordinates by treating them as equals, makes special efforts to help subordinates feel at ease, is easy to approach, puts subordinates’ suggestions into operation, and obtains subordinates’ approval on important matters before going ahead. Considerate leaders provide support that is oriented toward relationships, friendship, mutual trust, and interpersonal warmth. Participation and the maintenance of the group accompany such support (Atwater, 1988). In contrast, the inconsiderate leader criticizes subordinates in public, treats them without considering their feelings, threatens their security, and refuses to accept their suggestions or to explain his or her actions.
This second factor shows the extent to which a leader initiates activity in the group, organizes it, and defines the way work is to be done. Initiation of structure includes such leadership behavior as insisting on maintaining standards and meeting deadlines and deciding in detail what will be done and how it should be done. Clear channels of communication and clear patterns of work organization are established. Orientation is toward the task. The leader acts directively without consulting the group. Particularly relevant are defining and structuring the leader’s own role and the roles of subordinates in attaining goals. The leader whose factor score in initiating structure is low is described as hesitant about taking initiatives in the group. He or she fails to take necessary actions, makes suggestions only when members ask for it, and lets members do the work the way they think best.
The LBDQ, consisting of 40 statements, was designed to measure the two factors of consideration and initiation (Hemphill & Coons, 1957). An industrial version, the Supervisory Behavior Description Questionnaire (SBDQ), followed (Fleishman, 1972), as did LBDQ—Form XII, hereafter referred to as LBDQ-XII (Stogdill, 1963a). Each of the LBDQs had instructions such as these:
The following … are items that may be used to describe the behavior of your leader or supervisor. Each item describes a specific kind of behavior, but does not ask you to judge whether the behavior is desirable or undesirable. This is not a test of ability. It simply asks you to describe, as accurately as you can, the behavior of your supervisor or leader. Group refers to an organization or to a department, division, or other unit of organization that is supervised by the person being described. … Members refer to all the people in the unit of organization that is supervised by the person being described.
THINK about how frequently the leader engages in the behavior described by the item.
DECIDE whether he (A) always, (B) often, (C) occasionally, (D) seldom, or (E) never acts as described by the item.
Typical items were: “He lets group members know what is expected of them.” “He is friendly and approachable.”
The intentions of the developers of the LBDQs and the SBDQ, particularly with regard to initiating structure, differed somewhat in the construction of the different versions. The LBDQ contained a subset of 15 items that asked subordinates to describe the actual structuring behavior of their leader. This structuring behavior was the leader’s behavior in delineating relationships with subordinates, in establishing well-defined patterns of communication, and in detailing ways to get a job done (Halpin, 1957b). The Supervisory Behavior Description Questionnaire (SBDQ) consisted of 20 items that included asking subordinates about their leader’s actual structuring behavior. Initiating structure, as measured by the SBDQ, was intended to reflect the extent to which the leader organizes and defines interactions among group members, establishes ways to get the job done, schedules, criticizes, and so on (Fleishman, 1972). The SBDQ items for the factor of initiation of structure included a wider variety of structuring behaviors drawn from the loadings on this factor. Several were close to Misumi’s (1985) Production Pressure. Items on the LBDQ came mostly from original conceptualizations about communication and organization (Schriesheim, House, & Kerr, 1976). The revised LBDQ-XII had 10 items that measured initiation of structure in terms of the actions of leaders who clearly define their own role and let followers know what is expected of them (Stogdill, 1963a).
As noted, ratings of the consideration and initiation of structure by leaders are highly stable and consistent from one situation to another (Taylor, Crook, & Dropkin, 1961; and Philipsen, 1965a). According to Schriesheim and Kerr’s (1974) review of the psychometric properties of the LBDQ and SBDQ, the descriptions maintain the high internal consistency that was the basis for their construction. That is, items on the “consideration behavior” in each instrument correlate highly with all the other consideration items and do not correlate with items on the “initiation behavior” factor. Conversely, items on the “initiating structure” factor are independent of the “consideration” items and are highly intercorrelated with all the other structuring items. Using a sample of 308 public utilities employees, Schriesheim (1979a) found that consideration and initiation were so psychometrically robust that it did not make much difference in a supervisor’s scores whether one asked a subordinate to describe how the supervisor behaved toward him or her personally (the dyadic approach) or how the supervisor behaved toward the whole work group (the standard approach).
Nonetheless, the factor scores left a lot to be desired. They suffered from halo effects and were plagued by a variety of other response errors, such as leniency and social desirability, as well as a response set to agree rather than to disagree (Schriesheim, Kinicki, & Schriesheim, 1979). It was not known whether they were valid measures of true consideration and initiation of structure. Most important, as research with the original LBDQ and SBDQ instruments continued, it became apparent that some of what leaders do had been missed. A great deal of the behavior of leaders was being lost in the emphasis on just two factors to account for all the common variance among items describing this behavior. Therefore, for LBDQ-XII, various additional factored scales were constructed, possibly lacking complete independence from structuring and consideration, yet likely to include much of the missing information of consequence.
Comparison of the Three Forms. All three versions—LBDQ, SBDQ, and LBDQ-Form XII—have been used extensively, and each has been subjected to additional factor analyses (Bish & Schriesheim, 1974; Szilagyi & Sims, 1974a; Tscheulin, 1973). A direct comparison of all three became possible after a survey and factor-analytic study of 242 hourly employees by Schriesheim and Stogdill (1975). This comparison study was necessary, since, as Korman (1966) and others had noted, the content of the scales varied, which caused differences in outcomes. The original LBDQ and particularly the SBDQ contained several items such as “Needles subordinates for greater effort” and “Prods subordinates for production” that measured punitive, arbitrary, coercive, and dominating behaviors and that affected the scores for initiation of structure. The LBDQ-XII was considered to be most nearly free of such autocratic items (Schriesheim & Kerr, 1974). As has usually been found, internal consistency reliabilities were high for scores on both factors that were derived from the items drawn according to their use in the LBDQ, SBDQ, or LBDQ-XII. For consideration and initiation, reliabilities were .93 and .81 for the LBDQ, .81 and .68 for the SBDQ, and .90 and .78 for the LBDQ-XII. (The reliability of .68 for initiation of structure on the SBDQ was raised to .78 when three SBDQ punitive items were removed from the scoring of the scale.) The primary factors that were extracted indicated that all three versions contained some degree of arbitrary punitive performance (“The leader demands more than we can do”). But, as expected, the pattern was most marked in the SBDQ. A hierarchical factor analysis disclosed the existence of a higher-order factor of rater bias, which appeared in all three questionnaires.
According to Schriesheim and Kerr, the items loading highest on the SBDQ in initiating structure were: “He insists that he be informed on decisions made by people under him” (.60), “He insists that people under him follow standard ways of doing things in every detail” (.65), and “He stresses being ahead of competing work groups” (.64). In fact, Atwater (1988) found that two items—“Demands a great deal from his workers” and “Pushes his workers to work harder”—correlated highest with the SBDQ factor she obtained and labeled demanding behavior. But on the LBDQ, the two items loading highest on initiating structure were: “He maintains definite standards of performance” (.59) and “He lets members know what is expected of them” (.54).
Matters further were complicated because many researchers deleted items or modified the wording of items for use in a particular study, as Podsakoff and Schriesheim (1985) found in their review of studies that used the LBDQ. Also, many researchers failed to specify which version of the LBDQ they had used or how they had modified the scales (Hunt, Osborn, & Schriesheim, 1978).
The items ask how the leader acts toward the work group rather than toward specific individuals. Critics, such as Graen and Schiemann (1978), assume that there are large variations in the leader’s behavior toward different individual members of a work group, and therefore that the wording of items should allow for this. Individual differences among raters also underlie the ratings. Although WABA analysis could handle this, it was not introduced until 1984 (see Chapter 16). The previously cited findings of Schriesheim (1979a) suggested, however, that the matter may be overblown. D. M. Lee (1976) asked 80 students to judge the initiation and consideration of their En glish professors over an eight-week period. The results indicated that the individual students differed widely in the cues they used as the basis of their ratings of the same professors. The questionnaires fail to weight the timing, appropriateness, importance, and specificity or generality of responses. They may assess the circumstantial requirements of the job, rather than the leader as a person with discretionary opportunities to behave in the manner indicated.
Leniency Effects. Seeman (1957) reported that the LBDQ scales suffered from halo effects. Even making items more detailed was of no help in reducing the halo. Schriesheim, Kinicki, and Schriesheim (1979) completed five studies of the extent to which consideration and initiation of structure were biased by leniency effects. They inferred from the results that leniency response bias—the tendency to describe others in favorable but probably untrue terms—did not particularly affect descriptions of initiation of structure. But even though consideration and leniency are conceptually distinct, they concluded that (1) consideration items were not socially neutral and were susceptible to leniency, (2) consideration reflected an underlying leniency factor when applied in a field setting, and (3) leniency explained much or most of the variance in consideration. Leniency may explain why consideration tends to correlate higher with other evaluative variables than does initiation of structure (Fleishman, 1973).
Implicit Theories. D. J. Schneider (1973), among others, suggested that respondents express their own implicit theories about and stereotypes of leaders and leaders’ behavior, rather than the behavior of the specific leader they are supposed to be describing with the LBDQ. That is, respondents describe their idealized prototype of a leader, rather than the actual leader they should be describing (Rush, Thomas, & Lord, 1977). Eden and Leviatan (1975) noted that leader-behavior descriptions of a fictitious manager using the Survey of Organizations resulted in a factor structure highly similar to that obtained from descriptions of real managers reported by Taylor and Bowers (1972).1 Rush, Thomas, and Lord (1977) found a high degree of congruence between factor structures obtained from descriptions of a fictitious supervisor using LBDQ-XII and descriptions of real leaders from a field study by Schriesheim and Stogdill (1975). In both studies, the authors concluded that since practically identical factor structures emerged for fictitious and specific real leaders, the actual behavior of a leader is relatively unimportant for behavioral descriptions, because descriptions are based mainly on implicit theories or stereotypes.
One might suggest that this tendency to project or to use implicit theories tells more about the subordinate than about the leader. But the tendency is a consequence of ambiguity and of a lack of specific information about the leader to be rated. Schriesheim and DeNisi (1978) studied 110 bank employees and 205 workers in a manufacturing plant who used the LBDQ-XII to describe supervisors in general after first describing their own supervisors. The investigators, as expected, found comparable factors emerging from the general and specific descriptions when each was analyzed separately. However, separate real and imaginary factors emerged when the combined data were subjected to an analysis that provided an opportunity for statistical differentiation. As before, in both real and imaginary descriptions, initiation of structure and consideration was correlated above .50, with reliabilities ranging from .84 to .87.
Schriesheim and DeNisi discovered that although satisfaction with one’s real supervisor correlated between .51 and .75 with descriptions of the actual consideration and initiation of the real supervisors, it correlated only .23 and –.03 with scores for initiating structure and consideration for the stereotypes of supervisors. This was a particularly important finding.
Subsequent experimentation by Schriesheim and DeNisi (1978) with 360 undergraduates strongly supported the contention that as more specific information became available to them, the respondents’ LBDQ responses became more accurate and were less likely to depend on implicit theories. These results are consistent with those of Bass, Valenzi, Farrow, and Solomon (1975), who found that subordinates describing the same real leader were in much more significant agreement with each other than with subordinates who described other leaders.
Self-Ratings Unrelated to Subordinates’ Ratings. As many researchers reported for self-rated directive versus participative leadership, for autocratic versus democratic leadership, and for task-versus relations-oriented leadership,2 little relation was obtained between leaders’ self-descriptions of their own initiation and consideration and their subordinates’ descriptions on the LBDQ. Similarly, there was only a weak relation between what leaders say they should do, according to scores on the Leadership Opinion Questionnaire, and what they actually do, according to their subordinates’ descriptions of them on the LBDQ (Schriesheim & Kerr, 1974).
Theoretically, given the original orthogonal factor structure, consideration and initiation of structure should be independent, but this is not the case. Schriesheim, House, and Kerr (1976) reexamined Weissenberg and Kavanagh’s (1972) review of the data, along with work published subsequently. In 11 of 13 studies using the LBDQ, a positive correlation was reported. The median correlation for the 13 analyses was .45. Likewise, for the LBDQ-XII for 10 studies, the median correlation was .52 between consideration and initiation of structure. The correlation was even higher in situations where job pressure was strong. In addition, the correlations between consideration and initiation were positive both when group-by-group values were correlated and when the scores for individuals within the groups were correlated (Katerberg & Hom, 1981). Of the 16 studies that used the SBDQ, which includes some “autocratic” items, 11 studies yielded some significant negative correlations between consideration and initiation. However, the median correlation was –.05. Without the punitive items, consideration and initiation on the LBDQ tend to correlate more positively. A comprehensive review (Fleishman, 1989a), which included 32 studies with the SBDQ, found a median correlation of –.02 between the score for consideration and the score for structure.
Oldham (1976), among others, developed alternative and additional scales to provide a more detailed profile of leader behavior. The scales included behaviors such as these: personally rewarding, personally punishing, setting goals, designing feedback systems, placing personnel, and designing job systems. These scales were higher in relation to effectiveness than were measures of consideration and the initiation of structure. Seltzer and Bass (1987) found that the transformational leadership factors of charisma, individualized consideration, and intellectual stimulation added substantially to the effects of consideration on subordinates’ satisfaction and effectiveness. Seltzer and Bass conceived of consideration and initiation of structure as primarily transactional.
