Psychological safety, born of an unexpected result
Studying medical errors, a researcher found that the best teams reported more of them. The explanation founded an entire field.
The initial observation
Amy Edmondson studied hospital teams in the 1990s, expecting the best-rated teams to make fewer medication errors.
She found the opposite: the best-rated teams reported more. The hypothesis adopted is that those teams do not make more errors but report them, which the others do not.
The result illustrates a classic measurement problem: an indicator based on reporting measures both the phenomenon and the willingness to report it. A team that reports little may be safe, or simply silent.
The notion
She defines psychological safety as the shared belief that the team is a place where interpersonal risk can be taken: asking a question, flagging a problem, admitting a mistake, without adverse consequence.
The term is not entirely new. William Kahn had used it in 1990 in his work on engagement, at the level of the individual. Edmondson's contribution is to make it a property of the group, measurable at team level.
It is neither a pleasant climate nor an absence of demands. Edmondson insists on crossing the two dimensions: a team can be demanding and safe, or permissive and unsafe.
Later work
A meta-analysis published in 2017 by Alexander Newman, Ross Donohue and Nathan Eva surveys the measured associations between psychological safety and indicators of performance, learning and reporting.
The associations are consistently found, with moderate sizes. They rest mostly on self-report measures collected from the same respondents, which limits causal interpretation: a successful team may feel safer, rather than the reverse.
The company study that popularised it
A large technology firm published in 2015 the results of an internal study of its teams, placing this notion at the top of the factors associated with collective performance.
That publication spread the term considerably. It is internal, not peer reviewed, and its methods have not been made public in detail, which calls for the usual caution about this kind of source. Yet it did more for the notion's fame than the academic work that preceded it.
The problematic use
The notion has become a management objective, with schemes designed to produce it. Evaluations of those schemes are rare and their effects poorly established, most being measured with the same questionnaires that were used to justify them.
The risk Edmondson herself flags is inversion: presenting psychological safety as a disposition of individuals rather than a property produced by managers' behaviour, which shifts responsibility onto those who do not dare speak.
Key points
The notion arose from an unexpected result about the reporting of medical errors.
It denotes neither a pleasant climate nor an absence of demands, but the possibility of interpersonal risk.
Measured associations rest mainly on self-reports from the same respondents.
This article describes documented mechanisms and disagreements. It recommends no position, no party and no vote, and commits to no reading other than that of the work it cites.
And you, where do you stand?
The Civix test situates your convictions on six dimensions and matches you with one of the twelve profiles, without ever telling you what to think.
Start my test