Our engagement survey came back. Psychological safety: 4.2 out of 5. Leadership celebrated.
Same week we held an incident postmortem. Forty minutes of silence. Nobody named the deploy shortcut everyone knew about. An hour later a junior DM’d me the real root cause. She would never say it in that room.
I do user research, and that DM broke my faith in the instrument, because a survey measures willingness to answer surveys, and the least safe people are least likely to tell the truth on one.
Google’s Project Aristotle ranked safety number one of five. So we all measure it. But what if the Likert scale just measures fear of the scale?
How do you actually know your team is safe? Is the whole instrument broken?