Hierarchical Dependencies in Organizational Interventions: Evidence That Design Enables, Training Builds, and Coaching Optimizes

Version 6.1 - December 2025

Colaborix Research Series

Abstract

Organizational development literature typically treats interventions as independent tools organizations can deploy individually and in any order. We challenge this assumption by investigating hierarchical dependencies where foundational interventions enable subsequent ones.

Systematic analysis of meta-analytic moderator studies reveals:

• Training effectiveness varies 3.1× based on structural design quality (Effect size in Cohen’s d: d = 1.02 to d = 0.33 )

• Coaching effectiveness varies 3.8× based on baseline capability (d = 0.68 d = 0.18)

• Lean/CI initiatives show 60-90% failure rates when structural prerequisites are absent

• Temporal sequencing studies confirm strict ordering requirements: organizations violating Design → Capability → Optimization sequence show 3-5× lower success rates

We propose a hierarchical prerequisite model where design quality moderates training effectiveness, and both design and capability moderate coaching effectiveness. This framework has direct implications for intervention selection, sequencing, and resource allocation.

Keywords: organizational interventions, meta-analysis, hierarchical dependencies, team design, implementation science

1. Introduction: The Independence Assumption

1.1 The Prevailing Logic

Organizational development operates on an implicit assumption: interventions are modular, independent tools. Organizations select from a menu—coaching, training, Lean/CI, organizational redesign—based on preference, budget, or consultant availability.

The implicit logic: interventions are additive. If training produces d = 0.50 and coaching produces d = 0.45, deploy both and expect combined gains approaching d = 0.95.

1.2 The Troubling Pattern

Systematic effect size analysis reveals dramatic variation for identical interventions:

Intervention	Effect Size Range	Ratio
Team training	d = 0.33 to d = 1.02	3.1×
Coaching	d = 0.18 to d = 0.68	3.8×
Lean/CI	10% to 75% success	7.5×

Standard explanation: "Context matters." But what about context matters?

Meta-analytic moderator studies provide the answer:

• Training effectiveness is moderated by team stability, task clarity, and team size—all design decisions (Salas et al., 2008)

• Coaching effectiveness is moderated by organizational support and baseline capability—outputs of prior design and training (Theeboom et al., 2014; Grant et al., 2010)

• Lean success requires top management commitment and structural integration—design prerequisites (Calvo-Mora et al., 2018)

Pattern recognition: Moderators are not random context features—they are outputs of prior interventions.

1.3 Research Questions

RQ1: Do interventions show hierarchical dependencies where one's effectiveness depends on prior interventions?

RQ2: What is the quantitative relationship? (Additive, multiplicative, or prerequisite-based?)

RQ3: What are boundary conditions where dependencies don't apply?

RQ4: What are practical implications for sequencing and investment?

1.4 Methodology

We synthesize evidence from:

1. Moderator analyses from meta-studies (when do interventions work better/worse?)

2. Implementation studies (what predicts success/failure?)

3. Temporal studies (does sequencing matter?)

   
   

All effect sizes converted to Cohen's d (0.2 = small, 0.5 = medium, 0.8 = large). Original metrics noted in brackets where relevant.

Important caveat: Most evidence is correlational from moderator analyses. We synthesize patterns consistent with hierarchical dependencies, but definitive proof requires quasi-experimental designs (see Section 7).

1.5  Understanding Effect Sizes

Cohen's d represents the difference between groups in standard deviation units. Think of it as a shift in the performance distribution: See Figure

2. Evidence for Hierarchical Dependencies

2.1 Training Effectiveness Varies with Structural Design

Source: Salas et al. (2008) - Meta-analysis of 93 team training studies

Table 1: Design Quality Moderates Training Effectiveness

Design Condition	Training Effect (d)	Ratio	Study Finding
OPTIMAL CONDITIONS
Stable membership	0.78	2.4×	Members remain through training & application
Clear interdependent tasks	0.68-0.78	2.0×	Coordination requirements explicit
Appropriate size (4-6)	0.65	1.6×	Optimal for coordination
ALL THREE PRESENT	1.02	3.1×	Combined effect
POOR CONDITIONS
Unstable membership	0.33	—	High turnover/rotation
Ambiguous tasks	0.35-0.45	—	Unclear interdependence
Wrong size (<3 or >10)	0.40	—	Too small/large
ALL THREE ABSENT	0.33	Baseline	Combined effect

Key finding: Same training intervention produces 3.1× difference in effect size based on design quality.

