This chapter is partly a revision and expansion of a longer version of a seminal publication by Ralph Stogdill (1948) concluding that leaders differ consistently in personal dispositions or traits and are propelled into leadership position if warranted by the needs of the situation.1 A review of research studies up to 1948, up to 1970, then up to 2005 will follow in this and the next chapter.
Up to the middle of the twentieth century, leadership research was dominated by the search for the traits of leadership. In Britain in the nineteenth century, personal traits such as height, weight, health, and education had been found to correlate with leadership. But as noted by Shackleton (2001), the powerful and wealthy were better fed, clothed, and housed than factory workers and farm laborers. They could pay for an education. Smith and Krueger (1933) surveyed the preceding studies on leadership to 1933. Developments in leadership methodology, as related especially to military situations, were reviewed by Jenkins (1947).
Stogdill (1948) reviewed 128 published studies that tried to determine the traits and characteristics of leaders. The results of the survey included those characteristics and traits that were studied by three or more inves tigators. When contradictory evidence occurred, those studies with positive, negative, and neutral results were presented. He concluded that to some degree the traits of leadership needed to match the needs of the situation. What follows first are shortened excerpts of Stogdill’s 1948 review.
Evidence of the relation of age to leadership in children was quite contradictory. Pigors (1933) observed that leadership did not appear in children before age two or three and then usually took the form of overt domination. Active leadership of a group seldom appeared before age nine or ten, at which age the formation of groups and gangs became a noticeable feature in the social development of children.
According to Pigors, four conditions were necessary for the appearance of leadership in children: (1) development of determination and self-control; (2) grasp of abstractions and social ideals; (3) awareness of personalities; (4) a sufficient memory span to pursue remote goals rather than immediate objectives. Arrington (1943), however, found no evidence from a survey of time-sampling experiments to support the proposition that leadership increases with age in preschool children.
Hunter and Jordan (1939) and Remmelin (1938) found leaders to be significantly younger than their followers. Bellingrath (1930) obtained results indicating that girl leaders were younger than nonleaders but boy leaders were older. However, leaders were found to be older than followers in 10 studies completed between 1915 and 1939 (Gowin, 1915; Zeleny, 1939). Correlations between age and leadership obtained in 13 studies ranged from 2.32 to .72 with an average of .21.
Gowin (1918) established that on average, outstanding executives were 12.2 years older than lower-level executives. But Ackerson (1942) and Brown (1933) failed to find leaders and followers different in age.
According to Caldwell and Wellman (1926), the relationship of age to leadership differed in various situations. Leaders in athletics were found to be close to the class average in age as were girl club leaders and student council and citizenship representatives. The correlation of chronological age with leadership appeared to depend on other variables. For instance, it was highly positive in an organization like the Roman Catholic church if policies dictated that rank was determined by age; but it was lower or negative in organizations that quickly promoted their “best and brightest” up through the ranks while their less endowed peers reached a plateau early.
Leaders were found to be taller in nine studies, shorter in two studies, and heavier in seven studies and lighter in two. No differences in height or weight were reported in four studies, and in two studies the results depended on the situation.
Height. Correlations between height and leadership ranged from 2.13 to .71. The general trend of these studies was a low positive relationship between height and leadership. The average correlation was about .30. Nevertheless, Hunter and Jordan (1939) and Garrison (1933) found student leaders to be somewhat shorter than non-leaders. Baldwin (1932) and Reynolds (1944), could find no relation at all between height and leadership in students.
Weight. The correlations obtained suggested a low positive relationship between weight and leadership. The average correlation was about .23. But again, Hunter and Jordan (1939) found the reverse—that leaders were significantly lighter than nonleaders. Garrison (1933) and Moore (1935) also reported leaders to be somewhat lighter than followers. Since height and weight correlated positively, the same pattern of results was obtained for weight as for height.
Physique was positively associated with leadership in five studies; athletic ability and physical prowess in seven; energy in five; and health in four. Moreover, health and physical condition were found to be a factor in four of the reviewed investigations.
Physique. Five studies between Webb (1915) and Bernard (1928) reported that superior physique was a characteristic of leaders. The correlations of .28, .18, .11, and .23 found by Kohs and Irle (1920), Nutting (1923), Sheldon (1927), and Webb (1915), respectively, suggested that this relationship was slight but reliable. But Bowden (1926) concluded from the results of his study of college students that leadership was not the result of a dominating physique, and Baldwin (1937) found that high school leaders did not differ from followers in freedom from physical defects.
Health. Leaders, according to Baldwin (1932), Bell-ingrath (1930), Reals (1938), and Stray (1934), appeared to have some advantage over nonleaders in possessing better health, although Ackerson (1942) and Hunter and Jordan (1939) failed to find that health was a differentiating factor.
Athletic Prowess. Athletic ability and physical prowess did appear to be associated with leadership status in boys’ gangs and groups. Evidence to this effect was presented in seven studies between Puffer (1905) and Flemming (1935). Correlations of .38, .62, and .40 between athletic ability and leadership were reported by Flemming (1935), Patridge (1934), and Webb (1915), respectively.
Energy. According to Bellingrath (1930), Brown (1934), Cox (1926), Stray (1934), and Wetzel (1932), leaders were also characterized by high energy. But Cox (1926) found that various groups of great leaders differed markedly from each other in physique, energy, use, and athletic prowess; only military leaders were outstanding in these traits.
Leaders presented a better appearance in 11 studies. They were better dressed in two others; but no relationship was found in one study, and appearance was negatively correlated with leadership in two studies. The results clearly depended on circumstances. Thus Dunkerley (1940) found that students who were chosen as leaders in social activities differed significantly from nonleaders in appearance and dress, but students chosen as leaders in intellectual and religious activities did not differ markedly from nonleaders in these respects. A correlation of .21 between attractive appearance and leadership in high school students was reported by Flemming (1935), but the correlation between leadership and being seen as beautiful was only .05. Yet Partridge (1934) found a correlation of .81 between ratings of appearance and leadership among adolescent boys. In Goodenough’s (1930) study a negative correlation of 2.20 was found between beauty and leadership in preschool children.
Tryon’s (1939) analysis suggested that appearance is more closely associated with leadership in boys than in girls. Tryon reported correlations with leadership of .49 and .06, respectively, for 15-year-old boys and girls; the correlations for 12-year-old boys and girls were .31 and .08, respectively. Ackerson (1942) obtained correlations of .12 and 2.06 between leadership and being seen as slovenly for boys and girls, respectively; a slovenly appearance and leading others into misconduct were correlated .32 and .31 for delinquent boys and girls.
Speech was positively associated with leadership in all 12 studies that examined the relationship, although a few of the results were marginal.
Tone of Voice. Baldwin (1932) reported a definite trend for teachers to rate the tone of voice of leaders as confident and the tone of voice of nonleaders as lacking in confidence. Flemming’s (1935) factor analysis of teachers’ ratings of high school leaders revealed “pleasant voice” as one of the four factors found to be associated with leadership. The correlation between “pleasing voice” and leadership in high school students was .28. Partridge (1934) reported that boy leaders could be reliably distinguished from nonleaders when in the presence of strange boys but hidden from view, so that judgments had to be made on speech alone. However, Fay and Middleton (1943), in repeating this experiment under somewhat similar conditions, found a correlation of only .08 between leadership ratings and degree of leadership, as estimated by voice alone. Likewise, Eichler (1934) reported a correlation of only .11 between voice and leadership.
Talkativeness. Talkativeness and leadership were reported by Tryon (1939) to be correlated .41 and .31 for 12-year-old boys and girls, respectively; the correlations for 15-year-old boys and girls were .15 and .44, respectively. In Goodenough’s (1930) study, a correlation of .61 between talkativeness and leadership was found. Thur-stone (1944) did not find that highly paid administrators surpassed their lower-paid associates in scores on tests of word fluency, but he did find a significant difference in their scores on tests of linguistic ability. Simpson (1938) also reported that verbal ability was correlated .45 with the capacity to influence others.
Fluency. Chevaleva-Ianovskaia and Sylla (1929) noted that child leaders were characterized by longer durations of verbal excitation. Terman (1904) reported that leaders were more fluent in speech, and Leib (1928) observed that leaders excel in speaking ability. The same skills were reported in adult leaders by Bernard (1928) and Merriam (1926). Zeleny (1939) found a correlation of .59 between leadership ratings and total remarks made in class. Interesting conversation and leadership were correlated .28 in Flemming’s (1935) study. Finally, Burks (1938) and Malloy (1936) found that vividness and originality of expression and facility of conversation were associated with successful social relationships. Considering the size of the experimental groups, the competence of the experimental methods employed, and the positive nature of the evidence presented, it was apparent that fluency of speech, if not tone of voice, was a factor to be considered in the study of leadership. It has long been recognized that effective leadership cannot be maintained in an organization without an adequate system of intercommunication. Thus it does not seem surprising that some of the most searching studies of leadership should reveal the capacity for ready communication as one of the skills associated with leadership.
