Trust, Rapport & Psychological Safety
Trust, Rapport & Psychological Safety
Three distinct but interlocking literatures govern professional relationships. Trust is dyadic willingness to be vulnerable. Rapport is the moment-to-moment relational tone. Psychological safety is a group-level climate belief. They are not the same thing — build all three deliberately.
This skill is about trust between people. For trust in machines/automation (over-reliance, algorithm aversion, calibration), use human-ai-interaction-psychology.
Meta-finding across all three: trust and safety are slow to build and fast to destroy (Slovic). Match every move to that asymmetry.
Deep detail, the repair decision aid, the full 7-item scale, and source citations live in references/trust-rapport-safety-detail.md. Load it when you need specifics.
1. Interpersonal trust models
| Construct | Core idea | Operator application (customer + team) |
|---|---|---|
| ABI model (Mayer, Davis & Schoorman 1995) | Trustworthiness = Ability + Benevolence + Integrity; trust = willingness to be vulnerable. | Integrity + ability read fast; benevolence is earned over time. New account/teammate: prove competence and keep promises now; don’t fake care. |
| Cognitive vs affective trust (McAllister 1995) | Head-trust (track record, competence) vs heart-trust (emotional bond). Cognitive usually precedes affective. | Earn the customer’s cognitive trust with reliability before expecting affective trust; a peer relationship needs both. |
| Trust Equation (Maister 2000) | (Credibility + Reliability + Intimacy) / Self-Orientation. | Highest-leverage lever = lower self-orientation (customer outcome over your quota). Reliability compounds; intimacy differentiates. |
| Swift trust (Meyerson, Weick & Kramer 1996) | Temporary teams import trust presumptively from role/credentials, then verify. Fragile. | First TAM engagement, new escalation bridge: borrow credibility from role/references/standards; protect it — one early miss does outsized damage. |
| Stages of trust (Lewicki & Bunker 1996) | Calculus-based → knowledge-based → identification-based (most resilient). | Early customer = calculus stage, breakable. Strategic, long-tenured accounts reach identification trust that survives an outage. |
2. Trust violation & repair
| Construct | Core idea | Operator application |
|---|---|---|
| Violation type → repair (Kim et al. 2004) | Competence breach → apology beats denial. Integrity breach → denial beats apology only when you’re innocent. | Blown SLA / wrong rec (competence) → apologize + show the fix. Honesty/commitment question (integrity) → if innocent, explain/deny with evidence; deny only if true. |
| Reticence (Ferrin, Kim et al.) | Silence is worse than a clear apology or a clear denial. | Never go quiet after a trust hit — unresolved suspicion festers. |
| Asymmetry principle (Slovic 1993) | Trust is built slowly, destroyed in an instant; one bad event erases many good ones. | Over-invest in prevention and rapid, visible recovery. You cannot average your way back. |
See the 6-step repair decision aid in the reference file.
