Is Leadership a Skill? What a 73-Sample Meta-Analysis Shows

Is leadership a skill, or is it a trait some people are simply born with? A 2002 meta-analysis in the Journal of Applied Psychology, pooling 222 correlations across 73 independent samples, found personality traits do predict leadership, but the size of that prediction depends entirely on what you mean by "leadership."

Why "born leader" explains less than it claims to

"Born leader" gets said in performance reviews the way "good chemistry" gets said about a team, as a label that closes the conversation rather than opening it. The research does not support closing it. The 2002 meta-analysis found the Five Factor Model of personality correlates with leadership at a multiple R of 0.53 when the outcome is leader emergence, who gets noticed, promoted, and put in charge. For leadership effectiveness, whether the person actually leads well once in the role, the same traits correlate at only R=0.39. Personality is a noticeably better predictor of who gets the job than of who is good at it.

What the meta-analysis actually found

Extraversion was the most consistent individual trait, correlating with leadership at r=0.31 across settings and criteria, ahead of Conscientiousness (r=0.28), Neuroticism (r=-0.24), and Openness (r=0.24). Agreeableness barely registered at r=0.08.

Read together, the gap between 0.53 and 0.39 is the honest finding. A confident, extraverted personality gets a person into the room faster than it makes them good once they are in it.

Whether leadership can actually be trained

If personality only partly explains who leads well, the practical question is whether the rest is buildable. A 2017 meta-analysis of 335 independent leadership-training evaluations, the largest synthesis of its kind, found training programs produced real improvement across every outcome measured: trainee reactions (δ=0.63), learning (δ=0.73), on-the-job behavior transfer (δ=0.82), and organizational results (δ=0.72). Transfer, whether a trained leader actually behaves differently at work, is the outcome that matters most, and it is also the largest effect in the set.

The same meta-analysis found programs built on a documented needs analysis, actually identifying the gap a leader has before designing the program around it, produced significantly stronger learning and transfer effects than programs skipping that step. That advantage did not extend to organizational results, where the two groups did not differ significantly, a useful reminder that a well-designed program still cannot promise a downstream business outcome, only a real change in the leader's own behavior. A fixed trait would not move like this under a training program with a comparison group attached.

Where this fits the six domains this site measures

Leadership is one of the six domains this site scores, described there as setting direction when the situation is ambiguous, and being willing to be the person who is wrong. The research above is why it gets scored as behavior under uncertainty rather than a personality reading: the same trait that gets someone chosen barely predicts whether they lead well once chosen. Teamwork, one of the other five domains, shows the identical pattern, a shared, buildable behavior rather than a fixed quality a person either has or lacks. New hires who struggle once in a role usually fail on attitude rather than raw ability, the same distinction a resume adjective like "born leader" cannot make.

None of this makes leadership easy to fake past a structured observer. It means the opposite: it is one of the few workplace words with a real, checkable behavior underneath it once someone bothers to measure the right outcome, the same standard this site applies everywhere else.