All but 5 of 23 studies presented evidence that the average leader surpassed the average member of his or her group in intelligence. However, five of the studies suggested that too great a difference in the IQ of the leader and the average member will militate against leadership. Statistically reliable differences were reported, for example, by Hunter and Jordan (1939), Remmelin (1938), and Sward (1933). In most of these studies, there was considerable overlapping of the scores of leaders and nonleaders on intelligence tests, indicating that superior intelligence was not an absolute requirement for leadership. Nevertheless, the general trend of the findings suggested that leadership status was more often than not associated with superiority in intelligence. The correlations revealed a consistently positive relationship. The average of these coefficients was approximately .28.
Factorial studies demonstrated a number of points that may be of considerable significance for the study of leadership. Cattell (1946), for example, reported that the intelligence factor is heavily weighted with such elements of character as being wise, emotionally mature, persevering, mentally alert, vigorous, and conscientious. These items correspond fairly closely to the factors discussed later, which were found in the present survey to be supported by an excess of positive over negative evidence. For example, Thorndike (1936) reported a correlation of .60 for 305 male members of European royal families between their intellectual ability and their esteem (which, in turn, is related to leadership). Thus it appears that high intelligence may be associated with other characteristics that contribute to a person’s value as a leader.
Limits. One of the most significant findings concerning the relation of intelligence to leadership is that extreme discrepancies between the intelligence of potential leaders and their followers militated against the exercise of leadership. Hollingworth (1926) found that among children with a mean IQ of 100, the IQ of the leader was likely to fall between 115 and 130. That is, the leader was likely to be more intelligent, but not too much more intelligent than the average of the group led. Observation further showed that a child with an IQ of 160 had little chance of being a popular leader in a group of children of average intelligence but might become a leader in a group of children with a mean IQ of 130. One of the difficulties in this connection seemed to be communication. The average child cannot comprehend a large part of the vocabulary used by a child of unusually superior intelligence to express exact meanings in relation to his or her more mature and complicated interests. Differences in interests, goals, and activity patterns also act as barriers to joint participation, which is a necessary condition of group leadership.
Hollingworth’s findings were confirmed by a number of investigations. Finch and Carroll (1932), studying groups of 66 gifted, 66 superior, and 66 average children, arrived at the conclusion that, “Given a superior group of children to lead, the leading will tend to be done by the gifted children,” even though the leaders as a group tend to be younger than the group led. However, in an early study of the formation of boys’ gangs, Warner (1923) found that leaders and followers differ much more in chronological age than in mental age. She observed that older boys who were mentally below normal tended to group with younger boys who had a mental age near their own and slightly higher, and that when groups of retarded delinquent boys contacted groups of brighter delinquents, the contacts were “so short and non-social that no noticeable event took place.” Mailer (1925), studying cooperation and competition among children, found that homogeneity of intelligence, rather than level of intelligence, was important in cooperative behavior. McCuen (1929) studied leadership in 58 organizations of college students. He concluded that “the crowd seems to desire to be led by the average person. Evidently in a democratic society, the leader must not be too far detached from the group.”
Two studies by Lehman (1937, 1942) are of interest in this connection. In the first study, Lehman determined the age intervals at which outstanding men in various professions made their best contributions. In the second study, he determined the optimal age intervals for eminent leadership. Chemists, for example, were found to make their best contributions during the age intervals 28–32 years, while the optimal ages for eminent leadership in chemistry were 45–49 years. Thus, it appears that even in science, individuals contributions and communications must be understood by, and in accord with the thinking of their contemporaries for them to rise to a position of leadership in their profession.
Leaders were reported to have better scholastic records in 22 studies and poorer records only once. No differences were obtained in four investigations. It is not surprising that leaders were found, with a high degree of uniformity, to have better average scholastic grades than did nonleaders, for, as was just noted, leaders had been found more intelligent than their followers. But the findings by Buttgereit (1932), Caldwell and Wellman (1926), and many others suggested that superior scholarship may not be a mere byproduct of superior intelligence, but may have direct importance for leadership when it is one aspect of a general ability to get things done. It was also suggested that superior accomplishment in areas valued by the group had prestige value, which may also contribute to leadership. At the same time, there was an abundance of evidence to indicate that a position of leadership was ordinarily not based on superior intelligence and accomplishment alone, since these two factors were present, to a high degree, in many persons who did not occupy positions of leadership. Thus, overall, the magnitude of the correlations suggests that intelligence and scholarship account for only a fraction of the total complex of factors associated with leadership.
The results of all 11 studies dealing with leadership and knowledge implied that persons who are chosen as leaders tended to know how to get things done. Of particular interest was Caldwell’s (1920) experiment in which he asked 282 high school students to nominate boy and girl leaders for three different situations: (1) a trip to a wharf; (2) the production of a program and its presentation at a neighboring school; (3) the reorganization of a program for administering athletics in the school. The nominations revealed “a clear judgment on the part of these pupils as to the members of the group best fitted to lead them.” The most important abilities ascribed to these leaders were intelligence and practical knowledge about the situations for which they were chosen as leaders.
In this connection, it seems worthwhile to consider the findings of Baldwin (1932) and Burks (1938) in relation to the association between leadership and the ability to make constructive and creative suggestions. Burks, for example, found that the ability to present constructive ideas for difficult situations was closely associated with successful social relationships.
The studies of Cox (1926), Drake (1944), Flemming (1935), Stray (1934), and Thrasher (1927) found that constructive imagination was a characteristic of leaders. Additional evidence related to the ability to get things done was presented by Bellingrath (1930) and Dunkerley (1940). Cox (1926) and Peck (1931) reported that great leaders were characterized and differentiated from the average by a greater intensity of application and industry. In summary, the results of these various studies seemed to indicate that specialized knowledge, imagination, and the ability to get things done were factors that contributed to leadership.
Soundness and finality of judgment were related to leadership in five studies; and speed and accuracy of thought and decision were related in four analyses. In view of the positive correlations found between intelligence and leadership, it was not surprising to find a similar relationship between judgment and leadership. Bellingrath (1932), Drake (1944), and Webb (1915) reported correlations ranging from .34 to .69 between common sense and leadership; Bellingrath (1930), Drake (1944), Flemming (1935), and Webb (1915) found correlations of 60, .34, .28, and .69, respectively, between judgment and leadership. Farsightedness and leadership were found to be correlated .55, .25, and .33 in the studies of Bellingrath, Drake, and Webb, respectively. Two of the factor-analytic studies—those of Cowley (1931) and Dunkerley (1940)—revealed that soundness and finality of judgment were common to leaders. In addition to judgment, Cowley (1931) also identified three factors that appeared to represent speed of decision. Hanawalt, Richardson, and Hamilton (1943) found that leaders used the “?”, or “undecided,” response on the Bernreuter Personality test significantly less frequently than nonleaders, and this tendency was especially noticeable on the most discriminating items. In spite of the small number of studies bearing on judgment and decision, the general competence of the methods lent confidence to the results obtained.
Leadership was found to be related to certain aspects of insight, as follows: “keenly alive to environment, alert” (six studies); ability to evaluate situations (five studies); social insight (five studies); self-insight (two studies); and sympathetic understanding (seven studies). Traditionally, insight has been regarded as one aspect of general intelligence. However, Jennings (1943) and others suggested that insight may be socially conditioned to a high degree. Brown (1931), Buttgereit (1932), Caldwell and Wellman (1926), Cox (1926), Dunkerley (1940), and Fauquier and Gilchrist (1942) found that leaders were characterized by alertness and keen awareness of their environment. The ability to evaluate situations was found to be a factor in the studies by Bowden (1926), Buttgereit (1932), Chevaleva-Ianovskaia and Sylla (1929), Merriam and Gosnell (1929), and Thurstone (1944). Less clearly defined was social insight, reported as a factor associated with leadership in the studies by Bowden (1926), Hanf-mann (1935), Jennings (1943), Pigors (1933), and Zeleny (1939). Brogden and Thomas (1943) and Guilford and Guilford (1939) found that a person who “studies the motives of others” was measured by Guilford’s T factor of being thoughtful, and was viewed as offering a kind of intellectual leadership.
The results of these various studies suggested that alertness to one’s surroundings and an understanding of situations were intimately associated with leadership ability, yet little was understood about the nature of these processes. No worker who is responsible for improving the social effectiveness of individuals can fail to be impressed by the persistent blindness of maladapted individuals to the social situations to which they are attempting to adjust. From the point of view of understanding personal qualifications for leadership, one issue that still needs thorough investigation is the fundamental nature of awareness and social insight.
Although only seven studies contained data on originality, the magnitude of the positive correlations use suggested that the relationship between originality and leadership was worthy of further investigation. The correlations reported by Bellingrath (1930), Drake (1944), Flemming (1935), and Webb (1915) ranged from .38 to .70 and were higher, on average, than those for any other trait except popularity. At the same time, Cox (1926) found that great leaders rated unusually high in originality.
The 10 studies involving adaptability and leadership suggested that ready adaptability to changing situations might be associated with leadership, although the correlations of .13 and .21 reported by Eichler (1934) and Flemming (1935) respectively were not impressive The ability to adjust to situations has also been regarded traditionally as an aspect of general intelligence but, as described in the investigations considered here, this factor appeared to have a large social component. This fact has long been recognized by clinical observers, who have repeatedly pointed out that persons of high intelligence may be rendered ineffectual in their vocational, social, and other adjustments through extreme self-preoccupation and inhibition to action. Such inhibition is negatively correlated with leadership.