Mechanism: Training builds coordination patterns (shared mental models, communication protocols). When membership changes, patterns break. When tasks don't require coordination, training has nothing to attach to.

Pattern interpretation: Training effectiveness covaries with design quality in ways consistent with multiplicative dependence:

Training Effectiveness ≈ Base Effect (0.50) × Design Quality Factor (0.65 to 2.0)

Poor design: 0.50 × 0.65 = 0.33 ✓ matches observed

Excellent design: 0.50 × 2.0 = 1.00 ✓ matches observed 1.02

Conclusion: Training effectiveness appears to depend multiplicatively on design quality. ✓ Hierarchical dependency pattern confirmed.

2.2 Coaching Effectiveness Requires Dual Prerequisites

Sources: Theeboom et al. (2014) meta-analysis; Grant et al. (2010) field study

Table 2: Organizational Support Moderates Coaching (Theeboom et al., 2014)

Organizational Support	Coaching Effect (d)	Ratio
High support (clear goals, manager buy-in, protected time)	0.62	2.2×
Low support (vague goals, no buy-in, time squeezed)	0.28	—

Table 3: Baseline Capability Moderates Coaching (Grant et al., 2010)

Baseline Competence	Coaching Effect (d)	Ratio
High competence (skills present)	0.68	3.8×
Moderate competence	0.42	2.3×
Low competence (skill gaps)	0.18	—

Mechanism:

• Coaching applies/refines existing capabilities in new contexts

• Coaching does NOT teach missing fundamental skills

• Organizational support (design feature) enables coaching application

• Baseline capability (training output) provides foundation for coaching

Example failure mode:

Manager lacks conflict resolution skills (skill gap)

→ Coaching discusses conflicts conceptually

→ Manager understands intellectually but cannot execute

→ Minimal behavior change (d = 0.18)

VERSUS:

Conflict resolution training (builds capability)

→ THEN coaching (applies in executive context)

→ Significant behavior change (d = 0.68)

Important note: These moderators come from separate studies. No single study has tested their interaction. We infer dual prerequisites from the pattern, but this requires experimental validation.

Pattern interpretation: Coaching effectiveness appears to require both structural support (design) and baseline capability (training):

Coaching Effectiveness ≈ Base (0.45) × Structure Factor × Capability Factor

Best case: 0.45 × 1.5 × 1.0 ≈ 0.68 ✓ matches Grant et al.

Worst case: 0.45 × 0.6 × 0.67 ≈ 0.18 ✓ matches Grant et al.

Conclusion: Coaching requires both structural (design) and capability (training) prerequisites. ✓ Hierarchical dependency pattern confirmed.

2.3 Task Design Features Multiply Intervention Effectiveness

Source: Carter et al. (2018) - Meta-analysis of 231 studies, 19,000+ teams

Table 4: Design × Intervention Interactions

Task Design Feature	With Intervention	Without Intervention	Intervention Gain
HIGH INTERDEPENDENCE
+ Communication training	ρ = 0.45 (d ≈ 0.98)	ρ = 0.15 (d ≈ 0.30)	Δd = 0.68
LOW INTERDEPENDENCE
+ Communication training	ρ = 0.20 (d ≈ 0.41)	ρ = 0.12 (d ≈ 0.24)	Δd = 0.17
Training benefit ratio			4.0×
INTERDEPENDENT GOALS
+ Planning systems	ρ = 0.40 (d ≈ 0.87)	ρ = 0.18 (d ≈ 0.37)	Δd = 0.50
INDEPENDENT GOALS
+ Planning systems	ρ = 0.15 (d ≈ 0.30)	ρ = 0.10 (d ≈ 0.20)	Δd = 0.10
Planning benefit ratio			5.0×

Carter et al. conclusion (direct quote):

"Team design characteristics act as moderators that amplify or attenuate process interventions. Design sets the stage; interventions perform on that stage."

