Psychological Safety Questionnaire

To measure psychological safety in a team, use this questionnaire. Part A is Amy Edmondson's seven-item scale, the standard instrument. Part B places the team on the safety-and-accountability grid, because safety without standards is the failure mode critics attack. Part C records behavioral indicators, which are harder to fake than a survey.

See psychological-safety.

Evidence: Strong. The underlying construct has meta-analytic support across three decades, linking it to learning behavior, error reporting, information sharing, innovation, and performance. The seven-item scale is the most widely validated instrument here. Two honest caveats carry into practice: it is self-report, and the construct is team-level, so an organization-wide average hides the variance that actually matters.

Before you run it

• Team level only. Psychological safety varies sharply between teams in the same organization — which is good news, because a local leader can change it without waiting for the enterprise. An organizational average destroys the finding.
• Never report individual responses. Reporting them destroys the thing being measured, immediately and durably.
• Minimum five responses before reporting any result.
• Do not run it if you cannot act. Asking people whether it is safe to speak up, and then doing nothing, is itself evidence that it is not.
• Say who will see the results, in what form, before anyone answers.

Part A: The seven-item scale

Rate each 1 (strongly disagree) to 7 (strongly agree).

Items 1, 3, and 5 are reverse-scored: subtract the response from 8 before averaging. They are worded negatively on purpose, to interrupt straight-lining.

1. If you make a mistake on this team, it is often held against you. (reverse-scored)

   • Rating (1–7): [ ]

2. Members of this team are able to bring up problems and tough issues.

   • Rating (1–7): [ ]

3. People on this team sometimes reject others for being different. (reverse-scored)

   • Rating (1–7): [ ]

4. It is safe to take a risk on this team.

   • Rating (1–7): [ ]

5. It is difficult to ask other members of this team for help. (reverse-scored)

   • Rating (1–7): [ ]

6. No one on this team would deliberately act in a way that undermines my efforts.

   • Rating (1–7): [ ]

7. Working with members of this team, my unique skills and talents are valued and utilized.

   • Rating (1–7): [ ]

Team mean after reversal: [ ]

Part B: The safety and accountability grid

Psychological safety is not comfort, and safety without standards produces pleasant teams that achieve nothing. Score both axes.

Safety — use the Part A mean.

Accountability — rate 1 (none) to 7 (high):

1. This team has clear, demanding standards for the quality of its work.
   • Rating (1–7): [ ]
2. People here are expected to deliver what they committed to.
   • Rating (1–7): [ ]
3. Underperformance is addressed rather than absorbed.
   • Rating (1–7): [ ]

Accountability mean: [ ]

Place the team:

Most organizations that believe they have a safety problem are in the anxiety zone: standards without safety. Most teams that have invested in safety alone land in the comfort zone, which is what critics of the concept are pointing at when they call it soft.

Part C: Behavioral indicators

Survey scores measure perception. These measure behavior, and they are harder to fake. Collect them by observation, not by asking.

1. Speaking distribution. In your next team meeting, record who speaks and for how long. What proportion of talking time belongs to the most senior person present?

   • Percentage: [ ]

2. Basic questions. In the last month, did anyone ask a question that revealed they did not know something basic, in front of a more senior person?

   • Count: [ ]

3. Time to disclosure. When something went wrong recently, how long between the first person knowing and the team knowing?

   • Hours: [ ]

4. Senior fallibility. In the last quarter, did anyone senior say "I don't know" or "I was wrong" in a group setting?

   • Count: [ ]

5. Challenge. In the last quarter, did anyone challenge a decision made by someone more senior, in the room rather than afterwards? What happened to them?

   • Count: [ ] Consequence:

6. Dissent record. Of the last five significant decisions, how many have a recorded dissenting view?

   • Count: [ ]

Analysis

• Read the distribution, not the mean. A team mean of 5.2 made of responses ranging 2 to 7 is a different problem from one made of responses clustered at

  5. Wide spread usually means safety is conditional — safe for some people, or on some topics, and the pattern of who is excluded is the finding.

• Compare items 1 and 2 against 6 and 7. Items 1–5 measure whether risk is punished; 6 and 7 measure whether contribution is valued. A team can be non-punitive and still ignore people. The remedy differs.

• Part C item 1 is usually the most startling number available. Measure it once. Where the most senior person holds more than half the talking time, the survey scores are optimistic regardless of what they say.

• Part C item 5 is the definitive test. One challenge that was punished teaches the whole team more than a year of survey improvement. If the consequence column contains anything negative, treat the Part A score as unreliable — people are answering the way it is safe to answer.

• Beware the ceiling. Teams frequently score 5.5 to 6 on first administration because they have nothing to compare against and because answering honestly is itself an interpersonal risk. Trend across administrations is more informative than the first absolute value.

• If accountability is low, do not invest further in safety. Add clarity, standards, and follow-through instead. See google-project-aristotle, where dependability was the second-ranked factor and the one most often neglected.

• Expect reporting to rise as safety improves. Incident reports, near misses, escalated concerns, and raised risks all increase when people become willing. This looks like deterioration and is success. Agree the interpretation with leadership before the numbers move. See measurement-and-metrics.

• The leader is the intervention. Where scores are low, the highest-leverage change is the leader's response to the next piece of unwelcome news, not a team workshop. See westrum-organizational-culture-typology-questionnaire for the organizational counterpart, and just-culture for what must happen after failure.