How do I know if people feel able to raise problems?
Measuring Whether People Will Speak Up
Psychological safety is a team-level property with a specific meaning. It is measurable, frequently misunderstood, and not the same as comfort.
Procedure
The term is used loosely to mean a pleasant atmosphere. The research meaning is narrower and more useful: whether people believe they can take interpersonal risks without penalty.
What it actually means
Can you admit a mistake without it being held against you.
Can you say you do not know.
Can you disagree with someone more senior.
Can you raise a problem without being treated as the problem.
Can you ask for help without appearing incapable.
It is not the absence of pressure, and not everyone being nice. High-performing teams with high safety have plenty of disagreement — the disagreement is about the work rather than about who is at fault.
Why it matters here
Problems surface earlier, which is the operational argument.
It determines whether any of your other measurement works. People who cannot say things out loud will not say them in a survey either.
It is the precondition for someone telling their manager they are struggling, which is the entire mechanism by which early intervention happens.
Where it is absent, wellbeing programmes measure nothing and reach nobody.
Measuring it
Established short instruments exist and the questions are consistent in substance.
If you make a mistake on this team, is it held against you?
Can you raise problems and tough issues?
Is it safe to take a risk here?
Is it easy to ask others for help?
Would anyone deliberately undermine your efforts?
Are your unique skills and contributions valued?
It is a team-level property, so aggregate at team level and expect wide variation within one organisation. Averaging it across a company discards the finding.
What drives it
Overwhelmingly, the behaviour of the most senior person present.
How they respond to bad news. One angry reaction to a problem being raised teaches a lasting lesson.
Whether they admit their own mistakes and gaps.
Whether they ask questions or only give answers.
Whether disagreement is engaged with or closed down.
Whether raising something produces a change, or disappears.
Consistency. Safety depends on predictability; a leader who is receptive some days and not others produces silence, because the risk is unquantifiable.
Building it
Respond well to the first hard thing you are told. This single moment sets the pattern for a team.
Ask specific questions rather than "any concerns?" — which reliably produces none. "What is the most frustrating part of this process?" produces answers.
Say what you got wrong, publicly, at your own level.
Separate the problem from the person after a failure, deliberately and out loud.
Thank people for raising things, particularly the inconvenient ones.
Close the loop. Report what happened as a result of what someone said.
What undermines it fastest
Punishing a messenger, even mildly, even once.
Public criticism of an individual.
Blame-focused incident reviews.
Reorganisation without explanation, which teaches people that speaking up is irrelevant to outcomes.
Ignoring something raised repeatedly, which is the slow version and is more common than the dramatic version.
The first hard thing
A team's safety level is set by a small number of moments, and usually by the first one.
Someone raises a problem, admits a mistake, or disagrees publicly.
Everyone watches what happens.
A defensive or irritated response teaches a lesson that persists for years, far beyond the individual involved.
A good response: thank them specifically, engage with the substance, separate the problem from the person out loud, and report back on what happened as a result.
The reporting back matters most. Safety is not just the absence of punishment; it is evidence that speaking produces an outcome.
New managers should be told this explicitly, because the moment arrives early and is not recognised as consequential at the time.
Asking questions that get answers
"Any concerns?" reliably produces none. Specific questions produce information.
What is the most frustrating part of this process?
What would you do differently if it were entirely up to you?
What are we pretending is fine that is not?
What did we get wrong last quarter?
If this fails, what will the reason turn out to have been?
The last one is unusually productive, because it lets people raise a risk without accusing anyone, and it surfaces concerns that have been circulating privately.