01 / 27
to navigate
Pathos Labs · Research

Executive Coaching Methods

What the Evidence Actually Shows

Sam Millunchick · 3 Meta-Analyses · 70+ Sources · February 2026

PL
The Bottom Line
g = 0.40–0.59

Moderate positive effect size across three independent meta-analyses of randomized controlled trials. Comparable to psychotherapy effect sizes.

De Haan & Nilsson 2023 · Nicolau et al. 2023 · Cannon-Bowers et al. 2023

The Field

A Young Science

Executive coaching sits at the intersection of organizational psychology, leadership development, and adult development theory. The first Coaching Psychology Unit was established in 2000. Rigorous research has accumulated meaningfully only in the last 15 years.

De Haan Grant Peterson Goldsmith Kegan Bachkirova Passmore
Harvard/McLean ICF Ashridge/Hult U. Sydney CCL
Key Findings
01 Coaching works. Three independent meta-analyses of RCTs converge on moderate positive effects. Established
02 The relationship is the strongest predictor. Working alliance predicts outcomes across all studied populations. Established
03 Specific methodology barely matters. No framework shows reliable superiority over any other. Strong
04 Behavior changes more than beliefs. Largest effects on observable behaviors, smallest on personality. Strong
Key Findings
05 Session count ≠ outcomes. No dose-response found — likely a measurement artifact, not a true null. Moderate
06 Virtual ≈ face-to-face. No significant difference (g = 0.48 vs g = 0.35). Moderate
07 ROI claims are inflated. The "788% ROI" is from a single industry-funded study, not independent research. Strong
08 Coachee readiness is a major predictor. Self-efficacy and motivation matter as much as anything the coach does. Strong
Section

Foundations

The 80/20 of what the research shows

02
Foundation 1

The Relationship Is the Method

The working alliance — agreement on goals, tasks, and the bond between coach and coachee — is the most consistent predictor across all studied populations. This mirrors 50+ years of psychotherapy research on common factors.

Graßmann et al. 2020 — 27 samples, N = 3,563 Established
De Haan et al. 2019 — 180 coachees, effect sizes up to d = 1.1 Strong

Personality matching between coach and coachee is not supported by the evidence. What matters is the relationship that develops, not demographic or personality similarity.

Foundation 2

Behavior Change Beats Belief Change

Cognitive activities
g = 1.28
Performance behaviors
g = 1.11
Leadership behaviors
g = 0.44
Attitudes
g = 0.34

Source: Nicolau et al. 2023. Behavioral wins create momentum, build self-efficacy, and often pull attitudes along behind them.

"What will you do differently in the next meeting?" is more powerful than "How do you need to think differently about leadership?"
Foundation 3

Readiness Trumps Technique

Coachee self-efficacy and readiness for change predict outcomes at least as strongly as anything the coach does. Coaching is not something done to someone.

Self-efficacy predicts outcomes independently (De Haan 2013) Strong
Readiness + core self-evaluations predict outcomes (MacKie 2015) Moderate

"Forced coaching" is a fundamentally different proposition. When readiness is low, the first phase of coaching should focus on building readiness — not jumping to behavior change.

Foundation 4

Methodology Is a Vehicle, Not a Destination

GROW, Co-Active, Solutions-Focused, Cognitive-Behavioral — no specific methodology has been shown to be reliably superior. Most share the same active ingredients: goal-setting, reflective questioning, accountability, a supportive relationship.

Process-oriented
g = 0.45
vs
Outcome-oriented
g = 0.39
No significant difference
"Commit yourself heart and soul to your approach, even if you know that it doesn't matter which professional approach you choose." Erik de Haan — Relational Coaching
Foundation 5

Vertical, Not Just Horizontal

Horizontal

Adding skills at the same level of complexity

Vertical

Increasing the sophistication of how someone makes meaning

Socialized Mind Self-Authoring Mind Self-Transforming Mind

Kegan & Lahey's framework (2009). Strong theoretical base, but limited RCT evidence specific to coaching applications. Emerging

Section

The Evidence

Three meta-analyses published in 2023

03
Meta-Analyses
De Haan & Nilsson
2023 AMLE
37 RCTs · N = 2,528
g = 0.59
Nicolau et al.
2023 Frontiers Psych
20 RCTs with pre-post tests
g = 0.43
Cannon-Bowers et al.
2023 Frontiers Psych
11 studies (10 after outlier removal)
Medium +

⚠ De Haan & Nilsson flag "indications of significant publication bias." True effect size may be smaller. Trim-and-fill adjusted: g = 0.27.

Outcome Predictors

What Predicts Coaching Outcomes?

Coach selection should focus on relational capacity and skill, not "fit" questionnaires or personality matching.

Counterintuitive Findings

Surprises From the Data

Session count doesn't predict outcomes

Appears in all three meta-analyses. Likely reflects measurement limitations — most studies cluster in the 6–12 session range. Quality of sessions likely matters more than quantity.

Virtual coaching works just as well

g = 0.48 face-to-face vs g = 0.35 virtual — not statistically different (Cannon-Bowers 2023).

