Where does psychological safety collide?

A shared belief held by members of a team that the team is safe for interpersonal risk taking: one can ask a question, admit a mistake, or raise a concern without being humiliated or punished. Edmondson fixed it in 1999 in a study of 51 manufacturing teams, measured it with seven items, and showed it predicted team learning behaviour and through that team performance. The construct is a property of the team, aggregated from members and justified by within-team agreement, and its criterion is learning behaviour. This is the meaning almost everyone now means, and the one the other five get mistaken for.[1][2]

Unit

A team; members' responses aggregated after checking agreement

Recovery means

Not applicable, the criterion is learning behaviour and error reporting, not return to a baseline

Measured by

Edmondson's seven-item team scale; learning behaviour and performance as criteria

How safe it feels, and how safe it is, to speak up in a clinical unit, but the object of the safety is the patient, not the speaker. Sexton and colleagues' Safety Attitudes Questionnaire, benchmarked across 203 clinical areas in 2006, carries a safety climate factor about whether errors are handled appropriately and whether staff would feel safe being treated there. Nembhard and Edmondson showed the same year that in neonatal intensive care units psychological safety in Edmondson's meaning tracked professional status, and that leader inclusiveness (physicians inviting and appreciating input) closed the gap. The healthcare literature runs both constructs together under one phrase: the climate item asks whether care is safe, the team item asks whether the nurse is.[6][7]

Unit

A clinical unit; responses aggregated by area

Recovery means

Not applicable, the criterion is adverse events and error reporting

Measured by

SAQ safety climate and teamwork climate factors; incident reports; quality improvement participation

Employees' shared perceptions of how much management values physical safety relative to production. Zohar coined safety climate in 1980 from a survey of 20 Israeli factories, found the perceived importance of safety training and management's attitude toward safety dominated the factor structure, and showed it separated plants that safety inspectors rated well from those they rated badly. Safety culture, the broader term used for nuclear plants and other high-reliability organisations, is the same idea extended to values and practices. The safety here is physical: nobody is measuring whether it is safe to speak, they are measuring whether the organisation cares if you get hurt.[8]

Unit

An organisation or plant; shared perceptions of management priorities

Recovery means

Not applicable, the criterion is injury and accident rates

Measured by

Zohar's safety climate questionnaire; accident rates and inspector ratings as criteria

A client's feeling that the therapist and the room are safe enough to disclose and to be changed. Bordin's 1979 working alliance (agreement on goals, agreement on tasks, and an affective bond) is the construct clinicians usually mean when they say a patient feels safe, and it has decades of outcome evidence. Porges's 2011 polyvagal theory relabelled this feeling as neuroception of safety, an automatic nervous-system detection of threat or safety said to gate social engagement. The polyvagal account is contested: Grossman and Taylor argued the vagal anatomy and the evolutionary story it rests on do not hold, and the theory has not been shown to add predictive power to the alliance it borrows from. The unit here is one person's body and one relationship, not a group.[9][10][11]

Unit

A person inside a dyadic relationship

Recovery means

Return of the capacity to engage and disclose; symptom change as the criterion

Measured by

Working Alliance Inventory; heart-rate variability in the polyvagal literature

The willingness to be vulnerable to the actions of another party, based on an expectation that they will act well regardless of one's ability to monitor them. Mayer, Davis, and Schoorman fixed it in 1995 with three antecedents in the trustee: ability, benevolence, and integrity. Edmondson and Lei drew the line in 2014: trust is about whether I will give you the benefit of the doubt, held one party at a time and looking forward; psychological safety is about whether others will give me the benefit of the doubt, held across the group and felt in the moment. The two correlate and are routinely merged in surveys, and the direction of the gaze is what keeps them apart.[12][2]

Unit

A trustor's stance toward a specific trustee

Recovery means

Trust repair after a violation

Measured by

Trust and trustworthiness scales; behavioural trust games in economics

The feeling of being able to show and employ oneself without fear of negative consequences to self-image, status, or career. Schein and Bennis used the phrase in 1965 for what a training group had to provide before people would risk changing; Kahn's 1990 ethnography of a camp and an architecture firm made it one of three psychological conditions, with meaningfulness and availability, that decide whether a person engages at work. In both, safety is a state of an individual, produced by relationships, group dynamics, and management style, and its criterion is personal engagement rather than team learning. Edmondson lifted the phrase from here and moved it to the team level; the individual version never went away.[4][3][5]

Unit

An individual's experienced condition at work

Recovery means

Not applicable, the criterion is personal engagement or willingness to change

Measured by

Kahn's qualitative conditions; later individual-level safety items pooled in meta-analyses