How to Improve Psychological Safety in Your Team

Why is psychological safety hard to improve?

Amy Edmondson's original research on psychological safety found that it forms through repeated interpersonal experiences within a team — specifically through how leaders respond when team members take interpersonal risks (speaking up, admitting mistakes, proposing unconventional ideas). The pattern of responses over time becomes the team's shared belief about how safe it is to take those risks.

This means improvement is necessarily slow and consistency-dependent. It also means regression is fast: one high-visibility incident where speaking up goes badly can reset months of progress.

What leader behaviours most reliably improve psychological safety?

1. Model fallibility — go first with a real mistake

   In your next team meeting, share something you got wrong recently and what you learned from it. Be specific ("I underestimated how long the API migration would take and didn't communicate the delay early enough — here's what I'd do differently") rather than vague ("I know I don't always get things right"). Specificity signals authentic vulnerability; vagueness signals performative humility. When leaders go first, the implicit message is: admitting error is safe here.

2. Respond to bad news with curiosity, not blame

   When a team member brings you a problem — especially one they contributed to — your response in that moment determines whether anyone brings you problems early in the future. Replace "Why did this happen?" (blame signal) with "Walk me through what happened" (curiosity signal). Replace immediate problem-solving with "What do you think we should do next?" Your facial expression and body language matter as much as the words.

3. Structure dissent into every team meeting

   Don't wait for people to volunteer disagreement — build space for it into the meeting structure. Before finalizing any significant decision, ask: "What am I missing?" or "Who sees this differently?" Then wait. Silence after these questions is not neutral — it is the moment where people decide whether speaking up is worth the risk. Hold the pause for at least 10–15 seconds before moving on.

4. Follow through visibly on what the team tells you

   The fastest way to destroy psychological safety is to ask for feedback and do nothing with it. After every feedback cycle — survey or otherwise — tell the team publicly what you heard and what you're changing because of it. Even if the answer is "I heard this concern and I'm not able to address it right now because of X" — that transparency is safety-building. Silence after feedback is safety-destroying.

5. Protect the person who speaks up in group settings

   When someone raises a concern or a dissenting view in a meeting, your public response determines whether anyone else will do the same. Thank them specifically for the challenge. Engage with the substance before responding to the position. Never signal — through tone, dismissal, or moving on immediately — that the challenge was unwelcome. Other team members are watching to calibrate their own risk assessment.

What team-level practices support psychological safety?

• Blameless post-mortems — after any significant failure or near-miss, run a structured debrief that asks "what happened in the system?" rather than "who made the mistake?". This frames errors as information about process gaps rather than evidence of personal failure — the exact reframe needed for psychological safety.
• Structured retrospectives — run regular retrospectives (monthly or after major projects) with explicit equal-voice mechanisms: round-robin input, anonymous sticky notes, or written submissions before the meeting so quieter team members aren't pre-empted by dominant voices.
• Peer appreciation rituals — practices where team members recognize each other's contributions (not just managers recognizing reports) reduce status hierarchy effects that suppress lower-status voices. Brief, regular, and specific beats annual and elaborate.
• Anonymous feedback channels — give the team a way to surface concerns without attribution pressure. Anonymous pulse surveys are the most reliable mechanism; they systematically remove the interpersonal risk from feedback without requiring the team to trust individual relationships enough to speak up directly.

How do you measure whether psychological safety is improving?

Timeline expectations for consistent leader behaviour change:

• Weeks 1–4 — No measurable score change expected. Team members are observing whether the new behaviours are consistent or performative. Trust is being evaluated, not yet given.
• Weeks 4–8 — Early behavioural signals may appear: slightly more questions in meetings, one or two people speaking up who hadn't before. Survey scores may tick up marginally.
• Weeks 8–16 — If behaviours have been consistent, survey scores should show a measurable improvement (typically 0.3–0.7 points on a 5-point scale). More people participating in retrospectives. Earlier escalation of problems.
• 6+ months — Psychological safety is becoming a team norm, not just a leader behaviour. New team members onboard into a culture where safety is the expected standard rather than the exception.

Frequently asked questions about improving psychological safety

Related guides

• What Is Psychological Safety and How Do You Build It? — the foundational guide to understanding what psychological safety is and why it is the strongest predictor of team performance.
• How to Run an Anonymous Team Survey (Step-by-Step) — how to measure psychological safety honestly using anonymous surveys that get around the interpersonal risk of direct feedback.
• 10 Warning Signs Your Team Health Is Declining — silence in meetings and declining survey participation are the top early signals that psychological safety is eroding.
• Team Health KPIs: Which Metrics Actually Matter? — how to track psychological safety as a quantitative KPI alongside the other three team health dimensions.