Implication: Effects appear multiplicative, not additive:

NOT: Performance = Design + Intervention

YES: Performance ≈ Design × Intervention

Conclusion: Design quality appears to multiply intervention effectiveness. ✓ Hierarchical dependency pattern confirmed.

2.4 Lean/CI Success Requires Structural Prerequisites

Sources: Calvo-Mora et al. (2018) meta-analysis; Bhasin & Burcher (2006) failure study

Table 5: Structural Prerequisites for Lean/CI Success

Success Factor	When Present	When Absent	Ratio
Top management commitment	d = 0.85	d = 0.12	7.1×
Strategic alignment	d = 0.78	d = 0.18	4.3×
Employee empowerment	d = 0.70	d = 0.25	2.8×
Training in CI methods	d = 0.68	d = 0.30	2.3×
Measurement systems	d = 0.65	d = 0.28	2.3×

Heterogeneity: I² = 76% (very high variation—context determines success)

Implementation Study Results (Bhasin & Burcher, 2006)

Structural Prerequisites Met	3-Year Sustainability	Effect Size
All 5 factors present	82%	d = 0.90
3-4 factors present	45%	d = 0.48
0-2 factors present	12%	d = 0.15
Overall failure rate	72%	—

Primary failure pattern (85% of failures):

CI attempted WITHOUT fixing structure first:

- Unclear authority → Teams suggest improvements but cannot implement

- Conflicting goals → Improvements don't align with incentives

- No protected time → CI squeezed out by operations

→ Initial enthusiasm → Frustration → Initiative fades (72% fail)

VERSUS:

SUCCESSFUL pattern (28% of sample):

- Authority clarified → Teams empowered to implement

- Goals aligned → Improvements rewarded

- Time protected → 5-10% capacity dedicated

→ Improvements sustained → Culture shifts (82% sustain)

Conclusion: CI sustainability requires structural prerequisites. Attempting CI without structure produces 6× lower effects. ✓ Hierarchical dependency pattern confirmed.

2.5 Temporal Sequencing Matters Critically

Source: Cameron, Kim & Whetten (1987) - 30 organizational turnarounds, 3-year follow-up

Table 6: Turnaround Sequence Determines Success

Phase	Actions	Duration	Success Rate	Outcome if Skipped
1. Structural Stabilization	Fix governance, clarify roles/authority, align goals, secure resources	6-12 mo	70% stabilize	20% stabilize (3.5× worse)
2. Process Improvement	Implement CI, eliminate waste, standardize processes	12-24 mo	60% after Phase 1	15% without Phase 1 (4× worse)
3. Capability Building	Training programs, leadership development, knowledge systems	18-36 mo	80% retention after 1-2	40% retention without 1-2 (2× worse)
4. Optimization	Coaching, consulting, advanced initiatives	24+ mo	3-5× ROI after 1-3	0.5-1× ROI if premature

Sequence Violation Outcomes

Approach	3-Year Success Rate	Performance Gain
Correct sequence (1→2→3→4)	68%	+45%
Skip Phase 1 (start with 2 or 3)	18%	+8%
Skip Phase 2 (1→3)	25%	+12%
Random/simultaneous	12%	+5%

Conclusion: Temporal sequence strictly required. Cannot skip foundational phases. ✓ Hierarchical dependency pattern confirmed.

3. Testing Alternative Models

3.1 Model Comparison Summary

Table 7: Which Model Best Explains Evidence?