Credentials matter inconsistently

Larger effects with qualified coaches in non-leadership contexts. But evidence that credentialing predicts better outcomes is thin. Emerging

Section

Practical Tactics

Evidence-grounded moves you can use

04
Tactics
01
Front-Load Relationship
Low difficultyHigh impact
Dedicate first 1-2 sessions to understanding the coachee's world. Ask about their pressures, wins, what keeps them up at night.
02
Behavioral Experiments
Low-Med difficultyHigh impact
Within the first 2-3 sessions, identify one specific behavioral experiment to run before the next session.
03
Peterson's Pipeline
Low difficultyMed-High impact
Diagnose which of 5 conditions is blocked: Insight, Motivation, Capabilities, Practice, or Accountability.
04
Screen for Readiness
Medium difficultyHigh impact
Assess: Is this executive here by choice? Do they believe they can change? What's their prior experience with feedback?
Tactics
05
Integrate the Body
Med-High difficultyMed-High impact
Somatic approaches for presence and performance. State management — breath, posture, grounding — operates below conscious thought.
06
Immunity to Change
High difficultyHigh impact
Surface hidden competing commitments when executives can't change despite genuine motivation. Kegan's 5-step mapping process.
07
Stakeholder Feedback
Medium difficultyHigh impact
Build in feedback from 3-5 key stakeholders. Makes results visible (sustains organizational support) and provides real-world accountability.
Deep Dive

Peterson's Development Pipeline

Where is the coachee stuck? Most coaching failures happen because coaches apply "more of the same" instead of diagnosing which condition is missing.

1 Insight Do they know what to change?
2 Motivation Do they care enough?
3 Capabilities Do they know how?
4 Practice Do they get to practice?
5 Accountability Are structures reinforcing it?

From David Peterson (Google). A coachee with insight and motivation but lacking capabilities needs teaching, not more reflective questions. Moderate

Deep Dive

Immunity to Change

For executives who know what to do but can't seem to do it. The closest coaching gets to therapeutic depth without crossing the therapy line.

1 Name the stated commitment "I want to delegate more"
2 Identify behaviors that work against it micromanaging, rescuing tasks
3 Surface the hidden competing commitment "never being seen as not having the answers"
4 Identify the big assumption "if people see me not knowing, they'll lose confidence"
5 Design small, safe tests of the big assumption

Kegan & Lahey (2009). Moderate

Section

Contrarian Views

What the critics say — and where they're right

05
Contrarian Views
"The evidence is weaker than the industry claims"
Total RCT pool across all 3 meta-analyses: just 37 + 20 + 11 studies. De Haan flags "significant publication bias." Most studies use self-report. Much research is by practitioners with financial interests. Strong critique
"ROI claims are marketing, not science"
The widely cited "788% ROI" is from a single MetrixGlobal industry-funded study. ICF's "7x return" traces to similar sources. The honest answer: "Reliable moderate positive effects on behavior. Specific dollar returns are not credibly established."
"Coaching models are post-hoc rationalizations"
GROW, Co-Active, CLEAR, OSKAR — developed from practitioner experience, not derived from research. No RCT tests any named model against alternatives. These serve training revenue more than clients. Strong critique
Contrarian Views
"The unregulated industry problem is real"
100,000+ practitioners globally with wildly varying quality. ICF's ACC credential requires just 60 hours of training — roughly 7 working days. The barrier to entry is lower than almost any other profession working intimately with people's psychology.
"Coaches without psychological training are dangerous"
Berglas (HBR, 2002): coaches who lack psychological training "make a bad situation worse" by treating narcissistic personality traits or clinical depression as "leadership challenges."
Negative effects are real but uncommon: dependency on coach, worsening supervisor relationships when the coachee changes but the system doesn't, reduced job satisfaction when coaching reveals values-job misfit. Schermuly & Graßmann, 2019
Section

So What?

Decision translation — what changes if this is true

06
Implications
If relationship > method → Invest in relational capacity, not certification collection. Your ability to make someone feel seen matters more than your framework.
If behavior first → Structure around experiments from session 2. "What will you do differently?" before "How should you think differently?"
If readiness matters most → Screen for it. When an org sends you a "problem" executive, clarify whether the executive actually wants to be coached.
If methodology barely matters → Master 2-3 approaches. Hold each loosely. Don't chase every new certification.
If ROI claims are inflated → Never promise dollar returns. Frame outcomes in observable behavioral terms that stakeholders can verify.
Open Questions

What We Don't Know Yet

References

Key Sources

Meta-Analyses
De Haan & Nilsson (2023) — AMLE. 37 RCTs, N = 2,528, g = 0.59
Nicolau et al. (2023) — Frontiers Psych. 20 RCTs, g = 0.43
Cannon-Bowers et al. (2023) — Frontiers Psych. 11 studies
Graßmann et al. (2020) — Human Relations. 27 samples, N = 3,563
Key Studies & Texts
De Haan et al. (2019) — AMLE. 180 coachees, 66 coaches
Kegan & Lahey (2009) — Immunity to Change
Peterson (2006) — The Development Pipeline
Goldsmith — Stakeholder Centered Coaching
Berglas (2002) — HBR: Dangers of Executive Coaching

Full source log with 70+ references available in the complete research document.

Pathos Labs
Sam Millunchick

Sam Millunchick

High-stakes communication coaching for founders and executives

Executive Coaching Methods — What the Evidence Actually Shows

Shelf-life: Monitor — active research area

February 2026 · 3 Meta-Analyses · 70+ Sources

PL