Leaders were found to be more extroverted in five studies and more introverted in three. No differences emerged in four studies. However, the only studies that reported a marked relationship between extroversion and leadership were those of Goodenough (1930) and Sward (1933). Goodenough reported a correlation of .46 between extroversion and leadership in children. Sward found that leaders scored reliably higher than nonleaders in extroversion on the Heidbreder scale. Richardson and Hanawalt (1943) observed that college leaders scored reliably lower in introversion than the Bernreuter norms and lower than nonleaders, although the difference between leaders and nonleaders was not significant. Hunter and Jordan (1939) and Remmelin (1938) also reported that introversion scores on the Bernreuter scale did not differentiate leaders from nonleaders. Middleton (1941) found that leaders scored low in extroversion. Bellingrath (1930) and Drake (1944) obtained no significant correlations between introversion-extroversion scores and leadership.
All the groups of great leaders except soldier-statesmen in Cox’s (1926) study were rated as introverted, with soldier-fighters rating very high in introversion. Thurs-tone’s (1944) study of administrators in Washington, D.C., revealed that successful administrators rated higher than less successful administrators in Guilford and Guilford’s (1939) T factor, which is measured by such items as “introspective, analyzes himself”; “often in a meditative state”; “analyzes the motives of others”; and “not more interested in athletics than in intellectual pursuits.” Brogden and Thomas (1943) added to this list such items as “he does not want anyone to be with him when he receives bad news”; “he does not try to find someone to cheer him up when in low spirits”; and “he prefers to make hurried decisions alone.” These items are of interest when considered in relation to the findings on mood control. In view of the diversity of findings, it appears doubtful that leaders can be described with any degree of uniformity in terms of introversion-extroversion.
Much the same situation exists with regard to self-sufficiency. Hunter and Jordan (1939) and Richardson and Hanawalt (1944) found that leaders had high self-sufficiency scores on the Bernreuter test, but Dunkerley (1940), Remmelin (1938), and Richardson and Hanawalt (1943) obtained no such significant results.
The evidence concerning the relationship of dominance to leadership is somewhat contradictory. Leaders were found to be more dominant and ascendant in 11 studies; potiential leaders were rejected if they were bossy and domineering in four studies; and no differences appeared in two studies.
Cox (1926) and Drake (1944) found “desire to impose will” to be associated with leadership, but Webb (1915) reported a zero-order correlation between those two factors. Ackerson (1942) reported a correlation of approximately .20 between bossiness and leadership in problem children. Leadership and bossiness were related, to some extent, in the children studied by Tryon (1939), who reported correlations of .28 and .29 between these two factors for 15-year-old boys and girls, respectively. Chapple and Donald (1946), Richardson and Hanawalt (1943, 1944), and Hunter and Jordan (1939) found leaders to be significantly more dominant than nonleaders. Small but positive differences in ascendance were reported by Bowden (1926) and Moore (1935). Eichler (1934), however, found that leaders and nonleaders did not differ in dominance. Still stronger contradictory evidence was presented by Broich (1929), Jennings (1943), and Hanfmann (1935), who concluded that bossy, domineering persons were rejected as leaders. Caldwell (1920) reported that high school pupils expressed preference for leaders who could keep order without being bossy. In all, these findings suggest that leadership cannot be defined in terms of personal dominance.
Initiative and a willingness to assume responsibility were related to leadership in 12 studies and persistence in the face of obstacles was related to leadership in 12 other inquiries. Ambition and desire to excel were of consequence to leadership in seven analyses, as were application and industry in six additional analyses.
Initiative and Assertiveness. All except one of the studies in which initiative was found to be a trait ascribed to leaders were investigations in which student leaders were nominated by their associates, and the traits that were thought to make them desirable as leaders were described. The study by Carlson and Harrel (1942) represented some departure from this method in that 53 Washington correspondents were asked to name the 10 ablest senators and the 10 ablest representatives in rank order and to rate them from 1 to 10 on integrity, intelligence, industry, and influence. A factor analysis of these ratings revealed Factor 1 to be heavily loaded with industry and influence and might also have been called push or assertiveness. Industriousness and leadership were correlated .55 and .16 in the studies of Bellingrath (1930) and Flemming (1935), Dunkerley’s (1940) factor analysis also revealed a trait cluster, identified as initiative, which was descriptive of intellectual and social leaders but not of religious leaders. Finally, Drake (1944) and Sheldon (1927) respectively reported correlations of .56 and .52 between aggressiveness and leadership.
Persistence. Cox (1926) found that great face-to-face leaders were characterized, to an outstanding degree, by “persistence in the face of obstacles,” “capacity to work with distant objects in view,” “degree of strength of will or perseverance,” and “tendency not to abandon tasks from mere changeability.” Pigors (1933) observed that the development of determination and a sufficient memory span to pursue remote goals rather than immediate objectives were necessary conditions for the appearance of leadership in children. The remainder of the studies that presented evidence on this point represented a variety of points of view.
Pinard (1932), in an experimental study of perseverance in 194 “difficult” children ages 8–15, found that of 24 leaders, 17 belonging to the “moderate nonperseverator” group were rated as more reliable, self-controlled, persistent, and as the most constructive leaders. Drake (1944) and Webb (1915) obtained correlations of .13 and .59 between leadership and strength of will. Webb (1915) reported a correlation of .70 between leadership and “persistence in overcoming obstacles” and of .53 between leadership and persistence. In Bellingrath’s (1930) study of high school students, persistence was correlated .68 with leadership. Eichler (1934) and Sheldon (1927) found correlations of .23 and .34 between leadership and persistence. An interesting sidelight was presented in Ackerson’s (1942) study of problem children: stubbornness correlated .15 for boys and .12 for girls with leadership.
Ambition. Cox (1926) also presented evidence to indicate that great face-to-face leaders, such as soldiers, religious leaders, and statesmen, were characterized to an outstanding degree by a “desire to excel” at performance. Hanawalt, Hamilton, and Morris (1934), in a study of 20 college leaders and 20 nonleaders, found that the level of aspiration of leaders was significantly higher than that of nonleaders. Correlations of .47 .29 and .64 between leadership and desire to excel were reported by Webb (1915), Drake (1944), and Bellingrath (1930), respectively.
Application and Industry. That leadership is related to willingness to work rather than to passive status or position is suggested by the fact that a number of investigators found leaders to rate high in application and industry. Cox (1926) observed that great leaders ranked unusually high in this respect. The correlations reported by Bellingrath (1930), Flemming (1935), and Webb (1915) ranged from .16 to .55.
All 17 studies found responsibility to be related to leadership. Thus student leaders were seen to rate somewhat higher than followers on dependability, trustworthiness, and reliability in carrying out responsibilities in the studies by Baldwin (1932), Bellingrath (1930), Burks (1938), Caldwell (1920), Dunkerley (1940), Moore (1932), Nutting (1923), Pinard (1932), and Wetzel (1932). Trustworthiness and leadership were correlated .64 in Webb’s (1915) study, .37 in Drake’s (1944) study, and .10 in Flemming’s (1935) study. Correlations of .42, .21, and .53 between conscientiousness and leadership were reported by Webb (1915), Drake (1944), and Bellingrath (1930), respectively. Partridge (1934) observed a correlation of .87 between dependability and leadership. Jennings (1943) observed that girls chosen as leaders tended to be those who inspired confidence. Cox (1926) found that all types of great face-to-face leaders rated high in trustworthiness and conscientiousness, with religious leaders rating outstandingly high in these traits. Additionally, Broich (1929), Jennings (1943), Leib (1928), Nutting (1923), and Pigors (1933) observed that leaders tend to be able to work for the group’s welfare, and Buttgereit (1932) noted that a sense of social responsibility is characteristic of leaders.
Integrity and fortitude were related to leadership in six studies, and strength of convictions was related to it in another seven analyses.
Integrity. Intellectual fortitude and integrity of character represent traits that are apparently associated with eminent leadership in maturity. All but one of the studies that contributed evidence on this point were concerned with outstanding adult leaders. Middleton (1941) obtained supporting evidence in college students.
Michels (1915) reported that strength of convictions was a characteristic of successful political leaders. Cox (1926) found that the great face-to-face leader was characterized to an outstanding degree by “absence of readiness to accept the sentiments of his associates.” This trait was especially conspicuous in revolutionary statesmen. Webb (1915) obtained a correlation of 2.32 between leadership and acceptance of the sentiments of others. Caldwell and Wellman (1926) noted that one characteristic of high school leaders was insistence on the acceptance of their ideas and plans.
Conviction. Adult leaders in a community studied by Chapin (1945) appeared to hold opinions that were generally similar to those of the group, but they “expressed the trends of opinion of the rank and file more sharply, more decisively, and more consistently.” Simpson (1938), in a study of those who influence and those who are influenced in discussions, found that influence scores correlated 2.41 with influenceability scores. It appears that persons in various types of groups may be valued as leaders because they know what they want to accomplish and are not likely to be swayed from their convictions.
