Overview
Based on "Leaders Eat Last" by Simon Sinek and Amy Edmondson's psychological safety research. Sinek's insight: humans are wired to withhold their best thinking in environments that feel unsafe. The Circle of Safety - the conditions under which people feel protected from external threats - is the leader's most fundamental responsibility. Without it, people use cognitive energy protecting themselves from internal threats (judgment, blame, embarrassment) rather than solving real problems. Edmondson's research found psychological safety is the single strongest predictor of team performance - more than talent, resources, or strategy.
The Four Stages of Psychological Safety
Edmondson and Sinek's work maps to a progression:
| Stage | Description | Signal |
|---|
| Inclusion | People feel accepted as members of the team | New members contribute within the first week |
| Learner | People feel safe asking questions and admitting they don't know | People say "I don't know" without hedging |
| Contributor | People feel safe doing their work and trying new approaches | People take initiative without asking permission |
| Challenger | People feel safe challenging the status quo and authority | People disagree with the leader openly |
Most teams that think they have psychological safety are at Inclusion or Learner. True high performance requires Challenger safety.
Workflow
Step 1: Diagnose the Current Level
Observable signals that safety is low:
- Meeting pattern: one or two people dominate, others are silent or give short answers
- Retrospectives: problems are stated vaguely or attributed to external factors, never to internal decisions
- Mistakes: people hide problems until they're unavoidable, then over-explain
- Disagreement: happens in side conversations (Slack DMs, hallway chats) but not in the room
- Questions: people ask questions privately in 1:1s that they won't ask in group settings
Observable signals that safety is high:
- People raise problems early, when they're still small
- People disagree with each other (and with you) in meetings, not just privately
- Someone says "I made a mistake" or "I don't know" without a defensive disclaimer
- New ideas come from multiple people, not just the same two people
- Postmortems focus on system failures, not individual blame
Run a quick check: In your last 5 team meetings, how many distinct voices contributed substantively? If the answer is 2-3 in a team of 8+, safety is low.
Step 2: Identify the Leader's Contribution to the Problem
Before prescribing interventions, the leader must assess their own behavior. Most low-safety environments are created by the leader, not by team dynamics.
Common leader behaviors that erode safety:
Response to mistakes:
- Does the team see you respond to failures with analysis or blame?
- When you or your team makes a mistake, do you own it publicly?
Response to disagreement:
- When someone challenges your idea, do you engage with curiosity or defend?
- Do you treat counterarguments as threats or as information?
Response to bad news:
- When someone brings you a problem, is your first reaction "thanks for telling me" or visible frustration?
- Do people wait to bring you problems until they have a solution? (This is a sign they don't feel safe raising raw problems)
Body language and signals:
- Do you check your phone or laptop in meetings?
- Do you visibly react to opinions you disagree with before the person finishes speaking?
Complete this honestly:
When was the last time I admitted a mistake to my team? [date or "can't remember"]
When was the last time I changed my position because of someone's argument? [date or "can't remember"]
When someone challenged me, what did I actually do? [describe]
What do I do when someone brings me bad news? [describe]
Step 3: Build Inclusion Safety (Foundation)
If people don't feel like they belong on the team, nothing else works.
Actions:
- Onboard new team members explicitly - make introductions, assign a peer buddy, check in weekly for the first month
- Use names in meetings. "What do you think, [name]?" signals that you know who they are and their input matters
- Acknowledge contributions publicly: "That point [name] made last week shaped how we approached this"
- Make eye contact and face people when they speak - physical attention is a signal of safety
Step 4: Build Learner Safety
People need to feel safe not knowing before they'll ask questions that surface real problems.
Actions:
- Model intellectual humility publicly: "I don't know the answer to that - let's find out" in front of the full team
- Praise questions, not just answers: "That's an important question" before answering it
- Create a "dumb question" norm - in meetings, explicitly invite the "basic" question that everyone is thinking but no one wants to ask
- Run "pre-mortems" before big decisions: "What don't we know yet that we should know?" This normalizes not knowing
Step 5: Build Contributor Safety
People need to feel safe taking initiative before they'll operate without constant guidance.
Actions:
- Give explicit permission to act: "You don't need my approval to [specific category of decisions]. Just do it and tell me."
- When someone makes a mistake while taking initiative, treat it as learning, not failure: "What did you try? What did you learn? What's the next experiment?"
- Debrief failed experiments with the same rigor as failed execution - distinguish between bad process and bad luck
- Publicly celebrate initiative, not just results: "I want to call out that [person] tried a completely new approach here, even though it didn't work as expected"
Step 6: Build Challenger Safety
This is the hardest and most valuable stage. It requires the leader to be genuinely comfortable with dissent.
Actions:
- Assign devil's advocate roles in major decisions - making it a role removes the personal risk of being "the difficult one"
- Use "two-hat" meetings: first half, generate ideas freely; second half, critique them. Keep the phases separate so people don't self-censor in the first half
- When someone challenges you, respond with questions first: "Tell me more about that" or "What are you seeing that I might be missing?"
- Change your position visibly when you're persuaded: "I was wrong about that. [Name]'s point changed my thinking." This is the highest-signal behavior for challenger safety
Step 7: Psychological Safety Health Check
Run this with the team quarterly (anonymous survey or group exercise):
Psychological Safety Check - [team] - [date]
Rate each statement 1-5 (1=never, 5=always):
1. I feel comfortable raising a problem before I have a solution.
2. I can disagree with [manager name] or senior team members without negative consequences.
3. When mistakes happen, we analyze the system, not blame the person.
4. I can say "I don't know" without it reflecting poorly on me.
5. New ideas are welcomed, even if they challenge how we currently do things.
Average score: [x/5]
Lowest-scoring statement: [this is where to focus]
One action we'll take in the next 30 days: [specific, behavioral]
Anti-Patterns
1. Confusing psychological safety with conflict avoidance
Bad: "We have a great culture - we never have conflict."
Good: Low-conflict teams are often low-safety teams where disagreement goes underground. Healthy psychological safety produces more productive conflict, not less. The goal is conflict that makes decisions better, not silence that protects them.
2. Responding to bad news with problem-solving before acknowledgment
Bad: Someone says "we're going to miss the deadline" and you immediately go to "OK what are our options."
Good: First: "Thanks for telling me early. What happened?" The problem-solving matters, but if people learn that bad news triggers immediate interrogation, they wait longer to share it next time.
3. Praising psychological safety without modeling it
Bad: Saying "I want you to speak up if you disagree" but visibly bristling when someone does.
Good: People watch behavior, not words. The leader's reaction to the first challenge in the room determines whether everyone else will challenge. One dismissive response can undo months of stated intent.
4. Safety as a morale initiative, not a performance tool
Bad: Running psychological safety workshops as an HR requirement and treating it as culture work unrelated to outcomes.
Good: Edmondson's research is unambiguous: psychological safety predicts team performance. Frame it as a performance lever, not a feelings exercise. Teams that feel safe make fewer mistakes, learn faster, and produce better outcomes.
5. Group surveys without action
Bad: Running a psychological safety survey and not acting on the results.
Good: Surveying without acting is worse than not surveying. It signals that the data doesn't matter. Every survey must produce at least one visible action. Share what you heard and what you're going to do about it.
Quality Checklist