Halpin and Croft (1962) also were not convinced that the behavior of leaders could be adequately described with just two factors. Using items containing additional content about school principals, as well as items from the LBDQ, they extracted four factors to account for the common variance in the obtained descriptions of school principals’ behavior: (1) aloofness, formality, and social distance; (2) Production emphasis—pushing for results; (3) Thrust—personal hard work and task structure; and (4) Consideration—concern for the comfort and welfare of followers. These factored scales for describing the behavior of school principals were supplemented by the following four scales used to describe the behavior of teachers: (1) Disengagement—clique formation, withdrawal, (2) Hindrance—frustration from routine and overwork, (3) Esprit—high morale, enthusiasm, and (4) Intimacy, mutual liking, and teamwork.
When Halpin and Croft classified 71 schools into six categories according to climate, they found that an open school climate was associated with the esprit of teachers under a principal who was high in thrust. An autonomous climate produced intimacy in teachers under an aloof principal. A controlled climate resulted in hindrance in teachers under a principal who pushed for production. A familiar climate was associated with the disengagement of teachers under a considerate principal. A climate with potential but with the disengagement of teachers resulted in a principal who exhibited consideration, along with an emphasis on production. A closed climate, also with the disengagement of teachers, was associated with an aloof principal. These results yielded considerably more insight into the dynamic interplay among the climate of schools, the behavior of leaders, and the response of teachers than could be produced by the use of just two factors to describe the behavior of leaders.
The most direct expansion from consideration and initiation of structure to a broader array of leader-behavior dimensions was the development of LBDQ-XII.
On the basis of a theoretical analysis of the differentiation of roles in groups, Stogdill (1959) proposed 10 additional patterns of behavior involved in leadership, conceptually independent of consideration and initiation of structure, to be included in the LBDQ-XII along with consideration and initiation of structure. These patterns were: (1) Representation—speaks and acts as the representative of the group; (2) Reconciliation—reconciles conflicting organizational demands and reduces disorder in the system; (3) Tolerance of uncertainty—is able to tolerate uncertainty and postponement without anxiety or upset; (4) Persuasiveness—uses persuasion and argument effectively; exhibits strong convictions; (5) Tolerance of freedom—allows followers scope for initiative, decisions, and action; (6) Role retention—actively exercises the leadership role, rather than surrendering leadership to others; (7) “Production emphasis”—applies pressure for productive output; (8) Predictive accuracy—exhibits foresight and the ability to predict outcomes accurately; (9) Integration—maintains a close-knit or-ganization and resolves intermember conflicts; and (10) Influence with supervisors—maintains cordial relations with superiors; has influence with them; strives for higher status. Subsequently, these patterns became the 10 scored factors of LBDQ-XII. The conception of the persuasiveness pattern anticipated the more recent focus on the measurement of charismatic and inspirational leadership.
Interdescriber Agreement. In a study of a governmental organization by Day (1968), high-ranking administrators were each described by two male and two female subordinates. Correlations were computed to determine the extent to which pairs of subordinates agreed with each other in their descriptions of their immediate superiors. The greatest agreement was shown by pairs of female subordinates describing their female superiors. Their correlations ranged from .39 (integration) to .73 (retention of the leadership role). The least agreement was shown by pairs of male subordinates describing their female superiors. Their correlations ranged from –.02 (tolerance of freedom) to .53 (retention of the leadership role). Pairs of female subordinates tended to exhibit higher degrees of agreement than male subordinates in descriptions of male superiors on 8 of the 12 scales. The four exceptions occurred for representation, production emphasis, integration, and influence with superiors. The scales with the highest degrees of interdescriber agreement across groups of raters, male or female, were demand reconciliation, tolerance of uncertainty, persuasiveness, role retention, predictive accuracy, and influence with superiors. The scales with the lowest degrees of agreement across samples were representation, tolerance of freedom, and integration.
Divergent Validities. To test the divergent validities of several scales of the LBDQ-XII, Stogdill (1969), with the assistance of a playwright, wrote a scenario for each of six scales (consideration, structure, representation, tolerance of freedom, production emphasis, and superior orientation). The items in each scale were used as a basis for writing the scenario for that pattern of behavior. Experienced actors played the roles of supervisor and workers. Each role was played by two actors, and each actor played two different roles. Motion pictures were made of the performances. Observers used LBDQ-XII to describe the “supervisor’s” behavior. No significant differences were found between two actors playing the same role. Still, the actors playing a given role were described as behaving significantly more like that role than the other roles. Stogdill concluded that the scales measured what they purported to measure.
Factor Validation of LBDQ-XII.3 Data collected by Stogdill, Goode, and Day (1963a, 1963b, 1964, 1965) used nine of the LBDQ-XII scales to obtain descriptions of the leadership behavior of U.S. senators, corporation presidents, presidents of international labor unions, and presidents of colleges and universities. For leaders in each setting, the scores for the scales were intercorrelated and factor-analyzed. In general, the results suggested that each factor was strongly dominated by a single appropriate scale. For example, in all four analyses the representation factor emerged with only the representation scale correlated highly with it. The representation scale correlated, respectively, with the representation factor, .80, .94, .92, and .92, in the four locales. Similarly, the role retention subscale correlated only with its own role retention factor, .89, .93, .81, and .92. However, production emphasis tended to load highly on initiation of structuring as well as on production emphasis.
“Reconciliation of conflicting demands” failed to emerge as a factor differentiating college presidents, presumably because they all were described similarly highly in this behavior. “Orientation to superiors,” of course, did not fit with the role of senators; nor did union presidents differ from each other in orientation to a higher authority. The differences in predictive accuracy generated the factor “predictive accuracy” for only the corporate and union leaders, not the senators or college presidents.
Slightly different results emerged when all the items of the LBDQ-Form XII were intercorrelated and factor-analyzed for three additional locales. Eight factors emerged: (1) General persuasive leadership; (2) tolerance of uncertainty; (3) tolerance of followers’ freedom of action; (4) representation of the group; (5) influence with superiors; (6) production emphasis; (7) structuring expectations; (8) retention of the leadership role. In addition, two distinct factors of consideration were extracted.
The most numerous and most highly loaded items on general persuasive leadership, aside from measures of persuasiveness, were scales with items about the reconciliation of conflicting demands, structuring expectations, retention of the leadership role, influence with superiors, consideration, and production emphasis. These items represented the followers’ general impression of the leaders. In fulfilling these functions, leaders were seen as considerate of their followers’ welfare. Each of the remaining factors tended to be composed of items from a single scale, but some contained stray items from other scales. Consideration broke down into two separate factors, to be discussed later in connection with J. A. Miller’s (1973b) hierarchical factor analysis.
The first nine factors showed similar loadings for senators, union leaders and college presidents. But “retention of the leadership role” appeared as a separate factor only in the ratings of the state senators. All scales except those dealing with tolerance of uncertainty, tolerance of freedom, and representation contributed some items to the general factor. However, all scales except persuasiveness and the reconciliation of conflicting demands emerged in separate factors differentiated from each other. These findings indicated that the behavior of leaders is indeed complex in structure and that followers are able to differentiate among different aspects of behavior. The general persuasion factor provided valuable additional insight into the nature of leadership, strongly suggesting that this general factor may be particularly useful given what was said in Chapter 6 about its opposite, laissez-faire leadership.
Initiation and Consideration as Higher-Order Dimensions. A. F. Brown (1967) used the LBDQ-XII to obtain scores on each of the 12 factors for 170 principals described by 1,551 teachers in Canadian schools. He found that two higher-order factors accounted for 76% of the total factor variance for the 12 primary factors. When the loadings for the two factors were plotted against each other, “production emphasis,” “structuring expectations,” and “representation of the group” clustered about an axis of “initiation of structure” and “tolerance of uncertainty.” “Tolerance of freedom” and “consideration” clustered about an axis of “consideration.” The loadings for the remaining factors fell between the clusters at the extremes of these two orthogonal axes. A plot of factor loadings obtained for descriptions of university presidents on the LBDQ-XII (Stogdill, Goode, & Day, 1965) produced similar results. “Representation,” “structuring expectations,” “emphasis,” and “persuasiveness” clustered around the first axis and “freedom,” “uncertainty,” and “consideration” clustered around the second axis. Marder (1960) obtained a somewhat different pattern of loadings when military rather than educational leaders were studied. The data consisted of 235 descriptions of U.S. Army officers by enlisted men. “Productivity emphasis” and “initiation of structure” centered on one axis and “tolerance for freedom” and “tolerance for uncertainty” clustered around the other. “Consideration” was displaced toward the central cluster of items.
There are different schools of thought regarding the use of factor analysis. One school maintains that as much of the total factor variance as possible should be explained in terms of a general factor. Another school holds that rotational procedures, such as the varimax that reduces the magnitude of the general factor, are legitimate. The former school, while admitting that systems of events in the real world may involve a variety of factors, maintains that human perception contains a large element of bias and halo that should be removed in the general factor before any attempt to determine the structure of measurements representing the real world. The second school argues that the apparent halo in the general factor has its equivalent in the opacity of the real world and that the purpose of research is to reduce this opacity by making full use of all the structure differentiated by human perception. The structure that is perceived should not be permitted to remain hidden in the general factor.
If one prefers a two-factor theory of leadership behavior, initiation of structure, production emphasis, or persuasiveness can define one of the factors; consideration, tolerance of freedom, and tolerance of uncertainty can define the other. A two-factor solution, which leaves a considerable amount of the total variance unexplained, can always be obtained in the analyses of descriptions of leaders’ behavior. However, a multifactor solution should not be rejected until its consequences have been thoroughly explored and it has been proved untenable. Furthermore, the dilemma can be reconciled, as J. A. Miller (1973b) and Schriesheim and Stogdill (1975) showed, by recourse to hierarchical factor analysis. The former used rotation and differentiation; the latter, the general evaluative bias factor. Finally, the positive association routinely found (as noted earlier) between consideration and initiation of structure, as measured by LBDQ-XII, suggests that a single, general factor solution may be warranted. Nevertheless, with reference to the contents of the LBDQs, the two-factor framework for describing leadership behavior—consideration and the initiation of structure—emerges consistently from factor analyses when no additional constraints are placed on the analyses, such as first requiring the isolation of a general factor that, no doubt, has strong connections with the respondents’ prototypes of leaders.
Consideration and initiating structure can be finely factored in a number of ways by adding detailed behaviors, other than those found on the LBDQ-XII, and pursuing reconceptualizations about consideration and initiation. As was just noted, Stogdill (1963a) added new content dealing with different domains of leadership behavior to obtain the 10 additional scales for LBDQ-XII. More detail about initiation and consideration can also intensify the analysis of the basic content of initiation and consideration and related measures. Yukl (1971) demonstrated the feasibility of a three-factor approach (consideration, initiation of structure, and centralization of decisions). Saris (1969) offered “responsibility reference” and Karmel (1978) offered “active engagement” as a third factor. Another three-factor approach—initiating structure, participation, and decision making—was pursued by R. H. Johnson (1973). Wofford (1971) expanded the framework of leadership behavior to five factors: (1) group achievement and order, (2) personal enhancement, (3) personal interaction, (4) dynamic achievement, and (5) security and maintenance.
Using several thousand members of a nationwide business fraternity, who described their leaders on both a new instrument (FFTQ) and the comparable scales of the LBDQ-XII, Yukl and Hunt (1976) demonstrated some degree of communality between factors and scales purporting to deal with similar dimensions; yet overall, unfortunately, the scales were not equivalent.
Hierarchical Factor Analysis. Because of earlier reported findings that the LBDQ and supposedly similar instruments were not equivalent,4 J. A. Miller (1973b) assembled 160 items from nine frequently cited standard instruments used in published research concerning leadership behavior described in this and preceding chapters. The objective was to gain a better understanding of similarities and differences in the measures of consideration and initiation of structure. The original pool included items from the following: the LBDQ (Halpin & Winer, 1957), Survey of Organizations (Taylor & Bowers, 1972), interaction process analysis (Bales, 1950), the Job Descriptive Index (Smith, Kendall, & Hulin, 1969), the Orientation Inventory (Bass, 1963), scale anchors use to describe a “continuum of leadership behavior” (Tannenbaum & Schmidt, 1958), six categorical statements describing a continuum of decision-making styles (Vroom & Yetton, 1974), five bases of social power (French & Raven, 1959), and adjectives used by Fiedler (1967a) for measuring the least preferred coworker (LPC). Miller drew 73 nonduplicative items from the pool of 160 that were most specific and that were descriptive, rather than evaluative, and then collected data from 200 respondents from 10 organizations, including social agencies, industrial firms, and military organizations.
The first step in the hierarchical solution was a factor analysis stipulating a two-factor solution. Then the process was repeated with a stipulated three-factor solution, then a four-factor solution, and so on (Zavala, 1971). Miller then successively rotated all 12 principal components, using the varimax (orthogonal) rotation algorithm. At each level, interpretable solutions reflecting familiar leader-behavior factors emerged. The two-factor solution clearly paralleled consideration and initiation of structure. Other clearly identifiable factors that had been discovered in previous research emerged when an additional factor in each successive level of analysis was called for. Production, goal emphasis, and close supervision split apart as subfactors of initiation of structure in the four-factor solution. Participation emerged at level 6, information sharing at level 7, and supporting (the narrowly interpersonal interpretation of consideration) at level 8. Enforcing rules and procedures emerged as a subfactor of close supervision at level 9, and so forth. The emergence of the factors and the hierarchical linkages are shown in Figure 20.1. Here, the two-factor solution appears at level 2, the three factor solution at level 3, and so on. A subsequent higher-order factor analysis, based on an oblique solution, obtained a higher-order factor of consideration and another of initiation of structure.
It can be seen from Figure 20.1 that consideration includes behavior ordinarily regarded as concern for the welfare of subordinates, such as supportive behavior and sharing information, but it also appears linked to participative group decision making, to abdication, and to delegation.