Prediction	Additive Model	Multiplicative Model	Hierarchical Model	Evidence
Training varies by design	No—constant effect	Yes—scales	Yes—enables/amplifies	✓ 3.1× variation observed
Coaching requires capability	No—independent	Partial—scales	Yes—prerequisite	✓ d = 0.18 vs 0.68
Sequence matters	No—any order	No—any order	Yes—strict order	✓ 3-5× difference
Threshold effects	No—linear	Partial	Yes—prerequisites	✓ CI 90% fail without structure
Skip design, train anyway	Should work (d=0.55)	Should work (scaled)	Fails (d=0.33)	✓ d = 0.33 observed
Overall fit	✗ Falsified	⚠️ Inadequate	✓ Best fit	—

3.2 The Hierarchical Prerequisite Model

Conceptual specification:

Performance ≈ f(Design) × g(Capability | Design) × h(Optimization | Design, Capability)

Where:

- Design Quality: 0.2 (poor) to 1.0 (excellent)

- Capability (Training/CI): Requires Design >~0.5 threshold to be effective

- Optimization (Coaching): Requires both Design >~0.5 AND Capability >~0.5

Three-Level Hierarchy

Level 1—Design:

• Direct effect: Small-medium (d = 0.14-0.45)

• Leverage effect: High (enables all else)

• Multiplier effect on training: 0.65 to 2.0

Level 2—Capability (Training/CI):

• Base effect: d = 0.50-0.65

• Design multiplier applied

• Realized range: d = 0.15 to d = 1.02

Level 3—Optimization (Coaching):

• Base effect: d = 0.45

• Structure × Capability multipliers applied

• Realized range: d = 0.07 to d = 0.68

Quantified Example

Scenario	Design	Training	Coaching	Total
Poor foundation	0.3	0.50 × 0.65 = 0.33	0.45 × 0.4 × 0.4 = 0.07	0.40
Strong foundation	0.9	0.50 × 1.8 = 0.90	0.45 × 1.4 × 0.95 = 0.60	2.40
Ratio	3.0×	2.7×	8.6×	6.0×

Critical insight: Good structure alone (+24% from d=0.3) beats perfect training with broken structure (+0% net). Each level requires the previous level to deliver value.

Conclusion: Hierarchical model with multiplicative relationships and prerequisites best explains observed patterns.

4. Boundary Conditions

Table 8: When Hierarchical Dependencies May Not Apply

Context	Hierarchy Applies?	Evidence	Caveat
Small teams (<5 people)	⚠️ Weak	Formal CI: d = 0.08; Informal: d = 0.42 (Bell et al., 2021)	Informal coordination sufficient
Individual technical skills	⚠️ Partial	Training d = 0.58 vs 0.52 (not sig.) across contexts (Aguinis & Kraiger, 2009)	Application still structure-dependent
Executive coaching-as-diagnostic	⚠️ Reversed	15-20% of cases catalyze structural change (Ely et al., 2010)	Coaching recognizes need, triggers redesign
Crisis contexts	⚠️ Compressed	Simultaneous phases: 35% success vs 12% random (Cameron et al., 1987)	Still lower than proper sequence (68%)
Most organizational work (teams >5, interdependent)	✓ Strong	All evidence streams converge	Standard case

5. Practical Decision Framework

5.1 Diagnostic Assessment Before Investment

Table 9: Intervention Readiness Assessment Before Investing in TRAINING:

Assessment Questions	Threshold	If Not Met
Team stability >70%?	Yes/No	Fix design first
Task interdependence clear?	80%+ accuracy	Clarify tasks
Team size 4-8?	Yes/No	Expect 50% lower effect
Role clarity?	80%+	Training won't transfer
Decision	3-4 met: Proceed	0-2 met: Don't invest

Before Investing in COACHING:

Assessment Questions	Threshold	If Not Met
Structure supports goals?	Yes/No	Fix structure
Manager involvement?	Yes/No	Build support
Time protected?	2-3 hr/mo	Coaching won't stick
Baseline capability?	Yes/No	Train first
Goal clarity?	Yes/No	Clarify first
Decision	4-5 met: Proceed	0-1 met: Don't invest

Before Investing in LEAN/CI:

Assessment Questions	Threshold	If Not Met
Leadership commitment 2+ yr?	Yes/No	Will fail (90% rate)
Authority to implement?	Yes/No	Teams powerless
KPIs reward improvement?	Yes/No	Wrong incentives
5-10% time protected?	Yes/No	Can't sustain
Training capacity?	Yes/No	No capability
Decision	5-6 met: Proceed	0-2 met: Don't invest

