Composite measure of psychological safety (observation and survey)

In short

An observation sheet for a team meeting plus a short survey. It records what people do, such as speaking up or staying silent, not only what they say they do. Use it on the teams an AI rollout touches first.

For whoever runs the assessment: an HR or operations lead, an internal project team, or an outside consultant.

What it measures

Psychological safety is the shared belief that you can speak up, ask a question or admit a mistake without being punished for it. This assessment measures it in a team, in two ways:

• Observation. A trained person sits in a team meeting and marks behaviours on a sheet: people speaking up, people going quiet, colleagues supporting or shutting down each other, the team stopping to learn from something.
• Survey. Team members answer 19 questions about the leader, their colleagues and the team as a whole.

It is the only assessment in this section that watches behaviour instead of asking about it.

How to use it

1. Get the team’s agreement first. Explain who will see the results.
2. Have a trained observer attend an ordinary meeting.
3. Keep the three questions that ask whether the meeting was typical. An observer in the room changes behaviour.
4. Report at team level only. Never name individuals.
5. Use it before and during a rollout. AI adoption has been linked to lower psychological safety and to staff hiding what they know.

The assessment comes from healthcare. Check that the behaviour groups fit your meetings before you rely on the sheet.

What it does not measure

• Anything about AI. The behaviours are general. You learn how safe the team is, not what it thinks of the tool.
• The whole organization. It measures one team at a time.
• Willingness to experiment or to take decisions. It does not cover those.

How it was tested

• The authors built both parts from earlier research and piloted them in 11 healthcare teams in Ireland and the Netherlands, 135 people in all.
• A separate team in Japan tested the survey part. Answers were consistent and related to other measures as expected. The statistical check of the survey’s structure fitted poorly, and the answers grouped into two parts: the leader, and the team.

Licence and access

The article is open access under CC BY 4.0. You may translate and adapt both parts with credit.

Alternatives

Other assessments with a similar job, and when to prefer each.

• Where do people feel unsafe to speak up: in the organization, the team or one-to-one? Perception of Psychological Safety scale (PPS) Prefer it when you cannot observe meetings and want a short survey that separates the organization, the team and one-to-one relations. Each employee · 14 questions Free to use and adapt Testing: Moderate Culture and speaking-up surveys
• How safe does each employee feel at work? Psychological Safety Inventory (PSI) Prefer it when you want a detailed survey of each person’s sense of safety at work. Each employee · 30 questions Ask the authors Testing: Moderate Culture and speaking-up surveys
• How well does the organization learn? Dimensions of the Learning Organization Questionnaire (DLOQ) Prefer it when the question is the learning culture of the whole organization. Staff, about their organization · 21 questions in the version tested in Germany Not established Testing: Strong Culture and speaking-up surveys

Practices that use it

• Protect psychological safety during an AI rollout AI projects can make staff afraid to speak up and unwilling to share what they know. Say early what AI means for jobs, and consult the people affected. Do first · Effort: medium (weeks, a few people) Evidence: Moderate Contested People and skills · Adopt

The details

Testing

Emerging Piloted in 11 healthcare teams. A Japanese study of the survey part found it reliable, but its structure differed from the original. The observation sheet has only the pilot behind it.

Looks at

Team

How the answers are collected

Observation by a trained person, plus a survey answered by team members.

Length

Observation of one team meeting (31 behaviours in 7 groups), plus a 19-question survey and 3 questions on whether the meeting was typical

Cost

Free. Needs a trained observer.

Licence

CC BY 4.0You may translate and adapt the observation sheet and the survey if you credit the authors.

Languages

English. The survey part has a Japanese version.

By

O’Donovan, van Dun and McAuliffe, 2020

Sources

1. O’Donovan, van Dun and McAuliffe (2020). Measuring psychological safety in healthcare teams: developing an observational measure to complement survey methods. BMC Medical Research Methodology. doi:10.1186/s12874-020-01066-z
2. Sasaki, Inoue, Asaoka, Sekiya, Nishi and colleagues (2022). The Survey Measure of Psychological Safety and Its Association with Mental Health and Job Performance: A Validation Study and Cross-Sectional Analysis. International Journal of Environmental Research and Public Health. doi:10.3390/ijerph19169879

Related studies on this site

• AI adoption, psychological safety and depression among employees Kim, Kim and Lee (2025). The dark side of artificial intelligence adoption: linking artificial intelligence adoption to employee depression via psychological safety and ethical leadership. Humanities and Social Sciences Communications, 12(1).
• Fear of losing a job to AI and the hiding of knowledge Kim and Kim (2024). How artificial intelligence-induced job insecurity shapes knowledge dynamics: the mitigating role of artificial intelligence self-efficacy. Journal of Innovation & Knowledge, 9(4), 100590.