Hemphill, Seigel, and Westie (1951) developed and Halpin (1957c) revised an ideal form of the LBDQ that asks respondents to describe how their leader should behave—not, as on the LBDQ, how they see their leader actually behaving. For example, in a study of 50 principals, J. E. Hunt (1968) found that teachers described principals as lower in actual consideration and structure than the teachers believed to be ideal. Such discrepancies between subordinates’ descriptions of what their leaders should do and what the leaders actually do are more highly related to various measures of group performance than are desired or observed leadership behavior alone. Such discrepancies are measures of dissatisfaction with the leaders’ performance and, as a consequence, are more strongly related to various group outcomes.
Stogdill, Scott, and Jaynes (1956) studied a large military research organization in which executives and their subordinates described themselves—and subordinates described their superiors—on the real and ideal forms of the LBDQ. When superiors were really high in initiation of structure, according to their subordinates, the subordinates described these superiors on the ideal forms as having less responsibility than they should, as delegating more than they should, and as devoting more time than they should to teaching. When superiors were really high in consideration on the LBDQ, according to their subordinates, the subordinates said on the ideal form that they expected the superiors to assume more responsibility than they perceived the superiors to assume, devote more time than necessary to scheduling, and devote less time to teaching and mathematical computation. When superiors were really high in initiating structure, as seen by their subordinates, the superiors perceived themselves to be devoting more time than they should to evaluation, consulting peers, and teaching and not enough time to professional consultation. When superiors were described as actually high in consideration, subordinates perceived themselves as having more responsibility than they should. The subordinates also reported that the superiors ought to devote more time to coordination, professional consultation, and writing reports but less time to preparing charts. The leaders’ initiation of structure was more highly related to subordinates’ actual work performance. The leaders’ consideration, by contrast, was more highly related to subordinates’ idealization of their own work performance.
Figure 20.1 The Hierarchical Structure of Leadership Behaviors
SOURCE: J. A. Miller (1973b).
In a similar type of study, Bledsoe, Brown, and Dalton (1980) showed that the actual behavior of school business managers, as described by 132 school superintendents, principals, and school board members, tended to differ considerably from the ideal for initiation and consideration Board members tended to describe the ideal school business managers as actually more considerate than did the school principals. Ogbuehi (1981) surveyed 270 Nigerian managers’ and administrators’ self-descriptions of their ideal behavior using the LBDQ-XII. The results were consistent with their superiors’ judgments of their effectiveness.
Vecchio and Boatwright (2002) examined the preferences of 1,137 employees in three organizations regarding leaders’ structuring and consideration. More highly educated and tenured employees preferred less structuring in their leaders; women preferred more consideration.
Fleishman’s (1989a) LOQ differs from the ideal forms of Hemphill, Seigel, and Westie (1951). It asks the leaders themselves to choose the alternative that most nearly expresses their opinion on how frequently they should do what is described by each item on the questionnaire, and what they, as a supervisor or manager, sincerely believe to be the desirable way to act. The LOQ also differs from the ideal form in that the LOQ scale for initiation of structure contains several items that Stogdill, Goode, and Day (1962) later found to measure production emphasis. Production emphasis correlates with initiating structure but is not identical with it.
Following a review, Schriesheim and Kerr (1974) concluded that the test-retest reliability of the LOQ had been adequately demonstrated over a one-to three-month period. Internal consistency reliabilities are also high (Fleishman, 1960, 1989b). In 60 studies, the median correlation between consideration and structure on the LOQ was –.06, with 57 of these correlations below .19 and only 9 above .20 found significant (Fleishman, 1989). According to a review of 20 LOQ validation studies (Fleishman, 1989b), a number of studies showed that supervisors higher in both consideration and structure were more likely to be higher on criteria of effectiveness, such as performance ratings, staff satisfaction, low stress, and less burnout of subordinates.
The internal consistency and test-retest reliability of the various scales of leader-behavior descriptions may be satisfactory, but to understand their effects, differences in content make it mandatory to distinguish whether the measures were based on the LBDQ, SBDQ, LOQ, or LBDQ-XII. This is essential in reviewing the antecedent conditions that influence the extent to which a particular behavior is exhibited and concurrent conditions are associated with such behavior.
Interpreting Concurrent Analyses. Most of the available research consists of surveys in which leadership behavior and other variables in the leader or the situation were measured concurrently. However, it seems reasonable to infer that a relatively invariant attribute such as the intelligence of the leader is an antecedent to the leader’s display of consideration or initiation as rated by their colleagues. The national origin of the leader’s organization is antecedent to the leader’s behavior; but the average leader’s behavior in the organization cannot affect its origin. Likewise, the leader’s educational level is obviously antecedent to his or her behavior, but that behavior cannot cause a change in the leader’s educational level. Similar inferences can be made about situational influences on the leader’s behavior with less confidence, because the leader often can influence the situation, just as the situation is influencing the leader. If an association is found between company policy and the behavior of first-line supervisors, it seems reasonable to infer that the policy has influenced the supervisors, but the policy may also reflect the continuing behavior of the supervisors and may be a reaction to it. If an association is found between the leader’s behavior and conflict in the work group, it is likely that the leader is a source of the conflict; but the continuing conflict is likely to be influencing the leader’s behavior as well.
Suppose that a positive association is found between a leader’s consideration and an absence of conflict within a group. The most plausible hypothesis is probably that the leader’s behavior contributes to the absence of conflict; but the harmony within the group makes it possible for the leader to be more considerate. Therefore, in examining concurrent results, the reader will have to decide what meaning to draw from the reported associations. Such criteria of effectiveness as the subordinates’ productivity, satisfaction, cohesion, and role clarity can be seen to be a consequence of the leader’s behavior, yet they may influence the leader’s behavior as well.
In a study of ROTC cadets, Fleishman (1957a) found that their attitudes toward consideration and initiation on the LOQ were not related to their intelligence or to their level of aspiration. But among school principals described by 726 teachers, Rooker (1968) found that principals with a strong need for achievement were described as high in tolerance of freedom and reconciliation of conflicting demands on the LBDQ-XII. However, Tronc and Enns (1969) found that promotion-oriented executives tended to emphasize initiation of structure over consideration to a greater degree than executives who were less highly oriented toward promotion. And Lindemuth (1969) reported that for college deans, consideration was related to their scholarship, propriety, and practicality.
Experience and sex have been correlated with LBDQ scores. For 124 managers of state rehabilitation agencies who were described by their 118 subordinates, Latta and Emener (1983) found that initiation of structure on the LBDQ increased directly with experience. Serafini and Pearson (1984) reported that at a university initiation of structure was higher only among the 208 male nonadministrative supervisors and managers. For the female leaders, consideration was higher. This is consistent with the expectation that females are likely to be more relations-oriented.
Personality, Values, and Interests. Although one might expect authoritarianism to coincide with initiation of structure, this does not appear to occur. For example, Fleishman (1957a) observed that the leader’s endorsement of authoritarian attitudes was negatively related to initiation of structure on the LOQ, but it was unrelated to consideration. Stanton (1960) also found no relation between consideration and authoritarianism. Flocco (1969), who studied 1,200 school administrators, showed that consideration and initiation of structure, as indicated by subordinates’ responses to the LBDQ, were unrelated to the administrators’ scores for dogmatism on a personality test.
Fleishman (1957b) also found that supervisors who favored consideration tended to have high scores on a personality scale of benevolence, whereas those favoring initiation of structure were more meticulous and sociable. Also, Fleishman and Peters (1962) obtained results for supervisors indicating that the trait of independence was correlated negatively with both initiation of structure and consideration, and the trait of benevolence was positively correlated with initiation of structure and consideration. Consideration was more highly related than initiation to ratings of social adjustment and charm (Marks & Jenkins, 1965). Litzinger (1965) reported that managers who favored consideration tended to value support (being treated with understanding and encouragement), whereas those who favored initiation of structure tended to place a low value on independence. Atwater and White (1985) reported that certain personal characteristics significantly correlated with demanding (SBDQ structuring) behavior by first-line supervisors; these included being inflexible, aggressive, uncooperative, harsh, strict, tense, ambitious, and unforgiving.
Newport (1962) studied 48 cadet flight leaders, each described on the LBDQ by seven flight members. Leaders who were rated high in both consideration and initiation of structure differed from those who were rated weak in desire for individual freedom of expression, resistance to social pressure, desire for power, cooperativeness, and aggressive attitudes.
In line with expectations, R. M. Anderson (1964) found, in a study of nursing supervisors, that those who preferred nursing-care activities were described as high in consideration but those who preferred coordinating activities were described as high in initiation. According to analyses by Stromberg (1967), school principals with emergent value systems were perceived by teachers as high in initiating structure, whereas those with traditional value orientations were perceived as high in consideration. Durand and Nord (1976) noted that 45 managers in a midwestern textile and plastics firm were rated by their subordinates as higher in both consideration and initiation of structure if the managers were externally, rather than internally, controlled—that is, if the managers believed that personal outcomes were due to forces outside their control rather than to their own actions.
Fleishman and Salter (1963) measured empathy in terms of supervisors’ ability to guess how their subordinates would fill out a self-description questionnaire. They found that empathy was significantly related to employees’ descriptions of their supervisors’ consideration but not their initiation of structure. L. V. Gordon (1963a) showed that personal ascendancy was positively related to initiating structure but negatively related to consideration. Neither score was related to responsibility or emotional stability, although sociability was correlated with initiating structure (but not with consideration). Rowland and Scott (1968) failed to find any relation between consideration on the LOQ and the social sensitivity of supervisors. Pierson (1984) failed to find any significant relations between the Myers-Briggs Indicator of perceptual and judgmental tendencies and consideration or initiation. Numerous other investigators5 also failed to find LBDQ and LOQ scores related to any of the personality measures they used. Situational factors, to be discussed later, may override or eliminate the effects of personality on initiation and consideration.
Cognitive Complexity. A number of studies of the influence of cognitive complexity on leadership behavior have obtained positive findings, particularly when the additional LBDQ-Form XII factors have been used. W. R. Kelley (1968) reported that school superintendents who were high in cognitive complexity were also described as high in predictive accuracy and in reconciliation of conflicting demands. Streufert, Streufert, and Castore (1968) found significant differences between emergent leaders whose scores on perceptual complexity varied in a negotiations game. Leaders who were lower in cognitive complexity scored higher on initiating structure, production emphasis, and reconciliation. Leaders who were higher in cognitive complexity scored higher on tolerance of uncertainty, retaining the leadership role, consideration, and predictive accuracy. Results obtained by Weissenberg and Gruenfeld (1966) indicated that supervisors who scored high in field independence endorsed less consideration than those who scored high in field dependence. But Erez (1979) found that the self-described consideration of 45 Israeli managers with engineering backgrounds was positively related to field independence and to social intelligence, whereas initiating structure was negatively related to these two factors.
Preferences for Taking Risks. Rim (1965) studied risky decision making by supervisors and reported that male supervisors who scored high on both consideration and initiating structure and head nurses who scored high on initiation of structure tended to make riskier decisions. Men and women who scored high on both tended to be more influential in their groups and to lead the groups toward riskier decisions. However, Trimble (1968) found that for a sample of teachers who described their principals as being higher in consideration than in initiating structure, neither of the principals’ scores was related to the principals’ perceptions of their own decision-making behavior.
Personal Satisfaction. Initiation and consideration are greater among more satisfied leaders, whose tendencies to make decisions and attempts to lead are also related. To some degree, these tendencies may be consequences rather than antecedents of initiation and consideration. Managers who are more satisfied with their circumstances tend to have higher LBDQ scores, according to Siegel (1969). Similarly, A. F. Brown (1966) reported that better-satisfied school principals were described as higher than dissatisfied principals on all subscales except tolerance of uncertainty.
Democratic versus autocratic, participative versus directive, and relations-oriented, task-oriented, transactional, and transformational leadership styles are discussed in Chapters 17, 18 and 19. It should come as no surprise that consideration and initiation are related to these other leadership styles. Miner (1973) noted that concepts used in other studies—such as providing support, an orientation toward employees, human relations skills, providing for the direct satisfaction of needs, and group-maintenance skills—are akin to consideration. Concepts similar to initiating structure include facilitation of work, production orientation, enabling the achievement of goals, differentiation of the supervisory role, and the utilization of technical skills. Miner further pointed out that with its emphasis on organizing, planning, coordinating, and controlling, initiating structure has much in common with the ideas of classical management.
Democratic and Autocratic Styles. Although factorially independent, the various scales of consideration and initiation contain the conceptually mixed bag of authoritarian and democratic leadership behaviors. Each scale contains a variety of authoritarian or democratic elements. Although empirically these elements cluster on one side or the other of authoritarian or democratic leadership, they are conceptually distinct. The industrial version, SBDQ, added strongly directive behaviors (“He rules with an iron hand”) to its factor of initiating structure (House & Filley, 1971).
Yukl and Hunt (1976) correlated Bowers and Seashore’s (1966) four leadership styles of support, the facilitation of interaction, emphasis on goals, and the facilitation of work with the LBDQ for 74 presidents of business fraternities. Support correlated .66 with consideration and .61 with initiation of structure. On the other hand, the emphasis on goals correlated .64 with consideration and .76 with initiation. Facilitation of work correlated .56 with consideration and .64 with initiation of structure. Clearly, a large general factor of leadership permeates all these measures, and leadership generally is or is not actively displayed. Karmel (1978) drew attention to the ubiquity of initiation and consideration in the study of leadership and in efforts to theorize about it. What she primarily added was the importance of the total amount of both kinds of activity by leaders, in contrast to inactivity.