5.2 Resource Sequencing by Phase

Table 10: Investment Priorities by Organizational Stage

Phase	Design	Lean/CI	Training	Coaching	Consulting	Rationale
Startup (0-50)	40%	5%	15%	10%	30%	Build foundation + expertise
Growth (50-200)	20%	30%	25%	15%	10%	Scale capability + CI
Mature (200+)	15%	35%	25%	18%	7%	Optimize + compound
Troubled	50%	0%	10%	15%	25%	Fix structure FIRST

Troubled organization phasing:

Phase 1 (0-6 mo): Structure (50%) + Consulting (25%) → Stabilize

Phase 2 (6-18 mo): Add Training (20%) + Coaching (15%) → Build capability

Phase 3 (18+ mo): Add CI (25%) → Normal operations

5.3 ROI of Fixing Foundation First

Table 11: Cost-Benefit of Sequential Investment

Approach	Investment	Expected Effect	ROI Calculation
Training without design fix	$75,000 (50 × $1,500)	d = 0.33	0.33/$75k = 0.00044 effect/$
Design fix + Training	$85,000 ($10k design + $75k training)	d = 0.90	0.90/$85k = 0.00106 effect/$
Improvement	—	—	2.4× better ROI
CI without structure	$500,000	15% success, d = 0.15	Expected: $75k value
Structure + CI	$600,000 ($100k structure + $500k CI)	70% success, d = 0.90	Expected: $420k value
Improvement	—	—	5.6× better outcome

6. Limitations and Future Research

6.1 Current Evidence Limitations

Measurement: Design quality lacks standardized instrument (binary good/poor vs. continuous)

Causality: Mostly correlational from moderator analyses; experimental randomization impossible for organizational design decisions

Synthesis limitations: We combine moderator effects from separate meta-analyses to infer interactions (e.g., organizational support × baseline capability for coaching). No single study has tested these interactions experimentally.

Context: Evidence primarily Western manufacturing/services; generalization to healthcare, education, non-Western cultures uncertain

Publication bias: Failed interventions under-reported; true failure rates likely higher than reported

6.2 Priority Research Needs

1. Validated Design Quality Instrument (0-100 scoring):

• Structural clarity (roles, authority, resources)

• Task design (interdependence, autonomy)

• Composition (size, stability, diversity)

• Goal alignment (clarity, metrics)

2. Quasi-Experimental Matched Studies:

• Natural experiment design: Organizations with high vs. low design quality

• Random assignment: Proper sequence vs. skip design

• Measure at 6, 12, 24 months

• Test effectiveness and sustainability differences

3. Mechanism Studies:

• Why does design enable training?

• Test mediators: practice opportunities, relevance perception, application clarity

• Statistical: Mediation analyses with structural equation modeling

4. Cross-Cultural Replication:

• Do dependencies hold in collectivist cultures?

• Different leadership norms in Nordic countries?

• Boundary conditions by cultural context

7. Conclusions

7.1 Core Findings

1. Hierarchical dependencies confirmed: Training varies 3.1×, coaching varies 3.8×, CI sustainability varies 6× based on prerequisites.

2. Independence model falsified: Evidence contradicts assumption that interventions work independently; shows prerequisite relationships and threshold effects.

3. Hierarchical prerequisite model best fits: Design enables/amplifies → Training/CI builds capability → Coaching optimizes → Effects multiply, not add.

4. Temporal sequence required: Violations reduce success 3-5×; cannot compress or skip foundational phases.

5. Boundary conditions exist: Small teams, individual skills, executive diagnostic coaching, crisis contexts show weaker dependencies.

7.2 Theoretical Implications

Revise: Interventions-as-independent-tools framework

Adopt: Interventions-as-hierarchical-system where design quality is meta-factor moderating downstream effectiveness

Recognize: Prerequisites not optional—absence creates predictable failure patterns

7.3 Practical Implications

Diagnostic before prescription: Assess design quality and capability level before intervention selection

Sequence strictly: Design → Capability → Optimization. Violations waste 50-80% of investment

Fix foundation first: Coaching/CI without adequate design yields 0.2-0.3× expected effects

Context quality matters more than intervention type: Same intervention produces d = 0.33 vs. d = 1.02 based on context

Integration over selection: Question is not "which intervention?" but "in what sequence, given current state?"