The evidence on liberalism or conservatism suggested that the attitudes regarded as acceptable in leaders are largely determined by the nature of the situation. Hunter and Jordan (1939) found college student leaders to be somewhat more liberal than nonleaders in attitudes toward social questions. Newcomb (1943) reported that in a college where liberalism is a tradition and an ideal, women students who had the most prestige were regarded as most liberal. But Middleton (1941) ascertained that campus leaders were low in radicalism. In Thurstone’s (1944) study of Washington administrators, the Allport-Vernon Study of Values was found to be the most effective of a battery of 75 tests in differentiating higher-salaried from lower-salaried administrators. Successful administrators scored significantly higher in social and theoretical values and significantly lower in economic and religious values. Drake (1944) and Webb (1915) obtained low positive correlations between leadership and interest in religion.
Almost all authors reporting data on the relationship of self-confidence to leadership, were uniform in the positive direction of their findings. Self-assurance was associated with leadership in 11 studies; absence of modesty, in six studies. The general trend of these findings suggested that leaders rate higher than their followers in self-confidence and self-esteem and slightly lower in modesty. The following correlations were reported between self-confidence and leadership: .58 by Bellingrath (1930), .59 by Drake (1944), and, .12 by Webb (1915). Cowley (1931) found self-confidence to be one of six factors possessed in common by three widely different types of leaders. Cox (1926) noted that great leaders were characterized to an unusual degree by such traits as self-confidence, esteem of their special talents, and a tendency to rate their talents correctly. Buttgereit (1932), Moore (1932), and Zeleny (1939) also reported that leaders rated high in self-confidence. Tryon (1939) described student leaders as assured in class and as assured with adults. Richardson and Hanawalt (1943, 1944) found that college and adult leaders earned higher self-confidence scores on the Bern-reuter test than nonleaders; but Hunter and Jordan (1939) and Remmelin (1938) failed to find that the self-confidence scores on the Bernreuter test differentiated between leaders and nonleaders.
Inferiority. Sward (1933) found that inferiority scores on the Heidbreder rating scale did not differentiate leaders from nonleaders, although women leaders rated themselves higher in inferiority attitudes than did their associates. But Ackerson (1942) reported correlations of only 2.02 and .08 between feelings of inferiority and leadership in boys and girls.
Modesty. The findings here suggested that leaders tend not to be handicapped by excessive modesty. Cox (1926) reported that great military leaders and statesmen were characterized to a greater than average degree by eagerness for the admiration of the crowd and desire for the limelight, although they exhibited offensive manifestations of self-esteem to a lesser degree than the average. Middleton (1941) also found leaders to rate low in modesty. A correlation of .09 between leadership and modesty was reported by Flemming (1935). Eagerness for admiration was correlated .16 with leadership in Webb’s (1915) study. Drake (1944) obtained a correlation of 2.11 between conceit and leadership. Ackerson (1942) and Tryon (1939) found correlations between leadership and attention-getting and “showing off” ranging from .15 to .30.
In all six studies on the subject, leadership and a sense of humor were positively related. In four of six studies, leaders were controlled in mood and seldom gloomy; in two studies, leaders were happy and cheerful, but happiness was not a factor in two of the studies. The scarcity of evidence concerning the relation of mood control to leadership cannot be regarded as confirmation of its unimportance. The evidence suggests that mood control may be significantly related to effective leadership, and a sense of humor is certainly relevant. The topic appears to warrant thorough investigation.
Mood. Jennings (1943) stated that one characteristic of girl leaders in an institution was the ability to control their own moods so as not to impose their negative feelings, depression, and anxiety on others. Caldwell and Wellman (1926) and Malloy (1936) also found leaders to be constant in mood. Webb (1915) reported a correlation of 2.45 between depression and leadership. Ackerson (1942) and Cox (1926), however, reported some association between leadership and moods of depression, although not to a significant degree, and the extent differed with different groups.
Drake (1944), Tryon (1939), and Webb (1915) found that a cheerful, happy disposition was associated with leadership. These authors reported correlations ranging from .29 to .60 between leadership and cheerfulness. However, Ackerson (1942) did not find cheerfulness to be a distinguishing factor in leadership. Unhappiness and leadership were correlated 2.03 for boys and .06 for girls. Baldwin (1932) also found a lack of correlation.
Humor. Drake (1944), Flemming (1935), Tryon (1939), and Webb (1915) reported correlations ranging from .34 to .64 between leadership and sense of humor. Stray (1934) also found leaders to be characterized by a sense of humor. Goodenough’s (1930) finding of a correlation of .53 between leadership and laughter was also relevant.
Leaders were found to be more stable and emotionally controlled in 11 studies and less well controlled emotionally in five studies. No differences were found in three other studies.
Self-Control. A number of manuals that outline practical techniques for gaining friends and becoming a leader regard self-control as a very important prerequisite for attaining these goals. The evidence relating to this contention is divided. Eichler (1934) reported a correlation of .18 between leadership and self-control. Baldwin (1932), Pigors (1933), and Wetzel (1932) also found self-control to be a factor related to leadership. Bellingrath (1930) and Drake (1944) reported correlations of .70 to .38, respectively, between leadership and stability. Leaders were found by Middleton (1941) and Terman (1904) to rate low in emotionality. Bowden (1926) and Caldwell and Wellman (1926) found leaders to be well balanced and self-composed in comparison with their followers. Webb (1915) reported correlations of 2.25 between irritability and leadership, and 2.36 between readiness for anger and leadership.
Excitability. Despite the results above, Cox (1926) found that great face to-face leaders rated high in excitability. This trait was present to an unusual degree in revolutionary statesmen. In problem children, Ackerson (1942) reported correlations of .12 for boys and .36 for girls between irritability and leadership. A correlation of .16 between leadership and excitability was found by Sheldon (1927). Fauquier and Gilchrist (1942) also noted that leaders were more excitable than nonleaders. According to Chevaleva-Ianovskaia and Sylla (1929), leaders were characterized by a predominance of excitation over inhibition. But Zeleny (1939) could find no difference between leaders and nonleaders in degree of emotional control, and Drake (1944) and Flemming (1935) reported correlations close to zero between leadership and excitability.
Anger. The data on the relationship between leadership, anger, and fighting cast further light on this subject. Cox (1926) found that great face-to-face leaders, except statesmen, were characterized by a tendency to anger and “a tendency to flare up on slight provocation.” Ackerson (1942) reported that in children “temper tantrums” and “leader” were positively correlated, but “temper tantrums” and “follower” were negatively correlated. Webb (1915), however, found a correlation of 2.12 between leadership and occasional extreme anger. Tryon (1939) obtained correlations of .59, .48, .25, and .40 between fighting and leadership for 12-year-old boys, 15-year-old boys, 12-year-old girls, and 15-year-old girls, respectively. Ackerson (1942) ascertained that fighting and leadership were correlated .13 for boys and 2.17 for girls. Fighting and leading others into bad conduct were correlated .20 for boys and .36 for girls. Incorrigibility and defiance were also positively correlated with leadership, and to a still higher degree with leadership in misconduct, while these traits were correlated negatively among followers.
These studies did not lend convincing support to the view that leaders are necessarily characterized by a high degree of self-control or an incapacity for emotional expression.
In 15 studies the leaders came from higher socioeconomic backgrounds, but in two studies no differences were found. Baldwin (1932) and Goodenough (1930) reported negligible differences. The differences in the social and economic status of leaders and nonleaders were usually not extreme. Only Remmelin (1938) obtained differences that were large enough to be statistically reliable. Nonetheless, taken as a whole, the evidence presented in studies from a wide variety of leadership situations indicated that leaders tend to come from a socioeconomic background superior to that of the average of their followers.
Leaders participated in more group activities in all 20 studies on the subject. They also exhibited a higher rate of social mobility in five additional studies.
Participation. Baldwin (1932), Brown (1933), Chapin (1945), Courtenay (1938), Richardson and Hanawalt (1943), Roslow (1940), Link (1944), Merriam and Gos-nell (1929), Reals (1938), Smith and Nystrom (1937), Sorokin (1927), and Zeleny (1939) all found that leaders surpassed followers in the number, extent, and variety of group activities in which they participated. Zeleny (1939) reported correlations ranging from .17 to .68 between leadership and participation in extracurricular activities. Leadership was defined by a number of authors as “occupying one or more positions of responsibility in group activities.” On the other hand, social detachment appeared to be a factor in the formation of the boys’ gangs studied by Thrasher (1927) and Warner (1923).
Mobility. Physical and social mobility were observed by Sorokin (1927), Sorokin and Zimmerman (1928), and Winston (1932) to be associated with adult leadership. Sorokin and Zimmerman reported that farmer leaders were characterized to a high degree by a tendency to shift from place to place and from one occupational or economic position to another. Winston (1937) observed the same tendency in inventors.
