Psychological Safety at Work: How to Measure and Improve It

Google's Project Aristotle — one of the largest studies of team effectiveness ever conducted — found that psychological safety was the single strongest predictor of whether a team performed well. Not talent density, not experience, not team size. Safety.

Despite this, most organisations either don't measure it at all, or measure it poorly. A survey question like "Do you feel safe at work?" is not measuring psychological safety — it's measuring physical safety and general comfort, which is a completely different construct.

What psychological safety actually means

The academic definition, from Amy Edmondson at Harvard Business School, is precise: psychological safety is "a shared belief held by members of a team that the team is safe for interpersonal risk-taking." This means employees believe they can speak up, ask questions, admit mistakes, offer dissenting opinions, and raise concerns — without fear of being embarrassed, dismissed, or punished.

It is specifically about interpersonal risk within a team context. It's not about job security (though those can interact). It's not about whether the office feels welcoming. It's about whether people feel they can be honest without consequences.

This distinction matters for measurement. If you ask "I feel safe to take risks at work," you'll get responses heavily influenced by job security anxiety. Edmondson's validated scale asks about very specific interpersonal behaviours instead.

How to measure psychological safety accurately

The gold standard is Edmondson's 7-item scale, which has been validated across hundreds of organisations. The items (1–5 agreement scale):

1. "If you make a mistake on this team, it is often held against you." (reverse scored)
2. "Members of this team are able to bring up problems and difficult issues."
3. "People on this team sometimes reject others for being different." (reverse scored)
4. "It is safe to take a risk on this team."
5. "It is difficult to ask other members of this team for help." (reverse scored)
6. "No one on this team would deliberately act in a way that undermines my efforts."
7. "Working with members of this team, my unique skills and talents are valued and utilised."

Note the reverse-scored items — they detect response bias from employees who click the same answer for every question without reading carefully.

If 7 questions is too many for your regular pulse survey, a valid 3-item short form covers the core construct:

• "In my team, it is safe to speak up and share concerns without fear of negative consequences."
• "If I make a mistake, it is not held against me."
• "I feel comfortable asking for help when I need it."

What low psychological safety looks like in practice

Low psychological safety is rarely dramatic. It's mostly invisible. It looks like:

• Meetings where no one disagrees with the manager
• Problems that everyone knows about but no one raises
• Post-mortems that focus on process, not on the people who made decisions
• New employees who take six months to share their first honest opinion
• High performers who leave because they felt their input wasn't valued

The invisible nature of low psychological safety is exactly why it needs to be measured directly. Leaders almost always believe their teams feel safe to speak up. The data usually disagrees.

The manager's role

Psychological safety is a team-level construct, not an organisational one. A company can have high overall psychological safety while individual teams — under specific managers — have dangerously low levels. This is why segment-level analysis matters. Aggregate scores hide the teams that need intervention.

Research is clear that managers are the primary driver of team-level psychological safety. The behaviours that build it:

• Modelling fallibility: Managers who openly admit mistakes signal that mistakes are survivable. This single behaviour has the largest effect size in the research.
• Asking questions rather than giving answers: "What do you think we should do here?" signals that the manager values input. "Here's what we're going to do" closes it off.
• Responding to bad news without punishment: How a manager responds the first time someone brings them a problem determines whether anyone brings them problems again.
• Explicitly inviting dissent: "What am I missing?" or "What's the strongest argument against this?" are specific phrases that create space for honest disagreement.

What to do when scores are low

Low psychological safety scores require a careful response. You cannot simply tell a team "you need to speak up more." If people aren't speaking up, it's because something in the environment has told them it's not safe to do so. The intervention needs to target that cause.

For managers with low-scoring teams: coaching is usually more effective than training. The manager needs to understand which specific behaviours are creating the problem — and that requires honest 360-degree input, not a generic leadership course.

For team-level patterns: consider whether there have been specific incidents that damaged safety (a public dressing-down, a colleague punished for raising a concern) and whether those need to be directly addressed rather than papered over with a team-building activity.

Tracking psychological safety scores over time — across team, manager, and tenure segment — is one of the most valuable things a people analytics function can do. TruePulse's engagement survey templates include Edmondson's validated psychological safety scale, with automatic segment breakdowns that show you exactly which teams need attention.