What Is Leadership? Definitions That Actually Help You Lead

Leadership is a process in which one or more people influence a group to work together toward a shared goal. Researchers define it in many ways, but virtually all of their definitions put influence at the core, Victor Vroom and Arthur Jago concluded in a 2007 paper in American Psychologist: if no one is following, no one is leading.¹

If you are asking what is leadership because you now have to practice it, the definitions that help most go one step further. They judge leadership by what the leader does and what happens to the group, not by a job title or a personality type.

What is leadership? The one idea every definition shares

Influence is the one idea nearly every research definition of leadership shares, according to Victor Vroom and Arthur Jago’s 2007 paper, and their own definition adds a purpose: motivating people to work together toward a goal they share. They describe leadership as “a process, not a property of a person.”¹

The wording varies because leadership has no formal, standardized scientific definition; a 1990 handbook of leadership research, which the paper quotes, suggested there were almost as many definitions as people who had tried to write one. Definitions split on two questions in particular: which kinds of influence count, since some researchers exclude coercion, and whether a group’s success belongs in the definition at all.¹

Defining leadership by what people do has evidence behind it. In a 2011 meta-analysis led by Scott DeRue, leaders’ traits and behaviors together accounted for at least 31 percent of the variation in leadership effectiveness, and behaviors tended to account for more of it than traits did. Effectiveness was measured four ways: leader effectiveness, group performance, followers’ job satisfaction and their satisfaction with the leader. The authors concluded that a model in which traits act through behavior is warranted.³ Because a meta-analysis like this pools correlations, it shows how strongly each goes with effectiveness, not what causes it.

Four definitions, and what each one tells you to watch

Four research definitions are worth keeping because each points your attention somewhere different: Vroom and Jago’s at influence, Robert Hogan and Robert Kaiser’s at the team’s results, the functional view at the team’s unmet needs, and the relational view at how people experience working with you.

The first two definitions disagree about results. Hogan and Kaiser argue that leadership should be judged by how the group performs over time, and they note that most studies instead define it by who emerges as leader in a group of strangers or by how senior managers rate someone.⁴ Vroom and Jago want results kept out of the definition, because a group’s success depends on many things besides its leadership.¹ A workable reading of both: define leadership by influence, then judge your own by results, knowing that results reflect more than leadership.

The functional view is the easiest to act on. It holds that leaders should give a team what it needs when it needs it, such as clear direction, a structure and context that enable the work, coaching and adequate resources, examples the review takes from J. Richard Hackman and Ruth Wageman, whose team research also underlies what makes a team effective.² A 2010 review by Frederick Morgeson and colleagues describes 15 team leadership functions and four distinct sources, inside and outside the team, that team leadership can come from, so it need not come only from the person in charge.⁵

The relational view holds that leadership is co-produced by leaders and the people around them, a phrase from Boas Shamir that guided the 2017 review’s look ahead.²

Looking like a leader is not the same as leading well

Researchers separate leader emergence, being seen as the leader within a group, from leadership effectiveness, influencing others in ways that help a group reach its goals. A 2017 review of a century of leadership research in the Journal of Applied Psychology, by Robert Lord and colleagues, warns that studies often blur the two, because effectiveness tends to be rated by the leader’s own supervisors rather than measured objectively.²

Personality research shows the gap. In a 2002 meta-analysis by Timothy Judge and colleagues, as the 2017 review reports, the five major personality traits together were more strongly related to emergence than to effectiveness (multiple correlations of .53 and .39).²

A 2015 meta-analysis led by Emily Grijalva found a sharper split for narcissism. It was positively related to emerging as a leader, a link explained by narcissists’ extraversion, but unrelated overall to effectiveness as rated by supervisors, subordinates and peers, although narcissistic leaders rated themselves as more effective. The flat overall link hid a curve, with a middling level of narcissism looking best.⁶

Myth

Confident, self-promoting people make the best leaders.

Fact

Narcissistic people are more likely to emerge as leaders, not to be rated as more effective: a 2015 meta-analysis linked narcissism to emerging as a leader but found no overall link to effectiveness as rated by others.

Hogan and Kaiser draw the practical conclusion: evaluate leaders by the performance of the group they are responsible for over time, not by how leaderlike they seem.⁴

Further reading

• 

  Leadership: Theory and Practice

  by Peter G. Northouse

  A standard textbook that walks through the trait, behavior, transformational and team approaches this article compares, with the research behind each.

As an Amazon Associate WiserHours earns from qualifying purchases.

What do effective leaders do differently?

Two broad families of leader behavior go with better outcomes, according to research summarized in a 2017 review by Robert Lord and colleagues: structuring the work, such as clarifying roles, rules and procedures, and showing consideration, such as being friendly and supportive toward followers. The focus on behavior began with the Ohio State studies of the late 1940s, which set out to discover what leaders do rather than who they are, and grew rapidly in the 1950s; these two families then dominated leadership research until the mid-1980s.²

1. To the 1940sTraits: who leaders are
2. Late 1940s-1950sBehaviors: what leaders do
3. 1960s-70sContingency: what the situation calls for
4. Mid-1980s onTransformational and charismatic leadership
5. Late 1990s onShared and collective leadership

A 2004 meta-analysis by Timothy Judge, Ronald Piccolo and Remus Ilies pooled more than 300 correlations and found both families linked to leadership outcomes, consideration more strongly (.48) than structure (.29). Consideration went more with followers’ satisfaction, motivation and leader effectiveness; structure went slightly more with the leader’s job performance and the group’s performance.⁷