Sociophysical Activity. Sociophysical activities were related to leadership as follows: “active in games” (six studies); “active, restless” (nine studies); “daring, adventurous” (three studies).2 Broich (1929), Brown (1931), Buttgereit (1932), and Reininger (1929) found that child leaders were more active in games than were nonleaders. In Tryon’s (1939) study, leadership and “active in games” were correlated .52 to .74 for groups of 12-and 15-year-old boys and girls. Terman (1904), Thrasher (1927), and Tryon (1939) found leaders to be more daring and adventurous than followers. Correlations of .57 to .78 between daringness and leadership were reported by Tryon (1939). Cowley (1931) ascertained that motor impulsion was a factor common to different types of leaders. Liveliness was reported by Leib (1928) and Brown (1931) as characterizing leaders. Flemming (1935) found a correlation of .47 between leadership and liveliness. Goodenough (1930) reported a correlation of .29 between physical activity and leadership. Ackerson (1942) and Tryon (1939) obtained correlations close to .20 between “restlessness” and leadership. These findings suggested that physical activity, mobility, and sociophysical activity were associated with leadership.
Sociability. Sociability was associated with leadership in 13 of 14 studies, and diplomacy or tact was associated with leadership in 8 others. Fairly high positive correlations between sociability and leadership were reported by Bonney (1943), Drake (1944), Flemming (1935), Good-enough (1930), Sheldon (1927), Tryon (1939), and Webb (1915). Correlations ranged from .33 to .98. But the correlation found by Eichler (1934) between social intelligence and leadership was only .10. Burks (1938), Malloy (1936), Middleton (1941), and Prosh (1928) also found that student leaders rated higher than nonleaders in sociability. Ackerson (1942) observed that belonging to a gang was correlated .26 with being a leader and .21 with being a follower. Being a leader and being within an intimate circle were correlated .39 in Webb’s (1915) study. Moore (1932) and Newcomb (1945) reported friendliness and social skills, respectively, as factors that distinguished leaders from followers. Cox (1926) also noted that, despite their higher introversion, great leaders were rated above average, but not to an outstanding degree, in fondness for companionship and social gatherings.
Tact. Courtesy, tact, and diplomacy were found by Bernard (1928), Wetzel (1932), Drake (1944), Flemming (1935), Hanfmann (1935), Parten (1933), Stray (1934), and Webb (1915) to be traits that distinguished leaders from nonleaders. Drake, Flemming, and Webb reported correlations of .08, .27, and .73, respectively, between tact and leadership. However, Flemming (1935) obtained a correlation of only 2.03 between rudeness and leadership for boys and girls. But the correlations between rudeness and leading others into bad conduct were .24 and .40 for boys and girls, respectively. Ackerson determined that both bashfulness and seclusiveness were negatively correlated with leadership.
Misconduct. Ackerson (1942), Goodenough (1930), and Webb (1915) obtained correlations ranging from 2.29 to .21 between offensive manifestations and leadership. Ackerson’s (1942) findings suggested that misconduct is not necessarily a bar to leadership. Stealing, for example, was correlated .12 and .21 with leadership; stealing and leading others into misconduct were correlated .46 and .16 for boys and girls, respectively.
Evidence from 10 diverse studies indicated that leaders tend to be rated higher than average in popularity. Evidence presented by Ackerson (1942), Bellingrath (1930), Carlson and Harrell (1942), Cox (1926), Garrison (1933), Michels (1915), Miller and Dollard (1941), Nutting (1923), Tryon (1939), and Zeleny (1939) all indicated that popularity and prestige were rather closely associated with leadership status. The correlations, ranging from .23 and .80, suggested that the relationship between popularity and leadership was fairly high. However, Nutting (1923) pointed out that popularity cannot be regarded as synonymous with leadership.
Cooperativeness was related to leadership in 11 studies. Work for the group and corporate responsibility were related to leadership in eight others. The ability to enlist cooperation was related to leadership in seven additional analyses.
Leaders were found by Baldwin (1932), Dunkerley (1940), Fauquier and Gilchrist (1942), Newcomb (1943), and Wetzel (1932) to rate higher than followers in cooperativeness. Drake (1944) and Webb (1915) reported correlations of .44 and .69 between cooperativeness and leadership. The ability to enlist cooperation and to control others in a group enterprise were found by Baldwin (1932), Caldwell (1920), Hanfmann (1935), Merriam and Gosnell (1926), and Nutting (1923) to be characteristics associated with leadership ability. Webb (1915) reported a correlation of .69 between leadership and corporate spirit. Cox (1926) also reported that great leaders rate outstandingly high in a sense of corporate spirit.
There was a preponderance of evidence from a wide variety of studies (19 in all) that indicated that patterns of leadership traits differed with the situation. Ackerson’s (1942) study revealed marked differences in the conduct and personality patterns of children who were regarded as leaders in general and children who were regarded as leaders in misconduct. Boys and girls in these two groups also differed somewhat. Bellingrath (1930) found marked differences in the extent to which leaders in athletics, student government, publications, and clubs participated in extracurricular activities and were chosen as leaders under various circumstances. The investigation by Caldwell and Wellman (1926) revealed athletic leaders to be tallest among the leaders and to excel in physical achievements, while editors were younger and shorter than average but ranked higher in scholarship than the other groups of leaders who were studied. Cowley’s (1928) study demonstrated large differences in the traits of criminal leaders, army leaders, and student leaders. The profiles of the average ratings of the traits of groups of great leaders studied by Cox (1926) differed markedly from one group to another, especially in physical and emotional traits, but much less so in traits that may be classified as intelligence, self-regard, and persistence. Dunkerley’s (1940) factor analysis of the intercorrelations of 15 variables representing trait ratings of 167 women college students revealed a factor identified as social leadership and two factors identified as religious leadership.
Hanfmann (1935) observed three types of leaders among preschool children: (1) objective leaders who engage in constructive play and get what they want by saying why they need it; (2) social leaders, whose goal is to play with others rather than play for the sake of play itself; and (3) gangsters, who get their way by force and a complete disregard for others. Schuler (1935) concluded that although teachers may ascertain with increasing reliability the dominant-submissive behavior of older adolescent boys in one situation, such as the school, it becomes less possible to predict those tendencies in another environment, such as the home.
Sward (1933) found that superior socioeconomic status, as well as higher intelligence and scholastic attainment, differentiated 125 campus leaders from 125 followers. However a classification of the leaders into subgroups demonstrated the following distinguishing differences: (1) bright, relatively unmotivated, unsociable, self-confident campus editors; (2) rather insecure, intellectual, and very intelligent debaters; (3) strongly socialized and intellectually mediocre campus politicians; and (4) extroverted women leaders.
Terman (1904) found that children who were leaders in one experimental situation may not have been leaders when matched against different children in other situations. Children who were “automatons,” or nonleaders, in most situations might achieve leadership in some situations. Children who were leaders in most situations were said by their teachers to be characterized by intelligence, congeniality, liveliness, and goodness.
In Tryon’s (1939) study, the clusters of traits that characterized boys and girls at age 12 differed from those found at age 15. This difference was especially noticeable for girls, who matured somewhat more rapidly than boys in social interests. The leadership cluster for 12-year-old boys was composed of the items: daring, leader, active in games, and friendly. The cluster for 15-year-old boys contained the items: daring, leader, active in games and fights. The leadership cluster for 12-year-old girls contained the items: daring, leader, and humor about jokes. The cluster for 15-year-old girls contained the following items: popular, friendly, enthusiastic, happy, humor about jokes, daring, leader. The total weight of the evidence presented in this group of studies suggests that if there were general traits that characterized leaders, nonetheless the patterns of such traits were likely to vary with the leadership requirements of different situations.
Six follow-up studies, although yielding somewhat variable results, intimated a certain degree of persistence or transferability of leadership. Levi (1930) studied 230 leaders in elementary and junior high school, 206 of whom were studied again in senior high school. The correlation between leadership in elementary school and leadership in senior high school was .19; the correlation between leadership in junior high school and leadership in senior high school was .52. For athletic leadership there was a low negative correlation between elementary school and high school, but a correlation of .44 between junior high school and senior high school.
Kohs and Irle (1920) completed a follow-up study of the military careers of 116 college students. Three faculty members rated these students on various traits. Correlations between the U.S. Army rank attained and various ratings in college ranged from .11 to .39. The best assessments for predicting military success were found to be the raters’ estimates of the assessees’ potential value to the service and raters estimates of the assessees’ intelligence. Assessments of leadership in college were correlated .11 with army rank attained, but scholarship was not predictive of army rank. Page (1935), studying cadets at West Point, found first-year leadership rank to be correlated .67 with fourth-year leadership rank. Rank in bearing and appearance was most highly correlated with rank in leadership, while the ranks in athletic activities, tactics, and academic standing were correlated with leadership rank to progressively lesser degrees.
Clem and Dodge (1933) conducted a comparative study of the success of 27 student leaders, 36 high-ranking student scholars, and 38 students selected at random after graduation from high school. High school student leaders ranked highest in subsequent outstanding achievements, number of honors received, and quantity of publications. The random group ranked highest in community leadership and the amount of money accumulated after graduation. In general, the student leaders tended to become more successful than the student scholars and the random group, although the differences were not impressive. Courtenay (1938) studied 100 women leaders and 100 nonleaders from 13 successive high school graduating classes. The two groups were matched in socioeconomic background, ethnic heritage, scholarship, and age at graduation. Courtenay found that 72 student leaders, but only 29 nonleaders, went to college and that twice as many high school leaders as non-leaders were engaged in professional work as adults. The average salary of the adults who had been high school leaders exceeded that of those who had been nonleaders. The high school leaders were more active as adults in community work. Shannon (1929) compared student leaders, scholars (honor-roll members), and a random group from five high school graduating classes. Although the honor graduates were little more successful than the random group, Shannon concluded that “whatever is required to excel in the extracurricular life of the high school seems to be the same thing that contributes most to success later.”