3. Rapport & relationship dynamics
| Construct | Core idea | Operator application |
|---|---|---|
| 3 components (Tickle-Degnen & Rosenthal 1990) | Rapport = mutual attentiveness + positivity + coordination; positivity leads early, coordination later. | First meeting: visible warmth + full attention. Long relationship: demonstrated sync. |
| Similarity-attraction (Byrne 1971) | Shared attitudes/values drive liking (proportion, not count). | Surface genuine common ground (tech philosophy, prior tooling) — never manufactured agreement. |
| Mere-exposure (Zajonc 1968) | Repeated positive contact increases liking; first exposures matter most. | A short, reliable weekly touchpoint can beat a rare big meeting for warmth. |
| Chameleon effect (Chartrand & Bargh 1999) | Unconscious mimicry/synchrony increases liking — but deliberate exaggerated mimicry backfires. | Let synchrony emerge; the safe deliberate move is matching pace/medium/formality, not posture-mirroring. |
| 5:1 ratio (Gottman) | Healthy relationships run ~5 positive : 1 negative; high-performing teams run positive-heavy. | Audit your interaction ledger with a tense account/teammate; deliberately add positives. (Exact team number is heuristic — see caveat.) |
| Active-constructive responding (Gable & Reis 2004) | How you respond to good news predicts bond quality more than bad news. ACR (enthusiastic + ask to elaborate) is the only style that builds it. | Highest-ROI micro-skill: when a customer/teammate shares a win, celebrate and ask them to say more. |
4. Psychological safety (team-level)
| Construct | Core idea | Operator application |
|---|---|---|
| Team psychological safety (Edmondson 1999) | Shared belief the team is safe for interpersonal risk-taking → learning behavior → performance. | An internal-team lever, not a customer one. Make it safe to ask, admit errors, experiment. |
| 7-item scale (Edmondson 1999) | 7-point Likert, 3 reverse-scored (held-against-you, reject-for-difference, hard-to-ask-help). | Measure team safety with the scale (full wording in refs). |
| Psych safety ≠ trust | Safety = group climate, near-term, “is it safe to be me?”; trust = dyadic, longer-term, “will they act well?”. | Don’t assume dyadic trust delivers team safety or vice versa — build both. |
| Leader behaviors (The Fearless Organization 2019) | Frame work as learning; model fallibility; invite participation; respond productively to bad news. | Erode it by blaming the messenger or punishing honest mistakes. |
| 4 stages (Clark 2020) | Inclusion → Learner → Contributor → Challenger safety. | Diagnose where it breaks; most teams stall before Challenger safety, killing dissent/innovation. |
| Project Aristotle (Google 2012-15) | Psych safety = #1 of 5 dynamics of effective teams. | How the team interacts beats individual talent. |
Safety is NOT niceness. Pair high safety with high accountability/standards (Edmondson’s learning zone). Never sell it as “be nice.” And trust is NOT blindness — mature trust is calibrated by evidence.
Caveats (knowledge gaps from the research)
- Edmondson 7-item exact wording varies across secondary reproductions; the version in refs is the most common, but verify against the 1999 ASQ original before any formal measurement.
- Team positive:negative ratio — the marital 5:1 is well replicated; the team threshold rests partly on Losada’s math, later critiqued (Brown, Sokal & Friedman 2013). The direction (high performers run positive-heavy) holds; treat a specific team number as heuristic, not law.
Reusable prompt — relationship diagnosis & action plan
Paste this, fill the brackets, and the assistant will diagnose against these frameworks and return concrete next moves.
You are a relationship strategist grounded in the trust/rapport/psychological-safety
literature (Mayer-Davis-Schoorman ABI; McAllister cognitive/affective; Maister Trust
Equation; swift trust; Kim et al. competence-vs-integrity repair; Slovic asymmetry;
Tickle-Degnen rapport; Gottman 5:1; Gable-Reis ACR; Edmondson psychological safety +
4-stages). This is about trust between PEOPLE, not trust in automation.
CONTEXT: [customer account OR internal team]
RELATIONSHIP: [who, tenure, current temperature, recent events]
SYMPTOM: [e.g. champion has gone cold / team won't surface bad news / trust dropped
after an outage / new account, no relationship yet]
Do this:
1. CLASSIFY the situation: dyadic trust vs team psychological safety vs rapport (or which mix).
2. If a trust BREACH occurred, classify it competence vs integrity and prescribe the
matched repair (apology vs evidence-based denial); never recommend silence; account
for the slow-build/fast-break asymmetry.
3. Score the relationship on the Trust Equation (credibility/reliability/intimacy/
self-orientation) and name the single highest-leverage lever.
4. If team-level, map it to Clark's 4 stages and name where safety breaks; recommend
leader behaviors; pair any safety move with an accountability move.
5. Return 3-5 concrete next actions for the next 2 weeks, each tied to the framework
that justifies it. Flag where evidence is heuristic (e.g. team ratio thresholds).