This general factor becomes most apparent when the LBDQ rather than the LOQ is used. Weissenberg and Kavanagh (1972) concluded from a review that although managers think they should behave as if consideration and initiating structure are independent, in 13 of 22 industrial studies and in eight of nine military studies, a significant positive correlation was found between these two factors of leadership behavior on the LBDQ as completed by subordinates. This was especially so when LBDQ-XII was the version used in the survey (Schriesheim & Kerr, 1974). Seeman (1957) noted that a school principal’s overall leadership performance was seen to be a matter of how much consideration and initiation of structure were exhibited. When Capelle (1967) asked 50 student leaders and 50 nonleaders to fill out the LOQ, he found that leaders scored significantly higher than did nonleaders on both consideration and initiation of structure. However, G. W. Bryant (1968) did not find that appointed and sociometrically chosen leaders (college students in ROTC) differed significantly in their conceptions of the ideal leader on the LOQ. These results fit with the general contention that, conceptually, initiation of structure is readily distinguishable from consideration, just as autocratic and democratic or relations-oriented and task-oriented leadership can be conceptually discriminated. But empirically, the same leaders who are high on one factor are often high on the other as well.
Task and Relations Orientation. Initiation of structure emphasized concern with tasks (“Insists on maintaining standards,” “Sees that subordinates work to their full capacity,” “Emphasizes the meeting of deadlines”), as well as directiveness (“Makes attitudes clear,” “Decides in detail what should be done and how it should be done”). Consideration emphasized the leader’s orientation to followers (“Stresses the importance of people and their satisfaction at work,” “Sees that subordinates are rewarded for a job well done,” “Makes subordinates feel at ease when talking with them”), as well as participative decision making (“Puts subordinates’ suggestions into operation,” “Gets approval of subordinates on important matters before going ahead”). Social distance was also minimized for considerate leaders (“Treats subordinates as equals,” “Is easy to approach”). Conceptually opposite to initiation of structure is destructuring behavior (J. A. Miller, 1973a)—that is, reducing the request for consistent patterns of relations within the group. The lack of initiation of structure implies allowing conditions to continue without structure, avoiding giving directions, and avoiding being task-oriented. Conceptually opposite to consideration is leadership behavior that is exploitative, unsupportive, and uncaring (Bernardin, 1976).
Among 55 corporation presidents according to ratings by a staff member, a correlation of .55 was found between task-oriented production emphasis and initiating structure. Similarly, these presidents’ consideration correlated .49 with the relations-oriented representation of their subordinates’ interests and .41 with toleration of freedom of action for their subordinates (Stogdill, Goode, & Day, 1963a). W. K. Graham (1968) found, as predicted, that high-LPC (relations-oriented) leaders were described as higher in consideration and initiating structure than low-LPC (task-oriented) leaders. Yukl (1968) also noted that low-LPC leaders tended to be described as high in initiation and low in consideration. However, Meuwese and Fiedler (1965) reported that leaders who were high and low on LPC tended to differ significantly only on specific items of the LBDQ, not in the total scores for consideration and initiating structure. Yukl (1971) and Kavanagh (1975) concluded that task-oriented behavior is implicit in initiating structure, but subordinates can still influence their superior’s decisions. Misumi’s Production and Maintenance (PM) style of leadership uses maintenance items that correlate highly with consideration. Production has two components: pressure and planning. Pressure correlates highly with initiation of structure when measured by the original SBDQ, which contained items such as “Prods for production” but is less correlated with initiation of structure as measured by the LPDQ without such autocratic items (Peterson, Smith, & Tayeb, 1993).
Power, Authority, and Responsibility of the Leader. Martin and Hunt (1980) obtained LBDQ evaluations by 407 professionals and quasi-professionals of their first-line supervisors in the construction and design units of 10 state highway department districts. Related data on morale were also collected. The expert power of the supervisors correlated .44 and .41 with their initiating structure and .48 and .51 with their consideration; but the other sources of power—referent, reward, coercive, and legitimate—correlated close to zero with the leadership measures. Foote (1970) found that members of the managerial staffs of television stations who tended to describe themselves on the RAD scales6 as high in responsibility and authority also tended to be described on the LBDQ-XII as high in tolerance of freedom. Those who delegated most freely were described as high in production emphasis and low in representation orientation toward superiors.
Attempts to lead, manifest in one’s emergence as a leader in a leaderless group discussion, were negatively related to consideration and positively related to initiation of structure (Fleishman, 1957a).
Transactional and Transformational Leadership. Cheng (1994) suggested that transformational school-teachers could be innovative and could demand clear rules in managing the classroom (structuring) as well as provide support to the students (consideration). Seltzer and Bass (1987) found that for 294 MBAs with full-time jobs who described their immediate supervisors’ Initiation of Structure on the LBDQ-XII correlated .53, .55, and .59, respectively, with charisma, individualized consideration, and intellectual stimulation on the Multifactor Leadership Questionnaire (MLQ); and .48 and .06, respectively, with the transactional measures of contingent reward and management by exception on the MLQ. However, consideration on the LBDQ-XII correlated .78, .78, and .65, respectively, with the same MLQ transformational leadership measures and .64 and –.23, respectively, with the same MLQ transactional leadership measures. Evidently, active leadership is common to initiation, consideration, and transformational and transactional leadership behavior, and there are particularly strong associations between transformational leadership and consideration. Miliffe, Piccolo, and Judge (2005) reported correlations of .46 and .27, respectively, of transformational leadership with consideration and initiation. Furthermore, adding the transformational scales to initiation and consideration increased substantially the prediction of outcomes of the rated effectiveness of leaders and satisfaction with leadership. Similar results were reported for 138 subordinates and 55 managers (Seltzer & Bass, 1990).
Peterson, Phillips, and Duran (1989) found that the MLQ scale of charismatic leadership correlated higher with measures of consideration than measures of initiation in a study of 264 retail chain-store employees describing their supervisors. Thus charismatic leadership correlated .48 with maintenance orientation and .74 with support but only .16 with pressure for production and .22 with assigning work.
Greenleaf (1977, 1991), an AT&T executive, conceived the idea of the servant leader from Herman Hesse’s Journey to the East. Leo, the servant to a band of men on a mythical journey, does their menial chores but also sustains them with his spirit and song. Leo disappears. The group falls apart and the journey is abandoned. It turns out later that Leo is the titular head of the order that sponsored the journey. For Greenleaf, as for Hesse, the leader is the servant, first. The leader ensures that the highest priorities of followers are met. He first listens and questions before suggesting initiatives. Instead of coercion and manipulation, the servant leader depends on awareness, empathy, and foresight. To illustrate:
John Woolman visited Quaker communities for 30 years asking Quaker slaveholders about what … slave-owning did to them morally. … What were they passing on to their children? Non-judgmental persuasion followed. The Quakers became the first religious group in America to denounce and forbid slavery among its members. (Greenleaf, 1991, p. 21)
Servant leaders encourage skill and moral development in followers. They are sensitive to the needs of organizational stakeholders and hold themselves accountable for their actions (Graham, 1991). Servant leadership can be a model for higher education. Faculty members need to curtail and redirect their ego and image. Power needs to be shared with students. They become collaborators in an ethic of collective study. “The faculty member who has never learned from his students is a failure” (Buchen, 1998, p. 130). In Block’s (1993) replacement of leadership with stewardship, service is chosen over self-interest. As noted by Heuerman (2002), caring about others, seen in servant leaders, underlies considerate leadership behavior. Followers are regarded as real people, not machines or expense categories. For leaders who care, service is an obligation, not a burden. At the extreme of the leader as servant, leaders sacrifice themselves for the perceived good of the group. In experiments with 357 students and 157 industrial personnel, Choi and Mai-Dalton (1998) found that self-sacrificing leaders were seen as more charismatic and legitimate, and generated intentions to reciprocate. Self-sacrificing leaders were also judged as more competent by the students. Block (1993) argued that self-interest, dependency, and control of others should be replaced by organizational stewards with service, responsibility, and accountability.
The self-sacrificing servant leader has to be a good steward. Abraham Feuerman was praised for continuing to pay full wages to his employees at Malden Mills from reserves and insurance during the period it was out of operation after it burned down. He might have been expected to use the money to rebuild overseas to make use of cheaper labor, as other textile companies had done. Instead, he continued to pay his employees, who were able to return with high morale to a beautiful new plant, for which he became extensively indebted. Unfortunately, his competitors used the time to catch and pass Malden Mills and gain some of its customers. Product leadership was lost. Bankruptcy ensued. In attending to the needs of his people, the servant CEO had not fully taken care of his organization’s financial future.
The organization, the immediate group that is led, and the task requirements affect the extent to which a leader initiates structure, is considerate, or both. For instance, when faced with a complex task and need for planning, a leader is likely to be more structuring in a group of low diversity. Structuring and the leader’s planning skills contribute to the quality and originality of the plans, according to a study of 195 participants working in 55 groups (Daniels, Leritz, & Mumford, 2003).
Organizational Policies. A clear example of the impact of the organizational context on the behavior of individual leaders within it was provided by Stanton (1960), who described two medium-size firms. In one company, which was interested only in profits, authoritarian policies were dominant, and subordinates had to understand what was expected of them. The personal qualities of leadership were emphasized, and all information in the company was restricted to the managers except when the information clearly applied to an employee’s job. The second firm, which had democratic policies, stressed participation as a matter of policy and was concerned about the employees’ well-being as well as about profits. This firm made a maximum effort to inform the employees about company matters. Supervisors in the firm with democratic policies favored more consideration, whereas supervisors in the firm with authoritarian policies favored more initiation.
The importance of the higher authority represented by the organization and its policies also can be inferred indirectly from results obtained in a progressive petrochemical refinery and in a national food-processing firm, where the extent to which supervisors felt they should be considerate was positively correlated with how highly they were rated by their superiors (Bass, 1956, 1958). Yet in other companies no such correlation was found (Rambo, 1958). Supervisors’ perceptions of their superiors’ and subordinates’ expectations affect their leadership behavior. That is, supervisors will be more supportive or more demanding, depending on what they perceive their superiors and subordinates expect of them (Atwater, 1988).
Organizational Size. Vienneau (1982) examined the LBDQ scores of 33 presidents of amateur sports organizations, obtained from the responses of 85 members of their executive committees. Although the presidents’ sex or language was of no consequence, both consideration and initiation were higher in the larger of the amateur organizations. The presidents agreed on what was required of their ideal leader on the Ideal Leader Behavior Questionnaire regardless of other organizational differences.
Functional Differences. D. R. Day (1961) found that upper-level marketing executives were described on the LBDQ-XII as high in tolerance of freedom and low in structuring, but upper-level engineering executives were described as low in tolerance of freedom and high in initiation. In the same firm, manufacturing executives were rated high and personnel executives were rated low in tolerance of uncertainty.
Military versus Civilian Supervisors. Holloman (1967) studied military and civilian personnel in a large U.S. Air Force organization and found that superiors did not perceive military and civilian supervisors to be different in observed consideration or initiation of structure, although they expected military supervisors to rank higher than civilian supervisors in initiation of structure and lower in showing consideration. Unexpectedly, Holloman found that subordinates—both military and civilian—perceived the military supervisors to be higher in consideration as well as in initiating structure than they did the civilian supervisors. Thus in comparison with civilian supervisors, military supervisors were seen to display more leadership by both their civilian and their military subordinates.
Halpin (1955b) administered the ideal form of the LBDQ to educational administrators and aircraft commanders. Subordinates described their leaders on the real form of the LBDQ. The educators exhibited more consideration and less initiation of structure than did the aircraft commanders, both in observed behavior and in ideal behavior. But in both samples, the leaders’ ideals of how they should behave were not highly related to their actual behavior as described by their subordinates.
Attributes of Subordinates. Atwater and White (1985) and Atwater (1988) found that supportive (considerate) behavior by supervisors correlated highly with the subordinates’ loyalty and trust. Kerr, Schriesheim, Murphy, and Stogdill (1974) concluded, from a review of LBDQ studies, that if the subordinates’ interest in the task and need for information are high, less consideration by the leader is necessary and more initiation of structure is acceptable to them. Consistent with this conclusion, Hsu and Newton (1974) showed that supervisors of unskilled employees were able to initiate more structure than were supervisors of skilled employees in the same manufacturing plant. In a large sample survey of employees, Vecchio (2000) found that less mature employees were more inclined to favor structuring by leaders.
Chacko (1990) noted in data from 144 department heads in institutions of higher education that subordinates were more likely to be assertive and appeal to higher authority when their supervisors were low in structuring and in consideration. Gemmill and Heisler (1972) and Lester and Genz (1978) analyzed the impact of subordinates’ locus of control—internal or external—on their perceptions of their supervisors’ leadership and their satisfaction with it. Internally controlled subordinates tended to see significantly more consideration and initiation in their supervisors’ behavior (Evans, 1974). Although internally controlled subordinates in a textile and plastics firm tended to see their supervisors as initiating more structure, they also felt that their supervisors were less considerate (Duran & Nord, 1976). But Blank, Weitzel, and Green (1987) found correlations close to zero between the psychological and job maturity of 353 advisers of residence halls and the initiation and consideration, respectively, of their 27 residence-hall directors.
Except for a few cross-lagged analyses and experiments, most of the results reported here come from concurrent surveys of leadership behavior and criteria such as subordinates’ satisfaction and productivity. For instance, Brooks (1955) found that all the items measuring consideration and initiation of structure differentiated managers rated as excellent from those rated as average or below average in effectiveness. Although one tends to infer that productivity and satisfaction are a consequence of leadership behavior, the effective outcomes modify the leader’s behavior to some extent as well. Greene and Schriesheim’s (1977) longitudinal study suggested that more consideration early on by a leader can contribute to good group relations, which in turn may later result in higher group productivity. In any event, Fisher and Edwards (1988), in a series of meta-analyses, established that subordinates were more satisfied with work and supervision if their supervisors were high in consideration.
Using an early version of the LBDQ scales, Hemphill, Seigel, and Westie (1951) found that leaders’ organizing behavior (initiating structure) and membership behavior (consideration) were both significantly related to the cohesiveness of the group. Likewise, Christner and Hemp-hill (1955) noted that subordinates’ ratings of the leaders’ consideration and initiation were positively related to ratings of the effectiveness of their units, but leaders’ self-descriptions of consideration and initiation were not.