These findings strongly suggested that leadership in school activities was somewhat predictive of later success. However, the extent to which leadership persisted and transferred was not clearly determined.
1. The following conclusions were supported by uniformly positive evidence from 15 or more of the studies surveyed: (a) The average person who occupies a position of leadership exceeds the average member of his or her group in the following respects: intelligence, scholarship, dependability in exercising responsibility, activity and social participation, and socioeconomic status. (b) The qualities, characteristics, and skills required in a leader are determined, to a large extent, by the demands of the situation, in which he or she is to function as a leader.
2. The following conclusions were supported by uniformly positive evidence from ten or more of the studies surveyed: (a) The average person who occupied a position of leadership exceeded the average member of his or her group, to some degree, in the following respects: sociability, initiative, persistence, knowing how to get things done, self-confidence, alertness to and insight into situations, cooperativeness, popularity, adaptability, and verbal facility.
3. A number of factors were found to be specific to well-defined groups. For example, athletic ability and physical prowess were found to be characteristics of leaders of boys’ gangs and play groups. Intellectual fortitude and integrity were found to be associated with eminent leadership in maturity.
4. The traits with the highest overall correlation with leadership were originality, popularity, sociability, judgment, assertiveness, desire to excel, humor, cooperativeness, liveliness, and athletic ability, in approximate order of magnitude of the average correlation.
5. Despite considerable negative evidence, the general trend of the results suggested a low positive correlation between leadership and such variables as chronological age, height, weight, physique, energy, appearance, dominance, and mood control. The evidence was about evenly divided concerning the relation to leadership of such traits as introversion-extroversion, self-sufficiency, and emotional control.
6. The evidence suggested that leadership exhibited in various school situations may persist into college and later vocational and community life. However, knowledge of the facts related to the transferability of leadership remains meager and obscure.
7. For understanding leadership, the most fruitful studies were those in which the behavior of leaders was described and analyzed on the basis of direct observation or the analysis of biographical and case history data. The factors associated with leadership could all be classified under the following general headings: Capacity (intelligence, alertness, verbal facility, originality, and judgment); Achievement (scholarship, knowledge, and athletic accomplishments); Responsibility (dependability, initiative, persistence, aggressiveness, self-confidence, and the desire to excel); Participation (activity, sociability, cooperation, adaptability, and humor); Status (socioeconomic position and popularity); and Situation (mental level, status, skills, needs and interests of followers, objectives to be achieved, and so on).
It is primarily by participating in group activities and demonstrating a capacity for expediting the work of the group that a person becomes endowed as a leader. A number of investigators were careful to distinguish between the leader and the figurehead and to point out that leadership is always associated with the attainment of group objectives. Leadership implies activity, movement, and getting work done. The leader is a person who occupies a position of responsibility in coordinating the activities of the members of the group in their task of attaining a common goal. This definition leads to a consideration of another significant factor. A person does not become a leader by virtue of some combination of traits; but the pattern of personal characteristics of the leader must bear some relevant relationship to the characteristics, activities, and goals of the followers. Thus leadership must be conceived in terms of the interaction of variables that are in constant flux. The factor of change is especially characteristic of the situation, which may be radically altered by the addition or loss of members, changes in interpersonal relationships and in goals, the competition of external influences, and the like. The personal characteristics of the leader and of the followers are, in comparison, highly stable. The persistence of individual patterns of human behavior in the face of continual situational change appears to be a primary obstacle not only to the practice of leaders but to their selection and placement. It is not especially difficult to find persons who are leaders. It is quite another matter to place these persons in different situations where they will be able to function as leaders. It becomes clear that an adequate analysis of leadership involves a study not only of leaders but also of situations.
The evidence suggests that leadership is a relationship between persons in a social situation and that persons who are leaders in one situation may not necessarily be leaders in other situations. Must it then be assumed that leadership is entirely incidental, haphazard, and unpredictable? Not at all. The very studies that provided the strongest arguments for the situational nature of leadership also supplied the strongest evidence to indicate that leadership patterns as well as nonleadership patterns of behavior were persistent and relatively stable. Jennings (1943, p. 210) observed that “the individual’s choice behavior, in contrast to his social expansiveness, appears as an expression of needs which are, so to speak, so ‘central’ to his personality that he must fulfill them whether or not the possibility of fulfilling them is at hand or not.” A somewhat similar observation was made by Newstetter, Feldstein, and Newcomb (1938, p. 92): “Being accepted or rejected is not determined by the cordiality or antagonism of the individual’s treatment of his fellows, nor evidently, is the individual’s treatment of his fellows much affected by the degree to which he is already being accepted or rejected by them. Their treatment of him is related to their acceptance or rejection of him. Their treatment of him is, of course, a reaction to some or all of his behaviors, but we have been completely unsuccessful in attempting to measure what these behaviors are.”
The authors concluded that these findings provided “devastating evidence” against the concept of the operation of measurable traits in determining social interactions. Although these findings do not appear to provide direct evidence either for or against a theory of traits, they do indicate that the complex of factors that determines an individual’s status in a group is most difficult to isolate and evaluate.
The findings of Jennings and Newsletter, Feldstein, and Newcomb suggested that selecting leaders should be much less difficult than training nonleaders to become leaders. The clinician or group worker who has observed the fruitless efforts of socially isolated individuals to gain acceptance or leadership status in a group is aware of the real nature of the phenomena just described. Some individuals are isolated in almost any group in which they find themselves, while others are readily accepted in most of their social contacts. A most pertinent observation on this point was made by Ackerson (1942, p. 45), who noted that “the correlations for ‘leader’ and ‘follower’ are not of opposite sign and similar magnitude as would be expected of traits supposed to be antithetical. These may not be the opposite poles of a single underlying trait.” Ackerson went on: “It may be that the true antithesis of ‘leader’ is not ‘follower,’ but ‘indifference,’ i.e., the incapacity or unwillingness either to lead or to follow. Thus it may be that some individuals who under one situation are leaders may under other conditions take the role of follower, while the true ‘opposite’ is represented by the child who neither leads nor follows.”
The findings suggest that leadership is not a matter of passive status or of the mere possession of some combination of traits. Rather, leadership appears to be a working relationship among members of a group, in which the leader acquires status through active participation and demonstration of his or her capacity to carry cooperative tasks to completion. Significant aspects of this capacity for organizing and expediting cooperative efforts appear to be intelligence, alertness to the needs and motives of others, and insight into situations, further reinforced by such habits as responsibility, initiative, persistence, and self-confidence.
But the studies surveyed offered little information as to the basic nature of these personal qualifications. Cattell’s (1946) analysis suggested that these qualifications could be based, to some degree, on basic intelligence, but Cat-tell and others also implied that these personal qualifications were, to a high degree, socially conditioned. The problems requiring thorough investigation relate to factors that condition social participation, insight into situations, mood control, responsibility, and the transferability of leadership from one situation to another. Solutions to these problems seem basic not only to any adequate understanding of the personal qualifications of leaders, but to any effective training for leadership.
Many new methods and measurements were introduced into the study of leadership in the decades after 1948. Experiments involving one variable at a time gave way to factorial and multivariate designs in which the effects of various treatments could be analyzed in the same experiment. Theory began to guide much of the data collection. Questionnaire methodologists introduced a variety of techniques to reduce errors such as the halo effect, leniency, and social desirability, and to increase the relevance and reliability of results, although their efforts often met with limited success. The critical-incidents technique, forced-choice checklists, behaviorally-anchored rating scales, and semantic differentials were just a few of the specific new methods used. Factor analysis became the basic tool in the search for and verification of the existence of traits of consequence. Other multivariate regression procedures also became commonplace in efforts to establish the relative importance of different traits to successful leadership. Varieties of statistical tests increased, to support or reject findings.
The internationalization of efforts also became widespread. Whether the same traits of leadership were relevant to rural agricultural leaders in Chile and Mali or led to promotion to higher management in Norway, Italy, and Japan were among the subjects examined. In the United States, a topic of considerable interest was whether different leadership traits would emerge as important for women and for racial and ethnic-minority leaders. The human-potential movement sparked awareness of the need to deal with leadership at a level of socioemotional feeling that was deeper than surface intellectual perception. More studies focused such traits as self-exposure, empathy, energy level, intuition, and interpersonal competence. The whole field of small-group research expanded, with many investigations in the 1950s. At the same time, much more rigor was introduced into measurements of individual role taking and behavior in small-group interactions. In addition, experimenters became much more aware of the many threats to the validity of their findings.
