Leadership theories: traits, behaviors, situations and transformations

In the second session Dr. Sato hands out a one-page list of leadership theories, from the trait studies of the early twentieth century to the authentic, servant and shared leadership models popular today. Owen counts nineteen of them and asks, half seriously, which one is right. Her answer frames the whole session. Each theory captured something real about how leaders influence groups, each was oversold in its time, and the useful question for a manager is not which theory is true but what each one tells you to pay attention to. A theory is a lens. The trait lens points at who you are, the behavioral lens at what you do, the situational lens at when and with whom, and the transformational lens at what you ask people to become.

She also warns the cohort about a problem that runs through the whole field. Much leadership research measures leadership by asking followers to rate their leader on questionnaires, and then correlates those ratings with other questionnaire answers from the same followers, such as satisfaction or commitment. Because the same people answer both, part of the relationship reflects their general feelings about the leader rather than any causal effect. The most credible studies use objective outcomes such as sales, turnover or safety incidents, field experiments in which leaders are randomly assigned to training, or ratings from different sources. Throughout the session she notes which kind of evidence supports each claim.

Traits: who leaders are

The earliest research looked for the traits that distinguish leaders from followers. A famous 1948 review by Ralph Stogdill concluded that no trait reliably did so across situations, and for several decades trait research fell out of favor. It returned with better measures and meta-analysis. A widely cited 2002 meta-analysis by Timothy Judge and colleagues found that the Big Five personality traits together had a multiple correlation of about 0.48 with leadership emergence and effectiveness. Extraversion was the most consistent predictor, followed by conscientiousness and openness to experience, while emotional stability showed smaller positive relationships and agreeableness was the weakest predictor overall.

General intelligence also predicts leadership, though less strongly than many people assume; a meta-analysis by Judge, Colbert and Ilies found a corrected correlation of about 0.27, with some evidence that very large gaps between a leader's intelligence and the group's can reduce the leader's influence. Traits matter most for who is chosen as a leader and somewhat less for how well the group then performs. The practical lesson is that personality is a starting point rather than a destiny. An introverted manager like Daniel cannot become an extravert, but he can learn the behaviors, such as speaking up early, structuring meetings and following up, that extraverts tend to display naturally.

Behaviors: what leaders do

From the late 1940s, researchers at Ohio State University and the University of Michigan shifted the focus from traits to behaviors. The Ohio State studies identified two broad dimensions. Consideration is the extent to which a leader shows concern and respect for followers, looks out for their welfare and expresses appreciation. Initiating structure is the extent to which a leader defines roles, organizes work and sets clear goals and standards. The Michigan studies found a similar pair, which they called employee-oriented and production-oriented behavior. Blake and Mouton later popularized the idea as a managerial grid, arguing that the best leaders score high on both concern for people and concern for results.

A 2004 meta-analysis by Judge, Piccolo and Ilies revived these two dimensions with impressive results. Consideration had a corrected correlation of about 0.48 with overall leadership outcomes and was particularly strongly related to follower satisfaction and motivation. Initiating structure had a correlation of about 0.29 and was more strongly related to leader job performance and group performance. Both mattered, for different things. This is good news for managers because both sets of behavior are concrete and learnable. Consideration looks like regular one-on-one conversations, remembering what matters to people and giving credit. Structure looks like clear priorities, defined roles and well-run meetings.

Situations: when and with whom

Behavioral theories left an obvious question: does the same style work everywhere? Contingency theories answered no. Fred Fiedler's contingency model proposed that a leader's style, measured with the least preferred co-worker scale, is relatively fixed, so the situation should be matched to the leader. Robert House's path-goal theory proposed that leaders should choose behaviors that clear the path to goals: directive leadership when tasks are ambiguous, supportive leadership when tasks are stressful or tedious, participative leadership when followers have expertise, and achievement-oriented leadership when followers want challenge. Paul Hersey and Ken Blanchard's situational leadership model, widely used in corporate training, prescribes more direction for less ready followers and more delegation as they gain competence and commitment.

The evidence for contingency theories is mixed. Path-goal theory has received partial support, especially for the idea that structure helps more when tasks are ambiguous and support helps more when tasks are stressful. Situational leadership is intuitive and popular but has weak research support for its specific prescriptions. Yet the core contingency insight has held up well: the effect of a leader's behavior depends on the task, the followers and the setting. A newer approach, leader-member exchange theory, adds that leaders develop different relationships with different followers. High-quality exchanges, built on trust and mutual obligation, are consistently associated with better performance, citizenship behavior and lower turnover, which raises the fairness question of who gets the leader's best attention.

Transformations: what leaders ask people to become

In 1978 the political historian James MacGregor Burns distinguished transactional leaders, who exchange rewards for performance, from transforming leaders, who raise followers' motivation and moral aspirations. Bernard Bass turned the idea into a measurable model. Transformational leadership has four components: idealized influence, in which the leader acts as a role model; inspirational motivation, in which the leader articulates an appealing vision; intellectual stimulation, in which the leader challenges assumptions and encourages new ideas; and individualized consideration, in which the leader attends to each follower's needs and development. Transactional leadership includes contingent reward and management by exception, and laissez-faire leadership is the absence of leadership.