A sample of 256 MBA students who were working full-time in many different organizations described the initiation and consideration of their immediate supervisors at work. They also completed a “burnout” questionnaire. Although their leaders’ initiation correlated only –.15 with the respondents’ feeling of being burned out, the leaders’ consideration correlated –.55 with this feeling. Thus considerate supervision appears to reduce substantially the sense of burnout in subordinates (Seltzer & Numerof, 1988). Considerate supervision also promotes creativity, according to Oldham and Cummings (1996). They found that supportive and noncontrolling supervision of 171 employees in two manufacturing plants correlated with subordinates’ pattern disclosures, contributions to suggestion programs, and rated creativity. But Williams (2001) reported that the amount of structure supervisors initiated reduced the divergent thinking of their subordinates.
In an extensive analysis of 27 organizations involving more than 1,300 supervisors and 3,700 employees, Stog-dill (1965a) ascertained that supervisors’ consideration was related to the employees’ satisfaction with the companies and to measures of the cohesiveness of the groups and the organizations. But as with authoritarian and democratic leadership, neither the supervisors’ consideration nor their initiation of structure was consistently related to group productivity. Organizational differences had to be considered. The contribution to effectiveness of initiation and consideration appears quite variable and hence requires further examination of the context in which the data are collected.
Leader Behavior. Fleishman (1989a) reviewed more than 20 validity studies of the SBDQ. Fleishman, Harris, and Burt (1955) found that production foremen rated higher in performance by their managers were higher in initiation of structure and lower in consideration. But absenteeism and turnover were greater in work groups that had foremen with this pattern. Many other studies with the SBDQ have confirmed this strong relation between leader consideration and worker job satisfaction in industry (Badin, 1974; Fleishman & Simmons, 1970; Skinner, 1969), hospitals (Szabo, 1981; Oaklander & Fleishman, 1964), educational settings (Petty & Lee, 1975), and government organizations (Miles & Petty, 1977).
The relationship between initiation of structure, as measured by the SBDQ, and performance criteria tended to vary with situations. Thus both initiation of structure and consideration were positively related to proficiency ratings in nonproduction departments (Fleishman & Harris, 1955). Using the SBDQ, Hammer and Dachler (1973) showed that the leader’s consideration was positively related to the subordinates’ perceptions that their job performance was instrumental in obtaining the desired outcomes. However, the leader’s initiation of structure was negatively related to such perceptions. Likewise, Gekoski (1952) found that supervisors’ initiation of structure, but not their consideration, was related positively to group-productivity measures in a clerical situation.
Lawshe and Nagle (1953) obtained a high positive correlation, for a small sample of work groups, between group productivity and employees’ perceptions of how considerate their supervisor was. In a study of the leadership of foremen, Besco and Lawshe (1959) found that superiors’ descriptions of a foreman’s consideration and initiation of structure were both related positively to ratings of the effectiveness of the foreman’s unit. However, subordinates’ descriptions of foremen’s consideration, but not initiation of structure, were positively related to such effectiveness.
In later studies, the LBDQ or LBDQ-XII was more likely to be used. Trieb and Marion (1969) studied two chains of retail grocery stores. They found that the supervisors’ consideration, as described by the workers, was positively related to productivity, cohesiveness, and satisfaction in both chains. The supervisors’ initiation of structure was related positively to the productivity of subordinates and to cohesiveness in one chain, but not in the other. In a study of two companies, House and Filley (1971) found that the supervisors’ consideration in both companies related significantly to the subordinates’ satisfaction with the company and with their jobs as well as to their freedom of action. In both companies, initiating structure was also related significantly to the subordinates’ satisfaction with the company and their jobs, along with favorable family attitudes toward the company and their jobs. Fleishman and Simmons (1970) showed that the effectiveness of Israeli supervisors was positively related to their initiation of structure and consideration.
M. G. Evans (1968) reported that supervisors’ consideration and initiation of structure were positively related to the importance of the goal to workers and to their job satisfaction. Under high supervisory consideration, a strong positive relationship existed between the supervisor’s initiation of structure and the group’s performance. In a later study, M. G. Evans (1970a) found that the supervisors’ consideration and initiation were related to the workers’ perception of opportunities to satisfy their need for security but not to their actual satisfaction with their job security. Weiss (1977) demonstrated that subordinates tended to be more likely to share values with their supervisors if the supervisors displayed considerate behavior toward them. Marks and Jenkins (1965) reported that initiation of structure was more highly related than was consideration to global ratings of effectiveness. However, supervisors’ initiation of structure and consideration were unrelated to the satisfaction of subordinates’ needs for social esteem, autonomy, and self-actualization.
In a study of insurance sales supervisors, W. K. Graham (1970/1973) found that supervisory consideration was positively associated with group performance. In a similar setting in seven retail discount department stores, Hodge (1976) obtained results for 21 second-level managers as reported on the LBDQ by 188 first-line managers. Satisfaction of subordinates’ needs was positively associated with the higher-level managers’ initiation of structure. But there was an unexpected negative correlation with consideration. These results were similar to what Patchen (1962) found for the supervision of manual workers, but contrary to what was reported by Fleishman and Harris (1962) for similar types of workers.
The source of the criteria makes some difference. Kofman (1966) reviewed research in which industrial supervisors’ scores for consideration and initiating structure were related to various criteria of the supervisors’ effectiveness and the work groups’ performance. Generally, he found that the peer ratings of the groups’ performance were unrelated to the peer ratings of the supervisors’ consideration and initiation of structure. However, evaluations of the supervisors’ effectiveness by superiors and subordinates, as well as evaluations based on objective criteria, tended to relate positively to the supervisors’ initiation and consideration as described by subordinates.
A number of industrial studies were completed using the additional scales of LBDQ-XII. D. R. Day (1961) obtained 165 ratings of executives in an aircraft manufacturing firm. The effectiveness of leaders correlated with a general factor, as well as with the predictive accuracy of LBDQ-XII, persuasiveness, role enactment, and the reconciliation of conflicting demands. According to R. E. Hastings (1964), leaders who were rated high in initiating structure and production emphasis supervised research teams that were rated high in the volume of work they completed. If the leaders were rated high in orientation to superiors, their groups were rated low in harmony. Leaders who were rated high in representation and role retention tended to supervise teams that were rated high in enthusiastic effort. Leaders who were high in persuasiveness supervised teams whose quality of work was rated as high. Even after the effects of many other morale variables were removed, freedom and consideration by supervisors on LBDQ-XII were important contributors to the amount of innovative behavior displayed by 309 federal R & D aerospace scientists and engineers (Dalessio & Davis, 1986).
M. Beer (1964) used the LBDQ-XII to test McGregor’s (1960) hypothesis that employees become motivated and are enabled to satisfy their higher-order needs (for autonomy, esteem, and self-actualization) only when supervisors allow them freedom from organizational structure and pressure. He found support in that the employees’ satisfaction of the need for autonomy, esteem, and self-actualization was positively related to the supervisors’ consideration and tolerance of freedom. However, contrary to his hypothesis, the leaders’ considerate behaviors that resulted in the satisfaction of higher-order needs were not the ones that led to employees’ strong striving for the needs. Rather, the leaders’ initiation of structure was the leaders’ behavior associated with such striving.
Leaders’ Opinions and Attitudes. As with the effects of leaders’ behavior on effective and satisfying outcome, findings about the effects of leaders’ attitudes were also mixed. Bass (1956) found that the effectiveness ratings by superiors of 53 supervisors were significantly related to their opinion scores on consideration but not on initiation expressed two years previously on the Leadership Opinion Questionnaire (LOQ). In a replication, a significant correlation of .32 was found between sales supervisors’ attitudes toward consideration and effectiveness ratings by superiors three years later (Bass, 1958). But Bass (1957a) also reported that neither attitudes toward consideration nor attitudes toward initiation of structure were related to peer ratings of sales supervisors on criteria such as popularity, problem-solving ability, and value to the company. Fleishman and Peters (1962) found no relation between LOQ measures of supervisors’ attitudes toward leadership and their rated effectiveness as supervisors. Rowland and Scott (1968) also noted that LOQ measures of supervisory consideration were unrelated to employees’ satisfaction. According to T. C. Parker (1963), 1,760 employees of a wholesale pharmaceutical company in 80 decentralized warehouses were more satisfied with their supervision, their recognition, and their job security when their supervisors felt that consideration and initiation were important, as measured by the LOQ. The correlations of subordinates’ satisfaction and supervisors’ attitudes were .51 with supervisory consideration and .22 with supervisory initiation. Although there was no relation of the supervisors’ attitudes to such objective measures of group performance as productivity and errors in filling orders, the supervisor’s favoring of initiating structure correlated .23 with pricing errors of the unit being supervised. Spitzer and McNamara (1964) also reported that managers’ attitudes toward consideration and initiation were not related to salary as a criterion of the managers’ success; nor were such attitudes related to superiors’ ratings of the managers’ success. However, Weissenberg and Gruenfeld (1966) found that supervisors who had favorable attitudes toward consideration and initiation were also more favorably inclined toward the personal development of their subordinates. Generally, satisfactory supervisors reveal this “hi-hi” pattern (Fleishman, 1989b), and it has been found to be related to low stress and low burnout among subordinates (Duxbury, Armstrong, Drew, & Henly, 1984).
In the first extensive use of the LBDQ with U.S. Air Force personnel, Christner and Hemphill (1955) found that changes in the attitudes of crew members toward each other over time were related to the leadership behavior of the crew commander. When crew members rated their commander as high in consideration, they increased their ratings of each other’s friendliness, mutual confidence, conversation on duty, and willingness to engage in combat. Crews that described their commander as high in initiation of structure increased their ratings of each other’s friendship and confidence. But Halpin (1954) found that superiors tended to evaluate positively those aircrew commanders they rated high in initiating structure and to evaluate negatively those rated high in consideration. In training, the satisfaction of crew members was positively related to rated commanders’ consideration (r = .48) and negatively related to commanders’ initiating structure (r = –.17). In combat, however, commanders’ consideration (r = .64) and initiating structure (r = .35) were positively related to the same crew members’ satisfaction. In a later study, Halpin (1957a) found that superiors’ ratings of the commanders’ effectiveness in combat were unrelated to crew members’ descriptions of their commander’s consideration. However, superiors’ ratings, as before, were positively and significantly related to the commander’s initiation of structure. Significantly, crew members’ ratings of their commander on confidence and proficiency, friendship and cooperation, and morale and satisfaction were positively related to both consideration and the initiation of structure by their commander.
Fleishman (1957a) found that the consideration and initiation scores of ROTC leaders were positively and significantly related to peers’ ratings of the leaders’ value to their groups, but superiors’ ratings were not related to either pattern of behavior. In a study of trainee leaders in the military, Hood (1963) ascertained that the trainees reported more affiliation and less communication when their superiors were higher in initiating structure and consideration. Enlisted personnel attained higher scores on a pencil-paper test of military leadership when their leaders structured the situation and pushed for production. However, such higher attainment by the enlisted personnel was not related to their leaders’ consideration.
Hooper (1968/1969) obtained results with the LOQ indicating that the attitudes of U.S. Air Force cadets toward consideration and initiation were not significantly related to their effectiveness ratings. However, the two factors differentiated significantly between those scoring high and those scoring low on a composite leadership criterion.
Group Effects. C. H. Rush (1957) reported the effects of leadership behavior on other dimensions of the performance of 212 aircrews. Crew members described the leadership behavior of the crew leaders and conditions in the crews on Hemphill and Westie’s (1950) Group Dimension Descriptions. The leaders’ consideration was associated with more intimacy and harmony and less control and stratification in the crews. The leaders’ high rating on the initiation of structure was related to greater harmony and procedural clarity and to less stratification in the crews.
In examining the satisfaction and the initiation and consideration of their leaders, according to 672 U.S. Army National Guardsmen in the study by Katerberg and Hom (1981), it was necessary to account statistically for both the group effects and the individual dyadic effects of LBDQ ratings of the first sergeants and the company commanders on the guardsmen’s satisfaction with the sergeants and commanders. First sergeants and commanders who earned high mean LBDQ scores on initiation and consideration from their collective unit of subordinates also had units with higher mean satisfaction. But beyond this, within the groups, individual subordinates were more satisfied if they saw their leader exhibiting more initiation and consideration. The adjusted correlations between satisfaction and the initiation and consideration scores on the LBDQ were .37 and .47 for the group-by-group analyses for the first sergeants and commanders, respectively. The individual-within-group analyses added substantially to the correlations between leadership and satisfaction. Although there were problems with Katerberg and Hom’s failure to separate the group and dyadic effects completely (Dansereau, Alutto, & Yammarino, 1984), the strong associations with satisfaction found for both initiation and consideration in the military were supported by similar LBDQ evidence obtained for 30,000 U.S. Army personnel in 63 national and overseas installations (Marsh & Atherton, 1981–1982).
Various investigations of the leadership behavior of college administrators, school administrators, school principals, and classroom teachers have been conducted. Hemphill (1955) used the LBDQ to study the leadership of heads of academic departments in a university. The department head’s reputation for administrative competence correlated .36 with consideration and .48 with initiation of structure. But Lindemuth (1969) failed to establish any relationship between a college dean’s initiation of structure and various measures of organizational climate. Superintendents who were rated as effective leaders by both their staff and school board members were described as high in both consideration and initiation of structure (Halpin, 1956a). H. J. Bowman (1964) asked school principals to describe the leadership behavior of higher-level school executives and themselves. Principals perceived themselves as exercising high degrees of responsibility and authority and as delegating extensively when they described their own superiors as being high in consideration but not in initiation. Among the school administrators studied by Flocco (1969), those described by subordinates as higher in consideration and initiation of structure were rated more effective. Those administrators who described themselves as being higher in consideration and initiation than their staff subordinates described them were rated ineffective.