As the era ended, situational leadership was the dominant theory used for management and leadership training, but some disquieting research that supported the primacy of individual differences in leaders appeared shortly afterward. There was a resurgence of interest in consistent individual differences across situations, braced with meta-analytic demonstrations of the validity and generalizability of a very limited battery of cognitive abilities tests for predicting successful performance in a wide variety of situations (Schmidt & Hunter, 1977). Reexamination of earlier data and meta-analytic theory itself suggested that situational variations frequently could and should be attributed not to substantive effects but to sampling error. Situational effects would have to be shown above and beyond what would be generated by the normal probability distribution of the means generated in the diverse situations. Strong inferences were drawn about personality and early developmental influences that had permanent effects on individuals and their behavior as leaders and followers.
Stogdill (1970) examined another 163 studies of the traits of leadership published between 1948 and 1970 and considered how findings could be meaningfully factor-analyzed and clustered. (These findings were first reported in full in Chapter 5 of the first edition of the Handbook of Leadership in 1974.) Stodgill summarized the physical characteristics, social background, intelligence and ability, personality, task-related characteristics, and social characteristics related to leadership. Between 1948 and 1970, the proportion of studies of adults in formal organizations increased and the proportion dealing with children and adolescents in informal groups decreased. In interpreting conclusions caution is urged on the basis of the published survey findings.3
Measures of physical characteristics, such as activity, age, and appearance were of positive consequence to leadership. Appearance can affect first impressions, which in turn have significance for emergent leadership. Fewer than three studies were found for height and weight; nevertheless, height or weight above the average of the peer group is certainly not a disadvantage in achieving leadership status. (Many organizations like to be represented by impressive physical specimens. When faced with taller opponents in television debates, shorter presidential candidates such as Jimmy Carter and Michael Dukakis stood on raised platforms. Frederick the Great required that all his soldiers be tall; however, Napoleon is often cited as an example that a man of small stature can rise to a position of great power. Thus physical stature may complement a leader, or a leader may compensate for lack of physical stature. Yet there is a rational element involved when coercive leadership is needed. Robert Peel, who introduced the “bobbies” to London, made a highly effective police force without firearms by choosing only large men who could dominate scenes of social conflict. Conversely, it has been noted that smaller policemen are more likely to suffer attack and injury.)
Activity, Energy, Stamina. The 1970 survey found 25 studies4 of these factors—many more than were found in the 1948 survey—which suggested that a leader tends to be endowed with an abundant reserve of energy, stamina, and ability to maintain a high rate of physical activity. Even when handicapped by physical disability or poor health, highly successful leaders tended to exhibit high energy.
Age. Age appeared in six studies reported between 1948 and 1970.5 Age continued to be related to leadership in a complicated way, as noted before. A survey by Lehman (1953) on the relation of age to achievement in science, art, politics, and other fields found that great men tended to exhibit signs of outstanding accomplishment at a relatively early age. Many, but not all, had the advantage of specialized education or training because of the early recognition of their talent. However, it usually takes time to rise to the top in a corporate or governmental structure. Standard and Poor’s (1967) reported that 74 percent of 66,336 American executives in its 1967 Register of Corporations, Directors, and Executives were over age 50. Only 168 executives were under age 30, while 8,085 were in the 71-to-80 age group. The two sets of findings indicate that the creative individual is likely to exhibit evidences of his or her ability at an early age; however, large organizations up to 1970 did not make much use of such creative gifts in administrative capacities. Rather, organizations tended to rely on administrative knowledge and demonstrations of success that come with experience and age.
Social Status. Studies of the socioeconomic background of leaders continued to proliferate between 1948 and 1970; 19 studies were found for this period.6 D. R. Mathews (1954) observed that from 1789 to 1934, 58 percent of the presidents, vice presidents, and cabinet members had fathers in the professional, proprietor, or official occupations; 38 percent had fathers who were farmers, but only 4 percent had fathers who were wage earners. It is apparent that high social status has provided an advantage in rising to high levels of political leadership. But Newcomer (1955) and Scientific American (1965) reported that, compared with 1900, more top executives in 1965 were beginning to come from the poorer and middle-income groups, rather than the wealthy strata of society. A. Porter (1965) found that the background of the father of an executive was significantly related to the executive’s level in the organization and his authority for making policy. However, the father’s background was not related to the size of the organization or to the executive’s status in the business world or satisfaction with the progress of his career.
Miller and Dirksen (1965) reported that highly visible community leaders were differentiated from their less visible peers by being business oriented, Republican, members of the chamber of commerce, and named in the mass media. However, the hidden community leaders were characterized as holding administrative or professional jobs; they were not owners of large businesses, nor natives of the city, nor did they come from families in the city that were prominent. R. M. Powell’s (1969) large-scale survey of the executive promotion process indicated that religious and ethnic background—usually linked with social status—were also important factors.
Studies of the social background of student leaders by Martin, Gross, and Darley (1952); Weinberg (1965); Krumboltz, Christal, and Ward (1959); and Kumar (1966) revealed few consistent relationships across samples, although Williamson (1948) found that fraternity members occupied a disproportionately large share of leadership positions on the college campus.
Mobility. Six studies dealt with the upward mobility of leaders.7 Jennings’s (1967a) study is illustrative of the work in this area; it presented an insightful analysis of the problems, stresses, and adaptations involved in rapid upward mobility in a large corporation. More specifically, Cussler (1958) found that once women executives reached middle management in an industrial organization, they found it difficult to rise higher. They hit the “glass ceiling.”
Education. Fourteen studies showed the importance of education to leadership.8 Reflecting a national rise in educational levels, senior managers with college degrees increased from 28.3 percent in 1900 to 74.3 percent in 1964 (Scientific American, 1965). In contrast, G. F. Lewis’s (1960) review of several studies indicated that small businessmen have less education than top executives in large firms and more often start their careers as unskilled or semiskilled workers. Johnson, Peterson, and Kahler (1968) studied 496 first-line foremen in a company from 1940 to 1961. They found that the average age of these foremen increased from 31.2 years in 1940–1944 to 41.2 years in 1955–1959 and that their years of schooling increased from 10.8 to 11.2 during the same period.
The following conclusions were drawn from these studies of social background and education: (1) High so cioeconomic status was an advantage in attaining leadership; (2) Leaders who rose to high positions in industry tended to come from a lower socioeconomic status than their counterparts of a half century earlier; (3) The leaders tended to be better educated than formerly. The rise in the general level of education of the population is common knowledge. Requirements for managerial and administrative positions increasingly demanded a graduate degree, such as an MBA. The trend toward reduced emphasis on social status and more emphasis on education was expected to accelerate as the effects of affirmative action manifest themselves. As firms and agencies aggressively promoted women, blacks, and other minorities, a considerable increase in the upward mobility of these groups was expected. For example, in 1960 hardly any women and blacks were observed in MBA on engineering programs.
Intelligence. In 1970, 25 reports of a positive relationship between leadership, intelligence, and ability were found to have been published between 1948 and 1970,9 compared with 17 studies in the 1948 review correlating scores on intelligence tests with leadership status. The average correlation of .28 in the 1948 review was corroborated in the 1970 survey. However, five of the competent studies of 1948 ascertained that a large discrepancy between the intelligence of potential leaders and that of their followers militated against the exercise of leadership. Ghiselli (1963b, p. 898) reported supporting evidence. In a study of three groups of managers, he found that “the relationship between intelligence and managerial success is curvilinear with those individuals earning both low and very high scores being less likely to achieve success in management positions than those with scores at intermediate levels.”
Thus leaders can be too able for those they lead. Persons with greater abilities may suffer from extreme self-preoccupation; their abilities may make it difficult for them to communicate with those they are attempting to lead; and their ideas may be too advanced to be accepted by their potential followers (pioneers are seldom outstanding leaders). The discrepancy in abilities is likely to be paralleled by discrepancies in interests and goals. Also, Korman’s (1968) extensive review on the prediction of managerial performance reported that “intelligence, as measured by verbal ability tests, is a fair predictor of first-line supervisory performance, but not of higher level managerial performance.”10 But one must also suppose that only those who already possess above-average intelligence are likely to have achieved top management positions in the organization. So there is a restriction in range, which makes it impossible for intelligence tests to discriminate the good from the bad performers at the top of the organization. Nevertheless, it should be clear that a high-level intelligence test that discriminates verbal intelligence among those at the upper end of the population’s intelligence, such as the Miller Analogies (used for predicting success in graduate and professional schools), is also likely to be a valid predictor of the potential to rise in firms, agencies, and institutions.
Other Intellectual Abilities. Uniformly positive findings for studies completed between 1948 and 1970 were found, which indicated that leaders are characterized by superior judgment, decisiveness, or both (e.g., Roadman, 1964), knowledge (e.g., Colyer, 1951), and fluency of speech (e.g., Burnett, 1951b).