Meta-analyses find that transformational leadership is positively associated with follower satisfaction, motivation and performance, and contingent reward is often nearly as effective, while laissez-faire leadership is consistently harmful. Field experiments are especially persuasive. In a well-known study, Israeli military platoon leaders randomly assigned to transformational leadership training led recruits who later scored higher on performance tests than recruits under untrained leaders. Critics note that the questionnaire measures blend the components together and include outcomes in the definition, so a leader rated as inspiring may simply be one whose team is doing well. Dr. Sato's summary is that the behaviors behind the model, a clear and meaningful purpose, a high bar paired with belief in people's ability to reach it, and genuine interest in individuals, are worth learning, whatever label is attached to them.

Newer models such as authentic, servant, ethical and humble leadership emphasize the leader's integrity and service to others. They capture real concerns about leaders who are charismatic but self-serving. However, a 2018 meta-analysis by Banks and colleagues found that measures of authentic leadership overlap so heavily with transformational leadership that they add little predictive power, and servant leadership showed somewhat more distinctiveness. For practice, the common thread is that followers judge leaders on whether they act in the group's interest, keep promises and treat people fairly.

# Scoring a short leader conduct survey and linking it to team outcomes
from statistics import mean, correlation   # Python 3.10+
teams = {   # team: ([consideration items], [structure items], turnover %, on-time %)
    "Costing":    ([4.2, 4.0, 4.4], [3.1, 3.4, 3.0], 9.0, 88),
    "Payables":   ([3.0, 2.8, 3.2], [4.3, 4.1, 4.4], 18.0, 95),
    "Payroll":    ([4.5, 4.6, 4.3], [4.2, 4.0, 4.4], 6.0, 97),
    "Reporting":  ([2.6, 2.9, 2.5], [2.8, 3.0, 2.7], 24.0, 81),
    "Treasury":   ([3.8, 3.9, 4.0], [3.6, 3.8, 3.5], 11.0, 92),
    "Plant cost": ([3.3, 3.5, 3.1], [3.9, 4.0, 3.7], 14.0, 90),
}
cons = [mean(v[0]) for v in teams.values()]
stru = [mean(v[1]) for v in teams.values()]
turn = [v[2] for v in teams.values()]
time = [v[3] for v in teams.values()]
for name, c, s in zip(teams, cons, stru):
    print(f"{name:<11} consideration {c:.2f}  structure {s:.2f}")
print(f"r(consideration, turnover) = {correlation(cons, turn):+.2f}")
print(f"r(structure, on-time close) = {correlation(stru, time):+.2f}")
print("Small sample: treat as a prompt for conversation, not proof.")

Costing     consideration 4.20  structure 3.17
Payables    consideration 3.00  structure 4.27
Payroll     consideration 4.47  structure 4.20
Reporting   consideration 2.67  structure 2.83
Treasury    consideration 3.90  structure 3.63
Plant cost  consideration 3.30  structure 3.87
r(consideration, turnover) = -0.97
r(structure, on-time close) = +0.94
Small sample: treat as a prompt for conversation, not proof.

Daniel built this toy analysis from a pulse survey across six finance teams in Northfield's parent group. Consideration is strongly and negatively related to turnover, and structure is strongly related to closing the books on time, which echoes the meta-analytic pattern. With only six teams and survey ratings from the same people, it proves nothing on its own, but it gives him a structured way to talk with his supervisors about what each team needs: the Reporting team appears short of both.

Who emerges as leader, and who is effective

Research distinguishes leader emergence, meaning who is seen as a leader by a group, from leader effectiveness, meaning whether the group performs well under that person. The two are related but not the same. People carry implicit theories, or prototypes, of what a leader looks like, and individuals who match the prototype tend to emerge as leaders even when they are no more effective. Narcissism is a striking example. A meta-analysis by Emily Grijalva and colleagues in 2015 found that narcissism is positively related to leader emergence, largely because narcissists appear confident and extraverted, but it is essentially unrelated to effectiveness overall, with some evidence that moderate levels are associated with the best ratings and very high levels with worse ones. Selection processes that reward confident self-presentation in a short interview can therefore choose people who look like leaders more reliably than people who lead well.

Gender and leadership has been studied extensively. Meta-analyses by Alice Eagly and colleagues found that women and men in comparable leadership roles differ only modestly in style, with women on average slightly more participative and slightly more likely to use transformational behaviors and contingent reward. Women leaders are rated as effective as men overall, but evaluations depend on context: women are often judged more harshly when they lead in a directive way or in roles stereotyped as masculine. Eagly's role congruity theory explains this as a clash between expectations of leaders and expectations of women, rather than a difference in competence. For managers who make promotion decisions, the lesson is to define what effective leadership looks like in behavioral terms before assessing candidates, so that a prototype does not substitute for evidence.