Why Leadership Training Shows Inconsistent Results

In the United States, leadership development takes the largest share of organizations’ training budgets. The implicit promise: train your leaders, and performance will follow.

The research tells a more complicated story. Leadership training can produce meaningful results. The largest meta-analysis, 335 studies, found an average effect of δ = .76—a substantial impact by social science standards. Training reliably improves what leaders know and can do.

But here’s the problem: few organizations think they are getting those results. In the survey the meta-analysis opens with, only 13% believed they had done a quality job of training their leaders. Many invest in programs, collect positive feedback from participants, and watch nothing change.

The gap between what leadership training can do and what it typically does suggests a fundamental misunderstanding about how development actually works. Training is necessary but insufficient. Without the conditions that enable transfer, even excellent programs change little back at work.

What the Research Shows

The evidence on leadership training effectiveness is more nuanced than either enthusiasts or skeptics suggest.

Meta-analyses consistently find that leadership training produces positive effects. The Lacerenza et al. (2017) meta-analysis of leadership training found an overall effect of δ = .76 across 335 independent samples and more than 26,000 people. That’s meaningful impact.

But the findings come with important caveats.

δ = .76 Average effect of leadership training across 335 studies—larger on the job where programs used practice, feedback and sessions spread over time

Effects hold at every level of what you measure. Using Kirkpatrick’s four-level framework:

Level 1 (Reaction): Did participants like the training? Effect size δ = .63, from only seven studies. Participants generally report positive experiences.

Level 2 (Learning): Did participants gain knowledge and skills? Effect size δ = .73. Training reliably produces learning.

Level 3 (Behavior): Did participants change their on-the-job behavior? Effect size δ = .82—the largest of the four, though the authors found the four did not differ significantly.

Level 4 (Results): Did training produce organizational outcomes? Effect size δ = .72. Within that, outcomes for the organization rose far more (δ = .75) than outcomes for the leader’s direct reports (δ = .22).

Here’s the catch: an evaluation that stops at Levels 1-2 measures satisfaction and knowledge, and never learns whether behavior actually changed or whether results followed. Most organizations collect reaction data, the meta-analysis notes, citing an industry report.

This can create a dangerous illusion. Positive feedback and post-test knowledge gains convince stakeholders that training “worked.” Meanwhile, leaders return to environments that don’t support new behaviors, nothing changes, and the cycle repeats.

The likely result: training without support back at work produces little lasting change. Yet organizations continue deploying programs as if knowledge acquisition automatically becomes behavior change.

The Transfer Climate Problem

Baldwin and Ford’s foundational review of training transfer (1988) traced whether learning becomes behavior to three inputs: the trainee, the design of the training, and the work environment. In the work environment, they named three factors:

Supervisor support. Do leaders’ managers expect, reinforce, and model the trained behaviors? If a participant’s boss doesn’t support the new approach, it is unlikely to stick.

Peer encouragement. Do colleagues reinforce trained behaviors or undermine them? Social pressure in the work environment can overwhelm training content.

Practice opportunities. Do leaders have chances to apply new skills with feedback? Without repeated application in real situations, skills fade.

A program that addresses none of these factors sends leaders to training, measures satisfaction, and expects transformation. The training itself might be excellent. The environment can make transfer impossible.

Consider a common scenario: A manager completes a program on coaching skills. They return to work Monday morning. Their inbox has 200 emails. Three direct reports need immediate decisions. A deadline looms. Their own manager asks why a project is behind schedule.

When, exactly, are they supposed to practice coaching? The organizational context—workload, expectations, time pressure—actively works against applying new skills.

“Training is an individual remedy, often deployed in systems that can make transfer impossible. We keep sending people to programs without changing anything about the environment they return to.”

The Evaluation Gap

The way organizations evaluate training can create systematic blindness to whether it works.

Much training evaluation stops at Kirkpatrick Levels 1-2:

Level 1 surveys ask: “Did you find this valuable?” “Would you recommend it?” “Was the facilitator engaging?” Participants generally say yes—they got time away from work, the content seemed reasonable, and there’s social pressure to be positive.

Level 2 assessments measure knowledge gained: pre-test vs. post-test comparisons, skill demonstrations during the program. Participants generally show improvement—they just learned the material.

Neither level tells you whether anything will change back at work.

Levels 3-4 require more effort:

Level 3 (Behavior) requires observation over time. Are leaders actually doing things differently? This means 30-day, 60-day, 90-day follow-ups. Manager observations. 360-degree feedback comparisons. Many organizations don’t invest in this measurement.

Level 4 (Results) requires connecting leader behavior to outcomes. Did team engagement improve? Did turnover decrease? Did performance metrics shift? This requires tracking over 6-18 months and controlling for other variables.

Levels 1-2 Where evaluation stops when it measures satisfaction and knowledge but never behavior change or organizational results

The likely result: organizations accumulate evidence that training is “well-received” while having no idea whether it produces results. Budget decisions get made based on participant satisfaction rather than organizational impact.

This isn’t laziness—it’s structural. L&D teams face pressure to demonstrate activity. Satisfaction scores are easy to collect and report. Behavior change measurement requires collaboration with operations, extended timelines, and methodological sophistication that many L&D functions don’t have.