Up to 31 studies in the 1970 survey dealt with personality and leadership. Several differences were noted between the 1948 and 1970 surveys, which might have been due to changes in personality methods and theories, as well as the change in participants in the studies. Uniformly positive findings found in the 1970 survey were the traits of adjustment (e.g., Terrell & Shreffler, 1958), aggressiveness or assertiveness (e.g., Hobert & Dunnette, 1967), independence (e.g., Hornaday & Bunker, 1970), objectivity (e.g., Argyris, 1953), enthusiasm (e.g., Gibb, 1949), and tolerance of stress (e.g., Lange & Jacobs, 1960). Characteristics that appeared with positive findings in both 1948 and 1970 were alertness (e.g., Porter, 1959), originality (e.g., Randle, 1956), personal integrity (e.g., Stephenson, 1959), self-confidence (e.g., Moment & Zaleznik, 1963), ascendance (e.g., Sanders, 1968), emotional balance (e.g., Harville, 1969), and extroversion (e.g., Harrell, 1966).
Task-Related Personality Characteristics. Both the 1948 review and the 1970 review produced uniformly positive results indicating that leaders are characterized by a need for achievement (e.g., Cummings & Scott, 1965) and a sense of responsibility (e.g., Gordon, 1952). Leaders tend to be task-oriented (e.g., Medow & Zander, 1965) and dependable in the pursuit of objectives (e.g., Powell & Nelson, 1969). They display enterprise and initiative (e.g., Helfrich & Schwirian, 1968) and are persistent in overcoming obstacles.
Social Personality Characteristics. The positive findings on social characteristics in both 1948 and 1970 suggested that leaders were active participants in various activities. They interacted easily with a wide range of personalities (e.g., Krumboltz, Christal, & Ward, 1959), and this interaction was valued by others. They not only were cooperative with others (e.g., J. S. Guilford, 1952) but were able to enlist cooperation (e.g., Bentz, 1964) and to execute (administer) projects (e.g., Kay, 1959). Interpersonal skills (e.g., D. S. Brown, 1964), including tact (e.g., Tarnapol, 1958), made them attractive to followers (e.g., Price, 1948). Leaders were valued by group members because they had characteristics such as nurturance (e.g., Roff, 1950) and popularity (e.g., Harrell & Lee, 1964) that fostered loyalty and cohesiveness in the group.
Stogdill’s (1974) review of factorial studies of leadership published between 1945 and 1970 was based on 52 factorial studies, including surveys of a large number of military and industrial personnel, studies of leadership in military and industrial groups, and reports on experimental groups. It should be emphasized that the factors which emerged depended on the variables originally included in the battery of traits measured or ignored by the investigators. For instance, if an investigator included many measures of social distance, a common factor of social distance could emerge. If no measures of social distance were included, no common factor of social distance could be extracted. Likewise, if only a single reliable measure of social distance was included, it would not appear in a separate common factor of social distance, as such. The most frequently occurring factors were descriptive of various skills of the leader.
They included social and interpersonal skills, technical skills, administrative skills, intellectual skills, leaders’ effectiveness and achievement, social nearness, friendliness, supportiveness of the group task, and task motivation and application. These factors indicated that leaders differed from each other consistently in the effective use they made of interpersonal, administrative, technical, and intellectual skills. Some leaders could be described as highly task-motivated; others were most capable of maintaining close, friendly, personal relationships. The best leaders were able to do both. The next most frequent set of factors was concerned with how leaders relate to their groups. The behaviors included maintaining the cohesiveness of the group, coordination, task motivation, task performance, and high quality of output. A concern for the group’s performance was softened by nurturant behavior and the use of informal controls. These factors were as follows: maintaining a cohesive work group, maintaining coordination and teamwork, maintaining standards of performance, maintaining informal control of the group (group freedom), and nurturant behavior. Next in frequency were factors concerned strictly with the personal characteristics of leaders. Leaders could be described in terms of emotional balance, willingness to assume responsibility, ethical conduct, ability to communicate readily, dominance, energy, experience, courage, and maturity.
In sum, successful leadership involves certain skills and capabilities—interpersonal, technical, administrative, and intellectual—that enable leaders to be of value to their group or organization. These skills allow leaders to maintain satisfactory levels of group cohesiveness, drive, and productivity. Leaders are further assisted in the execution of these functions if they possess a high degree of motivation to complete tasks, personal integrity, communicative ability, and the like. The 52 factorial studies as a whole seem to provide a well-balanced picture of the skills, functions, and personal characteristics of leaders in a wide variety of situations. The factors and the number of studies in which they emerged were as follows: ascendance, dominance, decisiveness (11); willingness to assume responsibility (10); ethical conduct, personal integrity (10); maintaining a cohesive work group (9); maintaining coordination and teamwork (7); ability to communicate, articulativeness (6); physical energy (6); maintaining standards of performance (5); creative, independent (5); conforming (5); courageous, daring (4); experience and activity (4); nurturant behavior (4); maintaining informal control of the group (4); mature, cultivated (3); and aloof, distant (3).
The differences between 1948 and 1970 may be due primarily to the larger percentage of studies in the 1970 survey from the world of work rather than from children’s, school, and social groups.11 The similar results made it reasonable to conclude that many cognitive, social, and emotional traits differentiated leaders from followers, successful from unsuccessful leaders, and high-level from low-level leaders. One practical application of this conclusion was the assessment center for determining leadership potential among candidates for managerial positions (Bray, Campbell, & Grant, 1974; Bray & Grant, 1966). By the 1970s, over 1,000 such assessment centers were in operation. In these centers candidates are observed for two to three days in interviews, leaderless group activities, and other situational tests. They are also tested individually with personality and aptitude tests. The “in-basket,” a sampling of managerial action requirements, is also often used. Observers meet to try to pool their results on the basis of inferences from the test results and their observations to yield a picture of the total personality of the candidate and his or her leadership potential in positions familiar to the observers. Much more will be presented about such assessment centers in Chapter 35.
The leader is characterized by a strong drive for responsibility and completion of tasks, vigor and persistence in the pursuit of goals, venturesomeness and originality in problem solving, a drive to exercise initiative in social situations, self-confidence and a sense of personal identity, willingness to accept the consequences of his or her decisions and actions, readiness to absorb interpersonal stress, willingness to tolerate frustration and delay, ability to influence other people’s behavior, and the capacity to structure social interaction systems to the purpose at hand. The clusters of characteristics discussed in this chapter differentiate leaders from followers, effective from ineffective leaders, and higher-echelon from lower-echelon leaders. In other words, different strata of leaders, and followers can be described in terms of the extent to which they exhibit some of the characteristics. Furthermore, research by Bass (1953), Moore and Smith (1953), and Tarnapol (1958) suggested that isolates and, to a lesser extent, followers and nonleaders can be described by the antonyms of trait names attributed to leaders. The characteristics generate personality dynamics that are advantageous to the person seeking the responsibilities of leadership.
The conclusion that traits are a factor in understanding leadership did not represent a return to the pure trait approach. It did represent a sensible modification of the extreme situationalist point of view. At first, the trait approach treated personality variables in an atomistic fashion, suggesting that each trait acts singly to determine the effects of leadership. Increasingly, models were tested describing how combinations of differentially weighted traits contributed to leadership. The wholly situationalist approach, on the other hand, denied the influences of individual differences, attributing all variance among persons to the demands of the environment. Again, it should be emphasized that some of the variance in who emerges as a leader and who is successful and effective is due to traits of consequence in the situation; some is due to situational effects; and some is due to the interaction of traits and situation. For example, suppose that candidates for management positions are tested in three situations: social service agencies, industrial firms, and military organizations. We are likely to find in the aggregate that individual interpersonal competence is predictive of successful performance. But it also may be more predictive in social service agencies and less predictive in military organizations. Considering the importance of competence to perform tasks and interpersonal competence at two stages in the careers of public accountants, engineers, and other technical specialists, one finds that both kinds of competence are important to performance. During a person’s early years with a firm, technical competence is most strongly indicative of successful performance; but after two to five years, interpersonal competence becomes more important.
The element of chance would appear to play a part in the rise of individual leaders. A given leader may be able to rise to the top of the hierarchy in competition with one group of peers but may be unable to do so in another group of peers. An individual’s upward mobility would seem to depend, to a considerable degree, on being in the right place at the right time. Finally, it should be noted that to some extent our conceptions of characteristics of leadership are culturally determined, as will be elucidated in Chapter 33. Situational contingencies will be examined more fully in chapters 25 through 29. These contingencies may be masked by consistent differences among individuals. This statement implies that there are regularities in individuals’ abilities, interests, orientations, values, and personality that endure across time. Other differences of consequence may include age, sex, family background, physique, and so on. Since the beginning of the study of leadership, the question has remained: how much do these individual differences account for the emergence of leadership and its effectiveness, and do the effects transcend situational circumstances? Some individuals will attempt to lead in most situations in which they find themselves, but others will avoid doing so whenever possible. Other individuals will attempt to be leaders only in certain situations, and their personal predispositions can be described. The same may be said about succeeding as a leader.
By 1970, there was plenty of evidence that particular patterns of traits were of consequence to leadership; these included determination, persistence, self-confidence, and ego strength. Additionally, the mass of research findings about the traits of leaders compared with nonleaders and of successful and effective leaders compared with unsuccessful and ineffective leaders included activity level, rate of talk, initiative, assertiveness, aggressiveness, dominance, ascendance, emotional balance, tolerance for stress, self-control, self-efficacy, enthusiasm, and extroversion. This was confirmed with qualification in studies in the years that followed.