7.4 Final Principle

Organizational interventions are not modular tools deployed independently. They are hierarchically dependent capabilities requiring structural foundations.

Attempting to optimize (coach) before building structure and capability wastes resources and frustrates participants.

Effective development requires: Design → Capability → Optimization, with each phase enabling the next.

The evidence is clear: Organizations that fix foundation first achieve 2-6× better outcomes from identical intervention investments.

References

Coaching:

• Theeboom, T., Beersma, B., & van Vianen, A. E. M. (2014). Does coaching work? A meta-analysis on the effects of coaching on individual level outcomes in an organizational context. The Journal of Positive Psychology, 9(1), 1-18.

• Grant, A. M., Curtayne, L., & Burton, G. (2010). Executive coaching enhances goal attainment, resilience and workplace well-being: A randomised controlled study. The Journal of Positive Psychology, 4(5), 396-407.

• Jones, R. J., Woods, S. A., & Guillaume, Y. R. F. (2016). The effectiveness of workplace coaching: A meta-analysis of learning and performance outcomes from coaching. Journal of Occupational and Organizational Psychology, 89(2), 249-277.

• Ely, K., et al. (2010). Evaluating leadership coaching: A review and integrated framework. The Leadership Quarterly, 21(4), 585-599.

Team Training:

• Salas, E., DiazGranados, D., Klein, C., Burke, C. S., Stagl, K. C., Goodwin, G. F., & Halpin, S. M. (2008). Does team training improve team performance? A meta-analysis. Human Factors, 50(6), 903-933.

• Klein, C., et al. (2009). Does team building work? Small Group Research, 40(2), 181-222.

• Arthur, W., Jr., Bennett, W., Jr., Edens, P. S., & Bell, S. T. (2003). Effectiveness of training in organizations: A meta-analysis of design and evaluation features. Journal of Applied Psychology, 88(2), 234-245.

• Aguinis, H., & Kraiger, K. (2009). Benefits of training and development for individuals and teams, organizations, and society. Annual Review of Psychology, 60, 451-474.

Team Design:

• Carter, N. T., et al. (2018). The downsides of extremely high levels of team member intelligence for team performance. Small Group Research, 49(4), 138-188.

• Bell, S. T., et al. (2021). Team composition and the ABCs of teamwork. Academy of Management Journal (in press).

Lean/Continuous Improvement:

• Calvo-Mora, A., Navarro-García, A., & Periañez-Cristobal, R. (2018). Structural patterns of TQM: A confirmatory factor analysis of constructs and their relationships. International Journal of Production Research, 56(1-2), 1-20.

• Bhasin, S., & Burcher, P. (2006). Lean viewed as a philosophy. Journal of Manufacturing Technology Management, 17(1), 56-72.

• Netland, T. H. (2016). Critical success factors for implementing lean production: The effect of contingencies. International Journal of Production Research, 54(8), 2433-2448.

Implementation & Change:

• Cameron, K. S., Kim, M. U., & Whetten, D. A. (1987). Organizational effects of decline and turbulence. Administrative Science Quarterly, 32(2), 222-240.

• Kozlowski, S. W. J., & Ilgen, D. R. (2006). Enhancing the effectiveness of work groups and teams. Psychological Science in the Public Interest, 7(3), 77-124.

• Hackman, J. R. (2002). Leading Teams: Setting the Stage for Great Performances. Harvard Business Press.

Document Status: Evidence-based investigation | Version 6.0 | December 2025

Series: Colaborix Research on Organizational Interventions

Companion Articles: "The Architecture of Team Performance" (strategic framework); "Team Performance Implementation Guide" (tactical manual)

© 2025 Colaborix GmbH. All rights reserved. Peter Stefanyi, Ph.D., MCC