Lead+D Lab · four independent evidence bases · 2026-07-28
One question, asked four independent ways: when a manager is abusive and the environment turns toxic, what actually happens? Each source has a different bias, so where they agree we can trust the finding, and where they diverge we know to be careful.
The strongest validation comes free: the themes people spontaneously describe are the items scholars put in their abusive-supervision scales. What gets measured is what gets lived.
| Scale construct (theory) | Voice theme (lived) | Voice count |
|---|---|---|
| Ridicule / hostility / yelling | Verbal abuse and emotional volatility | 87 |
| Public put-downs / scapegoating | Public humiliation and scapegoating | 49 |
| Retaliation / punitive exits | Retaliation, sabotage, and forced exits | 89 |
| Favoritism / injustice | Favoritism, nepotism, and exclusionary cliques | 40 |
| Intrusive control | Micromanagement and intrusive surveillance | 36 |
| Weaponized evaluation | Weaponized performance management | 23 |
All four sources agree on the shape: fairness rupture and retaliation are loudest; turnover is the quietest consequence.
| Consequence domain | Meta-analytic r | metaBUS r (k) | What the voices show | Rank |
|---|---|---|---|---|
| Justice / relational rupture | interactional −.51 to −.66 | −.554 (k=6) | Favoritism (40), humiliation (49), broken promises (17) — the fairness wound is named constantly | STRONGEST |
| Retaliation / counterproductive behavior | supervisor-directed ρ=.53 | +.431 (k=38) | Retaliation, sabotage & forced exits is the #1 voice theme (89) | VERY STRONG |
| Job attitudes (satisfaction, commitment) | −.20 to −.35 | −.301 (k=13) | Incompetent leadership & dysfunction (64) — disillusionment | STRONG |
| Wellbeing / strain | exhaustion ρ≈.36 | +.374 (k=7) | Verbal abuse (87), overwork (35), boundary breach (26) | MODERATE-STRONG |
| Performance / citizenship | task perf −.13 to −.16 | −.128 (k=32) | Rarely volunteered — people describe withdrawing effort, not output drops | MODERATE |
| Turnover / withdrawal | intent .30 / actual ~.04 | +.253 (k=6) | 221 quit BUT 141 wanted out and stayed — leaving is not the dominant response | WEAKEST |
| Claim | Confidence | Why |
|---|---|---|
| Bad management → injustice, retaliation, disengagement | SOLID | 4 sources converge: two meta-analyses, metaBUS (k=38 for CWB), and the voices where retaliation & unfairness dominate. |
| Wanting to leave ≫ actually leaving | SOLID | Meta-analytic intention r≈.30 vs actual r≈.04; voices show 141 wanted-out-but-stayed alongside 221 who quit. |
| The harm is psychological, not physical/financial-to-the-firm | SOLID | metaBUS health r≈−.06; Schyns org-performance r=.039 (ns). Damage is attitudes, fairness, strain, behavior. |
| Culture attenuates the exit response (high power distance) | MODERATE | Meta: West r=.26 vs Asia r=.04. Voices echo it (English quit-ratio 1.87 vs Thai 0.80) — but the voice data confounds language with platform (all Thai = Pantip), and the Asian meta cell is thin. |
| Psychological safety as the mechanism | THIN | Theory is strong (Frazier: leadership→psych-safety ρ=.44) but the abusive-supervision literature barely measures it — voice/psych-safety k=4 in metaBUS. |
| A single clean 'toxic boss' construct | CAUTION | Hershcovis: 72% of overlapping aggression constructs are empirically indistinguishable. The umbrella (destructive leadership) inflates turnover/performance vs the AS-specific grain (.22 vs .34). |