School Principals. A. F. Brown (1967) and Greenfield (1968) concluded from reviews of Canadian studies that the performance of pupils was associated with the principals’ LBDQ scores. Keeler and Andrews (1963) studied the relation of principals’ leadership to the performance of pupils and the cohesiveness of staffs in Canadian public schools. Both consideration and initiation of structure by principals, as described by teachers, were significantly and positively related to the pupils’ examination scores on a province-wide examination. Initiation of structure by the principals was positively related to the cohesiveness of their staffs, but consideration was not. Nevertheless, A. F. Brown (1966) reported that effective principals generally scored higher on the LBDQ-XII scales. Seeman (1957) found performance evaluations of the school principals’ leadership positively related to consideration, initiation of structure, communication, and willingness to change, and negatively related to domination and social distance. According to Fast (1964), consideration and initiation of structure by principals, as described by teachers, were positively related to the teachers’ satisfaction, although expected behavior was not. Stromberg (1967) obtained a significant relation between teachers’ morale and the attitudes of their principals toward consideration and initiation of structure.
C. C. Wall (1970) studied four effective and four ineffective principals in terms of their dialogue, decision making, and action. Effective principals were described by their teachers as higher than ineffective principals in consideration and tolerance of freedom on the LBDQ-XII. Ineffective principals were scored high in production emphasis. Teachers in seven of the eight schools studied believed that the principals ought to initiate more than they were perceived to do. Teachers in the ineffective schools believed that the principals should exhibit more persuasion and demand more reconciliation and more integration of the group than they were perceived to do. Mansour (1969) found that these discrepancies between the expected and actual behavior of principals were negatively relative to teachers’ job satisfaction and participation. Fast (1964) also obtained results indicating that the greater the discrepancy between the teachers’ expectations and observations of principals’ behavior, the lower the teachers’ satisfaction.
Among different schools, Punch (1967) found that the principals’ initiation of structure was positively related and the principals’ consideration was negatively related to a measure of bureaucracy in the school. Mathews (1963) reported that principals’ initiation of structure and consideration were significantly related to Hemphill and Westie’s (1950) measures of their staff’s stratification, control, homogeneity, cohesiveness, hedonic tone, and participation.
Hills (1963) obtained descriptions by 872 teachers of 53 principals. Both consideration and initiating structure were highly correlated with two representative functions of the principals: (1) representing the teachers’ interests to higher levels of the organization and (2) representing the teachers’ interests to pupils and their parents. Hills concluded that consideration and initiation of structure were not solely concerned with internal leadership but were reflected in how the principals, as leaders, dealt with outsiders and higher levels of authority.
However, Rasmussen (1976) failed to establish any significant relationship between the success of 25 elementary schools, the satisfaction of teachers, and the behavior of principals described on the LBDQ. Bailey (1966) studied four principals who were described by their superintendents and four teachers as higher in consideration than in initiating structure and four other principals who were described as being higher in initiating structure than in consideration. Each principal and four teachers played a decision-making game. Although the principals’ consideration was found to be significantly related to the teachers’ satisfaction with the decision and support of it, neither the principals’ consideration scores nor their initiating structure scores were significantly related to the ability of a group to arrive at a decision or to perceptions that teachers had helped make the decisions.
Classroom Teachers. Using the appraisal of the performance of their classroom teachers by 2,084 business students, Baba and Ace (1989) extracted stable initiation factors as well as factors of evaluation and effort. Class size and course level made no difference. In another large-scale Canadian study, Greenfield and Andrews (1961) obtained results indicating that consideration and initiation of structure by classroom teachers were positively and significantly related to the scores of their pupils on achievement tests. Cheng (1994) found that teachers high in initiating structure and consideration were more influential as a consequence of their greater expertise and attractiveness. Their classrooms had a more positive social climate.
In a study of nurses and their supervisors, Nealey and Blood (1968) found that the satisfaction of subordinates was related to the consideration scores of both their first-and second-level supervisors. The supervisors’ initiation of structure contributed to the subordinates’ job satisfaction at the first, but not at the second, level of supervision. Oaklander and Fleishman (1964) observed that when hospital administrators endorsed both high consideration and high initiation of structure on the LOQ, stress was lower in the units they supervised.
A path analysis by Sheridan and Vredenburgh (1979) for the descriptions by 372 nurses, practical nurses, and nursing aides of the behavior of their head nurses disclosed a positive effect of the head nurses’ initiation of structure on the subordinates’ group relations. But these group relations did not affect the subordinates’ performance or turnover rates. Good group relations also reduced the subordinates’ felt job tension. However, such job tension had only a slight positive association with job performance or turnover rates. Yet the head nurses’ consideration had a direct positive effect, as well as an indirect effect, on the subordinates’ performance. It also reduced the subordinates’ felt job tension. This result fits with Weed, Mitchell, and Moffitt’s (1976) laboratory finding that the leader’s consideration makes for a pleasant working situation, although it may not necessarily contribute to the group’s productivity.
The critical importance of the leader’s initiation and consideration to the performance of subordinates in health organizations was confirmed by Dagirmanjian (1981), Blaihed (1982), and Denton (1976). On the basis of a data analysis, Dagirmanjian (1981) concluded that the supervisors’ initiation of structure and consideration were the central links between the organizational structure in mental health services and the staffs. Blaihed (1982) obtained the LBDQ-XII scores of chief executive officers (CEOs) and staffs of hospitals in Los Angeles and related the results to criteria of hospital performance furnished by the California Health Facilities Commission. Initiation of structure by the CEOs and staffs contributed to the efficiency of the hospitals and the quality of care they provided, as did a number of the other scale scores of the LBDQ. Similarly, for 80 professional mental health workers and their directors, Denton (1976) found significant, direct relationships between supervisors’ consideration and initiation of structure, on the one hand, and the workers’ job satisfaction and satisfactory relations with clients on the other. Conversely, Duxbury, Armstrong, Drew, and Henly (1984) found that head nurses with low consideration and high initiation scores on the LOQ had the lowest staff satisfaction and highest staff burnout.
Bernardin (1976) found that for 501 police officers in a metropolitan department, consideration by supervisors was positively and linearly related to the police officers’ satisfaction but not to their performance or absenteeism. To account for the results more adequately, additional descriptive data about the supervisors were required, such as the supervisors’ specific reward orientation and punitiveness.
Stogdill (1965a) studied 10 regional organizations in a department of a state government. He noted that throughout the 10 organizations, executives who described their superiors as high in representation on the LBDQ-XII tended to manage groups that were rated high in support of the organization, and their subordinates tended to be satisfied with their pay. Tolerance of uncertainty by their superiors was related to harmony in the groups. Superiors who were rated high in initiating structure had subordinates who were satisfied with the organization. When the state employees described their first-line supervisors, Stogdill found that those first-line supervisors who were rated high in initiation of structure, consideration, and “influence upstairs” tended to have subordinates who were satisfied with the organization and groups that were rated strong in drive. Employees who described their supervisors as tolerant of freedom expressed satisfaction with their own freedom on the job. Klepinger (1980) collected LBDQ and outcome data from a stratified sample of 35 executive directors of social service departments who were rated by their 227 employees. Directors who scored high in initiation and consideration were also seen as highly effective managers, but their leadership behavior did not influence their employees’ job satisfaction.
Hood (1963) found that business trainees reported more affiliation and less communication when their leaders were rated high both in initiating structure and in consideration. Furthermore, the trainees attained higher scores on a paper-and-pencil test of performance when their leaders structured the situation and pushed for production. But such increased attainment was not related to the leaders’ consideration. Cunningham (1964) observed that the most effective agricultural agents and 4-H club agents were above the median in both consideration and initiation of structure on the LBDQ. Similarly, Osborn and Hunt (1975a, 1975b) obtained data indicating that most aspects of the satisfaction of members in 60 chapters of a business fraternity were positively associated with their presidents’ initiation and consideration. But Yukl and Hunt (1976) reported that in 74 chapters the correlations between LBDQ assessments of the chapter presidents’ initiation and consideration and the chapters’ efficiency in fulfilling specified requirements were only .12 and .10, respectively.
Fisher and Edwards (1988) completed a meta-analysis of studies in the computer abstracts of Psychological Abstracts. These studies presumably were published mainly after 1968. As can be seen in Table 20.1, after the mean findings were corrected for sample size, restriction in range, and unreliability, the adjusted mean correlation of leader consideration on the LBDQ with job performance ranged from .27 to .45. The adjusted correlations of employee job performance with LBDQ initiation of structure were similar, but this was not true of the SBDQ measure of initiation with its punitive elements.
Except for the SBDQ measure of leaders’ initiation, considerably higher adjusted correlations were obtained, as expected, between leaders’ initiation, consideration, and outcomes in job satisfaction, and satisfaction with supervision, as shown in Tables 20.2 and 20.3.
Table 20.1 Mean Correlations of Leader Consideration and the Initiation of Structure with Job Performance
| Mean Correlation | ||||
| Leadership Measure | Number of Respondents | Number of Correlations | Unadjusted | Adjusted |
| Consideration | ||||
| LBDQ | 1,486 | 19 | .19 | .45 |
| SBDQ | 1,953 | 21 | .19 | .46 |
| LBDQ-XII | 1,424 | 11 | .13 | .27 |
| Initiation of Structure | ||||
| LBDQ | 1,486 | 19 | .20 | .47 |
| SBDQ | 1,953 | 21 | −.02 | −.06 |
| LBDQ-XII | 1,424 | 11 | .09 | .22 |
SOURCE: Adapted from Fisher and Edwards (1988), p. 202.
The patterns of results lent support to Larson, Hunt, and Osborn (1976), who concluded after examining 14 samples of first-line supervisors involving 2,474 respondents that the multiple regression additive or interactive combining of initiation and consideration was unwarranted, since it did not add sufficiently to the prediction of outcomes beyond what each measure could do alone. Part of this lack of augmentation of one measure by the other was due to the mean intercorrelation of .52 between them found for 10 samples on the LBDQ-XII by Schriesheim, House, and Kerr (1976). For four samples, Schriesheim (1982) obtained results indicating that consideration alone accounted for most of the effects on the satisfaction of subordinates. For samples of 230 hourly employees, 178 college seniors and graduate students with employment experience, 96 middle managers, and 258 clerks and middle managers who described their supervisors’ leadership behavior by means of the LBDQ and the SBDQ, Schreisheim found that the supervisors’ consideration alone correlated between .62 and .77 with satisfaction with the respondents’ jobs and with their supervisors. The results for the supervisors’ initiation of structure, in multiple regression analysis, added only from 2% to 4% to the consideration effects in accounting for satisfaction. However, while Nystrom (1978) reached the same conclusion about consideration and the satisfaction of subordinates’ needs, with a sample of 100 junior and senior managers, he noted that managers who were high in initiating structure and in consideration had relatively lower salary levels and progressed more slowly in their careers. Low initiation of structure, rather than high consideration, contributed the most to higher salaries and career advancement. Nystrom’s results may say much about the culture and policies in the firm from which the 100 managers were drawn. Contingent factors need to be considered.
Table 20.2 Mean Correlations of Leader Consideration and Initiation of Structure with Overall Job Satisfaction
| Mean Correlation | ||||
| Leadership Measure | Number of Respondents | Number of Correlations | Unadjusted | Adjusted |
| Consideration | ||||
| LBDQ | 2,517 | 21 | .34 | .65 |
| SDBQ | 1,134 | 8 | .47 | .83 |
| LBDQ-XII | 4,347 | 25 | .38 | .70 |
| Initiation of Structure | ||||
| LBDQ | 2,517 | 21 | .26 | .51 |
| SBDQ | 1,134 | 8 | −.02 | −.04 |
| LBDQ-XII | 4,347 | 25 | .23 | .46 |
SOURCE: Adapted from Fisher and Edwards (1988), p. 202.
Table 20.3 Mean Correlations of Leaders Consideration and Initiation of Structure with Satisfaction with Supervision
| Mean Correlation | ||||
| Leadership Measure | Number of Respondents | Number of Correlations | Unadjusted | Adjusted |
| Consideration | ||||
| LBDQ | 632 | 7 | .63 | .99 |
| SBDQ | 1,048 | 10 | .79 | .99 |
| LBDQ-XII | 3,455 | 19 | .57 | .95 |
| Initiation of Structure | ||||
| LBDQ | 632 | 7 | .29 | .57 |
| SBDQ | 1,048 | 10 | −.15 | −.30 |
| LBDQ-XII | 3,455 | 19 | .39 | .73 |
SOURCE: Adapted from Fisher and Edwards (1988), p. 203.
Judge, Piccolo, and Remus (undated) conducted three meta-analyses involving over 20,000 cases of 209 correlations from 154 samples in 103 publications and dissertations; 203 correlations from 151 samples in 99 studies; and 181 correlations from 166 samples in 78 studies. The SBDQ, LBDQ, LBDQ-XII, LOQ, and other measures of initiation and consideration were included. The estimated true score correlation was .48 between consideration and measures of follower satisfaction with job, motivation, leader, and leader effectiveness. The highest correlation was with satisfaction with the leader (r = .78). The somewhat lower estimated true score correlation was .24 for initiation of structure with follower performance. The comparable estimated correlation between consideration and initiation of structure was .17, supporting their factorial independence. Although there was variability, the results were generalizable across many conditions. These were negligible for SBDQ, lower at supervisory than middle or upper hierarchical levels, and lower for cross-sectional compared with longitudinal studies. The consideration estimated correlations were higher for studies in the public sector and colleges (r = .56, r = .53) and lower in business and the military (r = .43, r = .40). The initiation estimates were higher in the military studies (r = .46) and lower elsewhere (r = .29, .28, .25).