The incentives favor programs that feel good over programs that work.

The Trickle-Down Problem

Leadership training operates on an implicit assumption: improve leaders, and their teams will benefit. The research on this “trickle-down” effect is sobering.

The largest meta-analysis gives a measure of it. Leadership training reliably changed leaders: after training, the average leader’s on-the-job behavior scored above about 78% of leaders before or without training, rather than half of them. For their direct reports the shift was much smaller, to about 58%. Training lifted organizational outcomes (δ = .75) far more than outcomes for direct reports (δ = .22).

The authors found this surprising. The tie between a leader and their team is the most direct one there is, so improvements should reach the team most readily. They reached it less than they reached the organization.

Why the disconnect?

Timing matters. Effects on a team are likely to take longer to show than effects on the leader, and a study that stops measuring early will miss them. Effects that might emerge at 12-18 months go unmeasured.

Dosage matters. In the meta-analysis, longer programs produced larger effects on results, though length made no measurable difference to learning or to behavior on the job. Shorter, cheaper options are the easier sell.

Context matters. Leader behavior change is unlikely to produce team benefits unless the organizational context supports it. A leader trying new approaches in a culture that punishes experimentation is unlikely to sustain the behavior—and their team won’t benefit.

What the Evidence Points To

The research points to conditions that make leadership training more effective:

Extended Duration with Spaced Practice

One-off workshops do less. In the Lacerenza meta-analysis, programs with several sessions spread over time produced more behavior change on the job, and better results, than a single massed session—though not more learning. Longer programs also did better on results.

Manager Involvement and Accountability

When participants’ managers are involved—setting expectations before training, debriefing after, observing application, providing feedback—transfer is more likely. Supervisor support is the work-environment factor the transfer literature has long pointed to (Baldwin & Ford, 1988). Without manager involvement, training competes against daily pressures and loses.

Behavioral Practice with Feedback

Knowledge acquisition is easy. Behavior change is hard. Programs that combine content with demonstration and practice—role-play, simulation, real-situation application with coaching—outperformed content-only programs in the meta-analysis, and feedback added to behavior change on the job.

Organizational Context Alignment

Training that contradicts organizational norms can produce cynicism rather than change. If you train leaders on empowerment while the organization rewards control, you’ve wasted everyone’s time. Effective development aligns with—or works to shift—broader organizational culture.

Multi-Level Evaluation

Programs that measure behavior change and results—not just satisfaction and knowledge—create accountability for actual impact. What gets measured gets attention.

“Training without environmental support is an individual fix for a systemic problem. The program might be excellent. The context can make transfer impossible.”

What This Means Practically

If you’re investing in leadership development, the research suggests several shifts:

Stop Buying Programs, Start Building Systems

Training is one component of development, not the whole thing. Effective leadership development requires: selection of the right people, quality training content, transfer climate support, practice opportunities, feedback mechanisms, and accountability for application.

It is easy to invest heavily in training content and neglect everything else. Flip the ratio.

Extend Evaluation Timelines

Demand Level 3-4 evaluation. If your L&D team says they can’t measure behavior change, that’s a capability gap to address—not a reason to accept satisfaction scores as success metrics.

Build 6-12 month evaluation windows into program design from the start. Budget for follow-up measurement.

Involve Managers in the Process

Before training: Have participants’ managers set expectations and identify specific behaviors to develop.

During training: Keep managers informed of content so they can reinforce it.

After training: Require manager check-ins on application. Include manager observations in evaluation.

Address Transfer Climate Explicitly

Ask: What in our environment will support or undermine application of this training? Workload? Competing priorities? Manager behavior? Cultural norms?

If the honest answer is “the environment will undermine it,” either change the environment or don’t waste money on training.

Match Intensity to Objectives

Brief awareness sessions can shift attitudes. Behavior change requires extended engagement with practice and feedback. Organizational culture change requires sustained intervention over years.

Match program design to what you’re actually trying to accomplish.

The Bottom Line

Leadership training works—under the right conditions. Meta-analyses show meaningful effect sizes. Well-designed programs produce real behavior change, and it reaches organizations and, to a lesser degree, teams.

But the argument of this piece is that much leadership training fails in practice, because organizations ignore the conditions that make it effective. They buy programs, collect satisfaction scores, and wonder why nothing changes.

The problem isn’t training. It’s the assumption that training alone produces transformation.

Organizations serious about leadership development need to think systemically: What conditions support transfer? How will we measure behavior change? What environmental factors need to shift? How will managers reinforce new behaviors?

Without those questions—and real answers—leadership training is an expensive way to feel like you’re doing something about a problem you’re not actually addressing.

References

Baldwin TT, Ford JK. Transfer of training: a review and directions for future research. Personnel Psychology. 1988;41(1):63-105. doi:10.1111/j.1744-6570.1988.tb00632.x

Kirkpatrick DL. Evaluating Training Programs: The Four Levels. Berrett-Koehler; 1994.

Lacerenza CN, Reyes DL, Marlow SL, et al. Leadership training design, delivery, and implementation: a meta-analysis. Journal of Applied Psychology. 2017;102(12):1686-1718. doi:10.1037/apl0000241