Two newer style labels are transformational leadership, which builds a sense of mission, and contingent reward, which means setting goals and rewarding people for meeting them.² They overlap heavily: in a 2004 meta-analysis of 626 correlations from 87 sources, Timothy Judge and Ronald Piccolo found that the two correlated .80 with each other, that transformational leadership correlated .44 with outcomes and contingent reward .39, and that contingent reward did better on several outcomes. Laissez-faire leadership had a negative link (−.37).⁸

The study

Moderate evidence

What links leader behavior to performance: 35 meta-analyses combined

Ryan Gottfredson and Herman Aguinis combined 35 meta-analyses, 26 published and 9 of their own, covering 3,327 studies, to test why four major leadership behaviors (consideration, structuring the work, contingent reward and transformational leadership) go with better follower performance. Of the explanations they compared, one was involved in the largest indirect links for all four: how followers rated their relationship with the leader, known as leader-member exchange.⁹

The authors read their results as suggesting that leader behaviors work mainly by improving the relationship with each person, the kind the most common questionnaire probes with items on whether your supervisor understands your problems and needs and whether you know where you stand. The caveats are theirs too: the underlying studies were cross-sectional, which limits causal conclusions, and relationship questionnaires share items with measures of leader behavior.⁹ A 2013 critique by Daan van Knippenberg and Sim Sitkin goes further for transformational leadership, arguing that its most-used questionnaires do not separate it from other leader behaviors and mix it up with its own effects.¹⁰

A 2021 meta-analysis by Henry Young and colleagues found the same relationship route for transactional leadership, with a twist: contingent reward was linked to better relationships but less empowerment, giving it both a positive and a negative indirect link with performance.¹¹

1. Leader behaviors: structuring the work, showing consideration, rewarding results, building a sense of mission
2. The relationship: how the follower rates it (leader-member exchange)
3. Follower performance: the job itself and help beyond it

Does leadership change a team’s results?

Bosses can make a large difference to output, estimates from a 2015 study by economists Edward Lazear, Kathryn Shaw and Christopher Stanton suggest. At one large services company, replacing a boss from the bottom 10 percent of quality with one from the top 10 percent was estimated to raise a team’s total output by about as much as adding a tenth worker to a nine-person team, or more.¹²

The researchers tracked daily productivity for 23,878 workers and 1,940 bosses between 2006 and 2010 in computer-monitored service jobs, where bosses mainly taught and motivated their teams, and workers assigned to better bosses were less likely to leave. By one of the authors’ normalizations, the average boss was about 1.75 times as productive as the average worker. Workers were not formally randomized to bosses, though the authors’ tests suggest assignment was close to random, and the findings come from one company and one kind of job.¹²

Google’s own manager research points the same way, with the caveats that apply to company research. Google says the work began with a hypothesis, voiced by some of its own people in the early 2000s, that managers don’t actually matter. It says its Project Oxygen then analyzed more than 10,000 data points from performance reviews and feedback surveys, plus interview notes, and found that teams with highly effective managers were more likely to achieve better results, be happier and have lower turnover. Its list of 10 manager behaviors mixes structuring and consideration, from being a good coach to focusing on productivity and results.¹³ It is one company studying its own managers, and it is not peer-reviewed.

Can leadership be learned?

Who ends up in a leadership role is partly heritable, but most of the variation is not genetic, and leadership training is followed by measurable gains, according to a 2013 twin study and a 2017 training meta-analysis. In the twin study of 346 same-sex twin pairs in the US, Jan-Emmanuel De Neve and colleagues estimated that genes accounted for about 24 percent of the variation in whether people supervised others at work, with wide uncertainty (6 to 46 percent). The authors note that supervising others sits at the lower end of leadership roles, so the estimate may understate heritability for senior ones.¹⁴

The training evidence is larger. A 2017 meta-analysis of 335 samples led by Christina Lacerenza found gains after leadership training in participants’ reactions, learning, transfer of skills to the job and organizational results. Gains were bigger when programs began with a needs analysis and used feedback, practice, spaced sessions and face-to-face delivery on site.¹⁵ The 2011 finding that behaviors accounted for more than traits points the same way, since behavior is what training targets.³

Do you need a title to lead?

A job title is neither necessary nor sufficient for leadership, research on shared leadership and Hogan and Kaiser’s 2005 review suggest. In a 2014 meta-analysis of 42 samples, Danni Wang and colleagues found that shared leadership, in which team members lead one another, was positively related to team effectiveness, more strongly when the work was complex and more for team attitudes and processes than for performance itself.¹⁶ Early research on problem-solving groups had already separated leadership functions from formal roles, on the view that any member could perform them.²

Titles and leadership also come apart at the top. Hogan and Kaiser argue that people rise in large organizations through hard work, brains, ambition, political skill and luck, but not necessarily through talent for leadership. Their competency model offers one way to separate leading from managing: it treats leadership skills, building and motivating a high-performing team, as one of four skill sets a manager needs, separate from the business skills of planning, budgeting, coordinating and monitoring work. They also rank leadership skills as the easiest of the four to train, a judgment rather than a tested result.⁴

Four questions that turn the definitions into practice

Each of the four research definitions of leadership becomes a question you can put to your own team: whether people share the goal (Vroom and Jago), how the team’s results compare (Hogan and Kaiser), what it needs that nobody provides (the functional view) and how each working relationship is going (the relational view and the 2017 synthesis of 35 meta-analyses).

The bottom line

Leadership is influence toward a goal a group shares, and the definitions that help you lead treat it as something you do rather than something you are. Judge yours by the group: its results over time, what it needs that nobody is providing, and how each person would describe working with you. Behaviors tended to account for more of leaders’ effectiveness than traits in the 2011 meta-analysis, and training research finds gains in how people lead on the job.³¹⁵