Generally, the consideration of supervisors seems to be associated with subordinates’ satisfaction with the supervisors, fewer absences, and less likelihood of quitting. But, as we have just seen, the correlations between a leader’s initiation of structure and the satisfaction and productivity of subordinates vary, in outcome, depending on the instruments used to measure them. The constraints and goals in the situation also make a difference. The personnel involved may be particularly important. Followers in a wide variety of groups consider it legitimate for the leader to exercise influence on matters related to the performance of tasks and the work environment (Fleishman & Peters, 1962). At higher executive levels, initiation of structure is seen in planning, innovation, and coordination; at lower levels, it is seen in the push for production (Brooks, 1955).
Although too much initiation often increases the likelihood of grievances, absenteeism, and turnover (Fleishman & Harris, 1962), a certain amount of pointing out the “paths to successful effort” (Bass, 1965c) is characteristic of the effective supervisor. This yields the greatest effectiveness and satisfaction in the work group—especially when workers are untrained, unmotivated, or both, for example, or when the group lacks cohesiveness. Untrained personnel need more help; trained people prefer less help.
The variety of mixed results noted in reviews7 of the effects of consideration and initiation suggests that to gain a better understanding of these effects, researchers need to specify the measures used and the conditions involved (House, 1971; Kerr, Schriesheim, Murphy, & Stogdill, 1974). For example, initiation of structure usually has been found to be associated with subordinates’ role clarity, but somewhat less frequently with subordinates’ performance. The correlation with subordinates’ satisfaction varies considerably from study to study (Fleishman, 1973). The leadership climate under which the supervisors work has been found to be related to the supervisors’ own consideration and initiating structure behavior and attitudes (Fleishman, 1953b).
As noted earlier, variations in results depend to some degree on whether the LBDQ, the SBDQ, the LBDQ-XII, or the LOQ has been used.
Schriesheim, House, and Kerr (1976) did a masterful detective job in reconciling the mixed results obtained with the various versions of the LBDQ and SBDQ in measuring initiation of structure. First, they pointed out that the LBDQ of Halpin (1957b) contained 15 items asking subordinates to describe the actual initiating structure behavior of their leader to establish well-defined patterns of communication and to set up ways to get the job done. But the revised LBDQ-XII (Stogdill, 1963a) contained 10 items for measuring initiation of structure that dealt with the actions of the leaders in clearly defining their own roles and informing followers about what was expected of them. Even more substantial differences were found in the leadership behavior tapped by the SBDQ and the LBDQ in a comprehensive item-by-item analysis of 242 employees’ descriptions of their supervisors. In addition to role clarification by the leader, as was noted earlier, the SBDQ included a cluster of items measuring punitive, autocratic, and production-oriented behaviors, such as “He rules with an iron hand” and “He needles those under him for greater effort.” Thus the three questionnaires—SBDQ and the earlier and later versions of the LBDQ—differed markedly in content. The initiation items of the LBDQ versions largely reflected communication and organization elements; the SBDQ initiating structure, however, contained domination and production pressure items.
An essential component of initiation of structure in all the instruments involved role-clarification behaviors. A specific aspect of role clarification—establishing methods to get the work done—is mentioned in the LBDQs, but other aspects, such as scheduling and criticizing, are found only in the SBDQ. Schriesheim, House, and Kerr (1976, p. 301) concluded:
When measured by the SBDQ, leader Initiation of Structure is generally positively related to performance ratings by superiors of manufacturing first-level supervisors’ subordinates.8 This generalization also holds with regard to noncommissioned … infantry officers and air force officers9 … with Initiating Structure being measured in these studies by a form containing items similar to the autocratic behavior items of the SBDQ. A similar although much weaker pattern of relationships has been found concerning non-manufacturing supervisors of clerical workers doing routine tasks … using selected items from the SBDQ in a laboratory experiment.10 When the revised LBDQ Initiating Structure scale is used to measure leader behavior of first-line supervisors of non-manufacturing employees performing routine tasks, correlations with subordinate satisfaction are positive, although generally [lower] … using a very modified version of the revised LBDQ.11
Meheut and Siegel (1973) demonstrated the differences by dividing the items on initiation of structure in the SBDQ into those concerned with role clarification and those concerned with the autocratic behavior of leaders. They obtained a correlation of .26 between leaders’ role clarification and subordinates’ satisfaction, but a correlation of –.21 between leaders’ autocratic behavior and subordinates’ satisfaction.
Another source of error that accounts for variations in results with the SBDQ, according to Schriesheim, House, and Kerr, is the SBDQ’s failure to provide opportunity for respondents to describe the timing or appropriateness of the structuring of the particular task or of the context in which respondents work, even though empirical evidence indicates that timing may be more important than the frequency of specific leadership behaviors (see, for example, W. K. Graham, 1968; Sample & Wilson, 1965). In addition, leaders who have adequate knowledge of the demands of their subordinates’ tasks may vary the amount of initiation of structure and the kind and timing of the structure they provide. Some tasks require more structure during the goal-setting (goal-clarification) stage, whereas others require more path clarification and feedback on performance. Furthermore, some subordinates need more administrative structure to relate their work to other employees. Other subordinates could benefit more from technical guidance. Leaders may have no control over standards. Their initiation of structure may depend on circumstances outside their purview. “He decides what shall be done and how it shall be done” may be physically impossible for leaders in some situations. Other items on initiation of structure, such as “He tries out his ideas in the group,” are less likely to be affected by circumstances.
Some items deal with specific behavior, and others deal with general tendencies. Such initiation of structure items as “He schedules the work to be done” refer to specific actions; items like “He encourages overtime work” relate to general practices. Even farther removed from specific behaviors and more concerned with the skills, traits, and personality attributed to the leader are such items as “He makes accurate decisions” and “He is a very persuasive talker.”
The impact of a leader’s initiation and consideration will depend on the organization in which they occur. House, Filley, and Gujarati (1971) found that both the leader’s consideration and the leader’s initiation of structure moderated the employees’ satisfaction with freedom on the job, job security, and family attitudes in one firm, but not in another. Similarly, Larson, Hunt, and Osborn (1974) found that in one state mental health institution, the leaders’ consideration was related to the performance of different groups of personnel, but that in another institution, initiating structure was more highly related to over-all groups’ performance. The report by Marsh and Atherton (1981–1982) of the results of a study of 30,000 U.S. Army personnel concluded that whether a military unit was mechanistic or organic moderated the extent to which the leader’s initiation or consideration was related to the subordinates’ satisfaction.
Differences in Function and Task. Fleishman, Harris, and Burtt (1955) noted that the leader’s greater initiation of structure contributed to absences and grievances of subordinates in manufacturing departments and to heightened rates of turnover in nonmanufacturing departments. At the same time, Fleishman, Harris, and Burtt found that supervisors in manufacturing departments or in other departments that were working under time constraints were likely to receive higher merit ratings from their own supervisors if they tended to exhibit more initiation of structure, whereas the reverse was true for supervisors in service departments. In addition, in the nonmanufacturing service departments, the more considerate supervisors were seen as more proficient. “Pressure for production” was the moderator.
In Cunningham’s (1964) study of county agricultural agents and 4-H club agents cited earlier, the agricultural agents’ consideration was significantly related to effectiveness, but initiation of structure was not. In contrast, initiating structure was significantly related to the effectiveness of 4-H club agents, but consideration was not. However, as was noted earlier, the most effective agents in both organizations were those who were described as above the median in both consideration and initiation. Mannheim, Rim, and Grinberg (1967) reported that manual workers tolerated more initiation of structure by their supervisors than did clerical workers. Only clerical workers tended to reject the high-structuring supervisor when they expected the initiation of structure to be low. When high consideration was expected, both groups chose the leader who conformed to their expectations.
Hunt and Liebscher (1973) showed that leaders’ consideration was more strongly associated with subordinates’ satisfaction in a construction bureau than in a bureau of design of a state highway department; the leaders’ initiation of structure did not vary as much in its effects on satisfaction. Dagirmanjian’s (1981) study of 126 mental health service personnel showed that when organizational differentiation was in force, the leaders’ consideration generated employee satisfaction with supervision, but when organizational integration was involved, the leaders’ initiation of structure generated such satisfaction.
Leaders in the Middle. Although Rambo (1958) failed to find any significant differences in the initiation and consideration of executives in different echelons of the hierarchical structure, Halpin (1956a) saw such hierarchical differences when he examined the reactions of organizational members at different levels to the leader in between them in level. Halpin studied the leadership of school superintendents described by staff members, school board members, and themselves on both the real and ideal forms of the LBDQ. He found that the board members agreed among themselves and the staff members agreed among themselves in their descriptions of the superintendents’ behavior, but the two groups differed significantly in their perceptions. Staff members saw the superintendents as less considerate than the superintendents saw themselves or the board members saw the superintendents. The board members described the superintendents as being higher in initiating structure than did the staff members or the superintendents. The staff members and the board members differed significantly regarding how considerate the superintendents should be, but they did not differ significantly about the extent to which superintendents should initiate structure. The board members expected the superintendents to act in a more considerate manner than the staff members considered ideal. There was a nonsignificant tendency for board members to expect more initiation of structure than either the staff members or the superintendents considered ideal.
Other studies in educational institutions by raters of a leader in the middle were conducted by Sharpe (1956), Carson and Schultz (1964), and Luckie (1963). Sharpe studied the leadership of principals as described by teachers, staff members, and the principals themselves. The three groups held similar ideals of leadership behavior, but the teachers and staff members perceived the principals as deviating less from the ideal norms than did the principals themselves. Occupants of high-status positions perceived the principals as deviating more from the ideal norms than did those in lower-status positions. Carson and Schultz (1964) obtained descriptions of junior college deans by college presidents, department heads, student leaders, and the deans themselves. The greatest discrepancies were found between the presidents’ and student leaders’ perceptions and expectations of the dean’s behavior. The evidence suggested that the greatest source of role conflict for the deans was the discrepant expectations of their behavior. Luckie (1963) obtained 434 descriptions of 53 directors of instruction by superintendents, staff members, and the directors themselves. The results indicated that the instructional directors actually behaved at a lower level of consideration than superintendents, the directors themselves, and the staff members rated as ideal. The superintendents and the staff members expected the directors to exhibit higher degrees of initiating structure than the directors considered ideal.
The conflict of the leader in the middle seems to reside in the question of how considerate to be, not how much structure to provide. Graen, Dansereau, and Minarm (1972b) obtained data indicating that at lower organizational levels, both superiors and subordinates evaluated the leader in between them more highly if the leader initiated more structure. But the leader’s consideration had more of an impact on subordinates than on superiors.
Lawrie (1966) used real and ideal consideration and initiation in a study of superiors’ and subordinates’ expectations of foremen in two departments. The convergence between real and expected behavior, as described by subordinates, was not related to ratings of the foremen’s effectiveness. However, in one of the two departments, the foremen’s ability to predict the superiors’ expectations and the congruence between the foremen’s and the superiors’ expectations were related to ratings of the foremen’s effectiveness.
Influence “Upstairs.” Leaders’ influence with higher authority has been found to affect the impact of their initiation of structure and consideration on their subordinates’ satisfaction and performance. Consideration often involves promises of tangible rewards, and leaders may need influence “upstairs” to deliver on their promises. Initiation of structure involves setting forth goals and plans; influence “upstairs” adds to the leaders’ ability to do so with authority and credibility. In one of two companies studied, House, Filley, and Gujarati (1971) found the expected strong positive relation between a supervisor’s influence with a higher authority and an increase in the correlation of the supervisor’s consideration with the subordinates’ satisfaction. As Wager (1965) found earlier, in both the companies studied, the greater a supervisor’s influence with higher authority, the greater was the supervisor’s tendency to be considerate. Presumably, the influential leader could offer support and rewards with more certainty of providing them. The more influential supervisor exhibited more initiation of structure in only one of the firms studied, but, generally, employees were more satisfied with most aspects of their jobs if their supervisors were more influential with higher authority.
Falling Dominoes Again. The falling dominoes effect, discussed at length in earlier chapters, was observed by Hunt, Hill, and Reaser (1973) for results with the LBDQ-XII. In a school for the mentally retarded, an increase was found in the association of considerate supervision and the performance of aides when the LBDQ scores of the second-level and first-level supervisors were combined. Hunt, Osborn, and Larson (1975), with data from three mental institutions, showed that leaders’ consideration had more of a positive impact on their groups’ performance if the leaders’ superiors were high in authoritarianism. However, group performance was higher if the leaders’ initiation of structure was low, whether the leaders’ superiors were authoritarian or egalitarian.
In his first study with the LOQ, Fleishman (1953a) found that the higher the supervisors’ positions in the hierarchy of a plant, the less considerate the supervisors thought they should be and the more structure they thought should be initiated. These attitudes had an impact on those below them. Foremen whose superiors expected them to lead with less consideration and with more structuring revealed high grievance rates among their subordinates.
Size of Group Led. Ordinarily, one would expect that when a group is enlarged, a leader would have to display more initiation of structure to be as effective as before the unit was enlarged (Bass, 1960). It would be expected that the leader would find it increasingly difficult to maintain the same level of consideration for the concerns of all subordinates as the group enlarged. Corollary results that are consistent with this expectation were found by Badin (1974), who showed that in the smaller of 42 work groups of 489 manufacturing employees, the initiation of structure by the leader correlated negatively with productivity, but that the amount of a supervisor’s initiation of Structure was unrelated to productivity in the larger groups.
Osborn and Hunt (1975b), in their previously cited study of presidents of chapters of a business fraternity, found that the size of a chapter moderated the positive effects of initiation and consideration on the members’ satisfaction. However, Sheridan and Vredenburgh (1979) failed to find any relation between the size of units led and the head nurses’ leadership behavior, although size correlated with perceived job tension, which, in turn, was related to the subordinates’ performance.
