Reading results
Why we don’t benchmark psychological safety
Your results are never compared with other teams: no percentiles, no “teams like yours”, no league tables. Here is why, and what to compare with instead.
The first thing many people ask of a psychological safety survey is how their team compares. It is a reasonable question, and we have built the tool so that it cannot be answered. This page says why.
What a benchmark promises
A benchmark promises context: 3.8 means little on its own, but 3.8 against an industry average of 3.4 feels like a result. It promises a target. And it promises a way to show progress to someone who was not in the room.
What it does instead
It turns a conversation into a target. Once a team knows it is being compared, the number becomes the thing to manage, and the quickest way to manage it is to answer more generously. The score rises and the conditions don’t. The measurement corrupts within a cycle or two.
It invites the wrong emotions. Complacency when the number is high (“we’re above average, we’re fine”), anxiety when it is low, and competition between teams in between. None of those helps a team speak more honestly.
The comparison isn’t real. A psychological safety score depends on the exact wording of the questions, the moment, the sector, the size of the team and who happened to answer. Questions adapted to a team, as ours are meant to be, are not comparable across teams at all. A percentile built on that is a number dressed as a fact.
It hides the spread. A benchmark compares averages, and an average hides the people at the quiet end. The team “above benchmark” with two silenced members has exactly the problem a benchmark cannot see.
It points the attention outward. The question that improves a team is “what would have to change for the lowest answer here to move?”. The question a benchmark asks is “are we better than them?”. Only one of those leads anywhere.
What to compare with instead
Your own past. The second survey of the same team shows what changed, theme by theme, after the team acted on the first. That is a comparison the team controls, cannot be gamed by anyone else’s numbers, and reads as a prompt for a conversation about what happened in between.
Within the team, compare statements with each other: where people agree, where they split, which theme sits lowest. And compare reasons: a team whose quiet people mostly report futility needs something different from one whose quiet people mostly report cost.
Across an organisation
An organisation licence pools what teams report: which reasons for holding back recur, what helps people speak up, how the picture moves as teams measure again. It never shows a league table, never ranks teams, and never opens a team’s own report, including to the people paying for it. If you want to know which teams are struggling, ask them. A number that outs a team to its own management is the fastest way to make sure the next survey tells you nothing.
If someone insists on a number
Boards and funders sometimes need one. Give them what the team is doing and when it will measure again, rather than a score against strangers. Never tie a psychological safety result to a manager’s pay or rating; the moment a score decides someone’s standing, everyone learns the number is a weapon and answers accordingly. Our principles of ethical measurement set this out at length.
Questions
Is there an industry benchmark for psychological safety?
Not one worth using. Scores depend on the exact wording, the moment and who answered, so a comparison across organisations compares different things. Measure never shows percentiles, league tables or “teams like yours”. A team’s only meaningful comparison is with its own past.
What is a good psychological safety score?
There isn’t one. Look at the spread of answers, the lowest answer and the reasons people give for holding back. A high average can hide two people who feel silenced, and that is not a good result whatever the benchmark says.
Can I compare teams within my own organisation?
You can pool patterns: which reasons for holding back recur and what helps people speak. Measure’s organisation view does that without ranking teams or opening any team’s own report. Ranking teams teaches everyone to manage the number instead of the conditions.
Why does Measure not show percentiles?
Because comparison turns a conversation into a target. Once a team knows it is being compared, the quickest way up is to answer more generously, and the measurement is corrupted within a cycle or two. We would rather the report stayed honest.