Structure of the Work Group. J. A. Miller (1973a) deduced that more initiation of structure would be contraindicated in a highly structured setting. Thus for a supervisor to tell skilled crafts personnel how to do their job was expected to be detrimental to their performance. Consistent with this expectation, Badin (1974) found that when first-line supervisors initiated structure a great deal in the previously mentioned 42 work groups at the manufacturing firm, effectiveness was reduced in the groups that were already most structured. In these highly structured groups, the correlation of group effectiveness and the supervisors’ initiation of structure was –.56. But in the less structured work groups, the effectiveness of the groups was correlated .20 with the extent to which their first-line supervisors initiated structure.
Also consistent with Miller’s argument, Jurma (1978) demonstrated, in an experimental comparison, that workers in 20 discussion groups were more satisfied when leaders provided discussants with task-related information, gave guiding suggestions, helped groups to budget their time, and established group goals. This was especially true when the workers were faced with a debatable, ambiguous discussion task rather than one with clear choices for making a decision.
Dyadic versus Group Relationships. By the late 1970s, considerable research had concentrated on the dyadic leader-subordinate relationship (Graen & Schie-mann, 1978), instead of on the leader and the primary work group (Hunt, Osborn, & Schriesheim, 1978). Substantive differences in results from standard leader-group investigations have failed often to emerge. Thus, for example, in the case of the LBDQ, for 308 managerial and clerical employees in 43 work groups in a public utility, C. A. Schriesheim (1979a) found correlations of .77 and .89 between dyadic and group LBDQ descriptions of supervisors for initiation and consideration, respectively. Nevertheless, Yammarino (1990) collected LBDQ-Form XII data in 13 groups from 54 members of a campus police department at various hierarchical levels from patrol officers and dispatchers to lieutenants and directors. Each group reported to a single superior. Yammarino found for that for initiation of structure, individual member and group correlations of supervisory control and satisfaction with rewards were due to differences between the groups but not within them. Individual supervisors were the reason. The same was true for the correlations between consideration and role ambiguity and effort. But the correlations of the LBDQ assessments with satisfaction with supervision were a matter of individual differences.
Cohesion. Among 308 low-and middle-level managerial and clerical employees, J. F. Schriesheim (1980) used a modification of the LBDQ-XII. She asked subordinates to indicate how their superiors acted toward them as individuals rather than toward the group as a whole. In line with her expectations, she showed that when the cohesion of the work group was low, the leader’s initiation of structure was positively related to the subordinates’ satisfaction with supervision, role clarity, and self-rated performance. But when cohesiveness was high in the work group, the leader’s consideration was positively related to the measures of satisfaction, clarity, and performance.
Alignment with Objectives. The contribution of initiation and consideration to productivity will depend on the extent to which the group of subordinates is supportive of the objectives of the productivity. In Howell’s (1985) experiment, detailed in Chapter 21, the leader who initiated structure generated more than average actual productivity in the experimental subject if the confederate fellow worker established high norms for productivity. The experimental subject produced much less if the confederate induced low norms for productivity. The leader’s consideration had the same effect, but to a lesser extent. Similar results occurred for the effect of initiating structure on satisfaction and freedom from role conflict, but here the effects of consideration were not moderated by productivity norms set by the confederates of the experimenter.
Hernandez and Kaluzny (1982) studied the leadership and performance of 20 work groups of public health nurses. They found that the supervisors’ initiation of structure had a strong positive relationship to productivity but that consideration did not. Likewise, satisfaction was enhanced by the supervisors’ and peers’ initiation of structure. Instead of the amount of communication flow and group processing correlating positively with the groups’ productivity and satisfaction, the reverse occurred. The explanation of these results lay in the lack of support by the nurses for the appropriateness of the services they offered.
Group Conflicts. Various kinds of conflict between the group and external agents, as well as within groups, moderate the extent to which consideration and initiation will be effective. R. Katz (1977) found that considerate leadership was most effective when the group faced external conflicts but that initiation of structure by the leader was most effective in dealing with internal interpersonal conflicts.
Stumpf (undated) completed a path analysis for questionnaire data from 144 professionals in a government R & D organization. Leadership behavior was not directly related to the subordinates’ job satisfaction or performance, but it was related through two moderating or intervening variables that linked the leader’s consideration and initiation to subordinates’ job satisfaction and performance. The leader’s initiation of structure correlated with the subordinates’ skill-role compatibility, which in turn correlated with the subordinates’ job satisfaction. Thus the R & D professionals’ satisfaction with their jobs depended on the extent to which their skills were not in conflict with the demands of their roles, which in turn depended on the structuring of the situation by their supervisors.
Curvilinear Effects. Skinner (1969) obtained results indicating that supervisors who scored high in consideration experienced lower than average grievance and turnover rates among their subordinates. As did Fleishman and Harris (1962), Skinner concluded that the consideration of supervisors bears a curvilinear relationship to turnover and grievances of employees. As consideration increases, grievances decrease to a point and then level off. Supervisory initiation also had a curvilinear relationship to grievances.
Interaction of Effects. The interacting effects of consideration and initiation of structure were observed to vary, depending on the situation. Thus considerate foremen could initiate structure without increasing turnover or grievances (Cummins, 1971). Fleishman and Harris (1962) found in using the SBDQ that consideration and initiation of structure by foremen interacted to affect the grievances and turnover of employees. Medium and high degrees of consideration, along with low degrees of initiation of structure by foremen, were associated with the lowest rates of turnover and grievances of employees. Graen, Dansereau, and Minami (1972a) found, among 660 managers of a large corporation, that for those managers who saw their leader as either extremely high or extremely low in initiating structure, the relationship between the leader’s consideration behavior and the managers’ performance evaluation was positive. But for those managers who saw their leader as intermediate in initiating structure, the relationship between the leader’s consideration and the managers’ performance was near zero.
But other studies failed to establish any interaction effects of initiation and consideration on outcomes. Filley, House, and Kerr (1976) studied three companies to test Fleishman and Harris’s (1962) hypothesis that the leader’s initiation of structure acts as a mediator of the relationship between the leader’s consideration and employees’ job satisfaction. Initiating structure was positively and significantly related to satisfaction with the company in all three organizations. Consideration was significantly related to satisfaction with the company and freedom of action in all three organizations. The data failed to support the mediating hypothesis. Similarly, as will be examined more fully in Chapter 27, M. G. Evans (1970a) tested the path-goal hypothesis in two organizations. He found that consideration and initiation of structure did not interact in path-goal facilitation but, rather, that consideration and initiation of structure each acted separately to enhance path-goal instrumentality. The mixed findings suggest that the occurrence of the interaction effect depends on particular circumstances.
Contingent Aspects of the Subordinates. As might have been expected from what has been presented regarding the effects of autocratic leader behavior compounded with authoritarian subordinates, Weed, Mitchell, and Moffitt (1976) found that leaders who initiated structure more with dogmatic subordinates were likely to achieve higher levels of performance; leaders who exhibited more consideration yielded greater performance among less dogmatic subordinates. In the path analysis by Sheridan and Vredenburgh (1979), greater job experience of the nursing personnel was found to be related to less considerate leadership from the head nurses, which, in turn, increased the subordinates’ tension and had mixed effects on their performance.
M. Beer’s (1964) industrial study of the effects of leadership behavior on subordinates’ fulfillment of needs disclosed that subordinates whose need for self-actualization, esteem, autonomy, production emphasis, and consideration was high were positively motivated by the leader’s initiation of structure, contrary to the hypothesis that the leader’s consideration would be more effective in motivating subordinates. Abdel-Halim (1981) collected data from 89 lower-and middle-level manufacturing managers and demonstrated that the subordinates’ locus of control moderated the effects of the ambiguity and complexity of tasks on the effectiveness of initiation and consideration.
Contingent Aspects of the Leaders. The effects of initiation or consideration appear to depend on whether the leader has a strong need for achievement. Leaders with a high need for achievement who initiated structure obtained a better performance from their subordinates than did leaders with a low need for achievement who also initiated structure (T. Mayes, 1979). Another illustration of a contingent effect that is due to the leader was provided by Miklos (1963), who observed that the longer the tenure of school principals who were high in the initiation of structure, the greater the consensus among teachers in their role expectations for teachers and the greater the agreement between teachers and principals. Agreement between teachers and principals was highest when principals were rated high in both consideration and initiation of structure.
The expert power of the leader also moderated the effects of initiation and consideration. Path analyses by Martin and Hunt (1980) that initiation and consideration, when combined directly or indirectly with expert power, linearly enhanced the group’s cohesiveness, job satisfaction, and intentions to remain on the job. Podsakoff, Todor, and Schuler (1983) helped to explain this effect of expert power and leadership. They collected data from 101 employees of a large nonprofit organization who used modified LBDQ scales to describe their supervisors. As the supervisors’ initiation of structure increased, if the supervisors were highly expert, the role ambiguity of the subordinates decreased. However, if the leaders who were high in initiation were low in expert power, the reverse occurred—the role ambiguity of the subordinates increased. A similar though less extreme pattern emerged with the increasing consideration of the leaders. Expert considerate leaders generated less role ambiguity in their subordinates, whereas inexpert considerate leaders generated more role ambiguity.
The limitations of the results of LBDQ surveys in ascertaining whether the described leadership is a cause, a consequence, or a coincidence of group effectiveness, satisfaction, or other valued outcomes has already been noted. Do considerate and structuring leaders promote the satisfaction and productivity of subordinates, or do leaders of satisfied and productive subordinates show them more consideration and tend to initiate less structure? Cross-lagged analyses suggest that the causality is reciprocal. Leaders affect subsequent outcomes, and outcomes affect the leaders’ subsequent behavior (Greene, 1973, 1974, 1975, 1979a).
Greene (1975) asked first-level managers in insurance, marketing, finance, and research and engineering to describe their own consideration and initiation of structure on the LBDQ. Two subordinates of each of the managers rated their own work satisfaction. Peers of these subordinates rated the subordinates’ productivity. The measures were picked up on successive occasions one month apart. Cross-lagged comparisons were made of the leaders’ behavior at time 1 correlated with the subordinates’ satisfaction or productivity at time 2 and contrasted with the subordinates’ satisfaction or performance at time 1 correlated lated with the leaders’ behavior at time 2. The results strongly suggested that considerate leadership resulted subsequently in the increased satisfaction of subordinates. At the same time, the subordinates’ productivity resulted subsequently in their managers’ increased consideration for them. Finally, the subordinates’ productivity subsequently resulted in a reduction in the managers’ initiation of structure.
Even greater confidence about the cause and effects of initiation and consideration comes from the rare experiment in which the leader’s initiation or consideration (or both) is arbitrarily manipulated experimentally and measurements are taken of the subsequent effects on subordinates. Some results were obtained to support the argument that initiation or consideration causes the subsequent improved performance or increased satisfaction of subordinates. Dawson, Messé, and Phillips (1972) arranged for teachers to increase their initiation of structure deliberately. They found that their students’ work group productivity improved as a consequence. When the teachers subsequently increased their consideration deliberately, the students’ work-group productivity again increased as a consequence. In a laboratory experiment by Gilmore, Beehr, and Richter (1979), 48 participants were subjected to either low or high consideration and low or high structure by specially coached student supervisors who were working with groups of four members each. The LBDQ completed by the members failed to discriminate the behavior of the four types of supervisors, even though the leadership behavior itself had differential effects on the quality of the participants’ performance. The quality of the participants’ performance was highest with leaders who were high in initiation of structure but low in consideration. But neither the amount of initiation nor the amount of consideration was related to various measures of satisfaction of the participants. Finally, as mentioned before, Howell’s (1985) experiment demonstrated that if group norms were supportive of high productivity, the leaders’ initiation and consideration resulted in a rise in productivity and satisfaction.
Hundreds of statements about leaders’ behavior formed two factors—consideration and initiation of structure—but it was recognized that some important elements in leadership might still be left out. Twelve factors were found to exhaust the list. To initiation and consideration were added representing, reconciling, tolerating uncertainty, persuasiveness, tolerating (followers’) freedom, role retention, production emphasis, predictive accuracy, integrating (of organization), and influence with supervisors. Although factorially and conceptually independent, the leader’s consideration and initiation of structure factors were ordinarily found to correlate moderately with each other in their revised measurement with LBDQ-Form XII scales.
The two factors are systematically affected by many personal and situational variables. Initiation of structure becomes more important when the group is not already highly structured. Psychometric reviews of results using the original and revised LBDQ and the SBDQ indicate that the generally negative associations between task-oriented initiation of structure, as measured by the SBDQ, and satisfaction and morale become positive when the coercive elements are removed. But leniency is likely to continue to bias the results, along with halo effects when single sources of variance are used to evaluate both leadership and outcomes. Self-ratings seem unlikely to indicate what leaders do according to their subordinates. Moreover, many contingencies can be cited as moderators of the relationships. A variety of expanded and alternative factor structures are available for a more detailed study of leadership behavior with the LBDQ-XII.
Causal analyses by cross-lagged surveys and experimentation suggest that consideration both increases the satisfaction of subordinates and is increased by it. Initiation of structure by the leader (if structure is low) improves the subordinates’ performance, which in turn increases the leader’s subsequent consideration and reduces the leader’s initiation of structure.
Several components that are variously involved in the different versions of scales that measure the leader’s initiation of structure must be taken into account for a full appreciation of their antecedents and their effects on the subordinates’ performance. The elements of role and task clarification are likely to have positive effects on satisfaction and productivity; the autocratic elements in the earlier LBDQs may have negative effects, especially on satisfaction. Consideration, likewise, may contribute differently in different situations to satisfaction and effectiveness as a consequence of its several components, including participatory and consultative decision making and concern for the welfare of subordinates. As we shall see next, together initiation and consideration can account for the transactional exchange between the leader and the led, but much more is involved when the leadership is transformational.