1. Full citation and link #
Citation (APA style)
Edmondson, A. C. (1999). Psychological safety and learning behavior in work teams. Administrative Science Quarterly, 44(2), 350–383. https://doi.org/10.2307/2666999 Harvard Dash+1
Online access
- Publisher page and abstract:
https://journals.sagepub.com/doi/10.2307/2666999 SAGE Journals - Commonly used teaching PDF (MIT mirror):
https://web.mit.edu/curhan/www/docs/Articles/15341_Readings/Group_Performance/Edmondson%20Psychological%20safety.pdf Massachusetts Institute of Technology
2. Overview #
This article introduces the construct of team psychological safety and tests how it relates to learning behavior and performance in work teams. Psychological safety is defined as a shared belief that the team is safe for interpersonal risk taking. In a multimethod field study of 51 manufacturing teams, Edmondson shows that psychological safety is positively associated with learning behaviors such as speaking up, asking for help, and discussing errors, and that these learning behaviors mediate the relationship between psychological safety and team performance. Team efficacy, by contrast, does not predict learning once psychological safety is taken into account.
3. Core ideas #
Psychological safety as a team level belief
Psychological safety is not an individual trait. It is a group climate. Teams high in psychological safety share the belief that they can raise concerns, admit mistakes, and ask questions without fear of embarrassment, punishment, or rejection. This belief shapes what people are willing to say or do in front of colleagues.
Difference between psychological safety and team efficacy
Team efficacy is a team’s shared belief in its capability to perform tasks. Psychological safety is about feeling safe to take interpersonal risks. Edmondson finds that psychological safety predicts whether teams engage in learning behaviors, while team efficacy does not, once psychological safety is controlled for. A team can be confident and still avoid difficult conversations if the climate does not feel safe.
Learning behavior as the mechanism
The study models learning behavior as the bridge between psychological safety and performance. Teams with higher psychological safety report more behaviors such as feedback seeking, error discussion, and experimentation. These behaviors in turn are associated with higher performance ratings from external observers. Learning behavior mediates the link between psychological safety and team performance.
Antecedents: context support and leader coaching
Context support and team leader coaching emerge as important antecedents of psychological safety. Teams whose leaders coach rather than control, and whose context supports access to information and resources, show higher psychological safety and more learning behavior than teams with similar tasks but less supportive conditions.
4. Key findings #
- The study uses survey data, interviews, and observations from 51 work teams in a manufacturing company to test the model that links contextual factors, psychological safety, learning behavior, and performance.
- Statistical analyses show that team psychological safety is significantly related to team learning behavior, and that learning behavior is significantly related to performance assessed by independent observers. Team efficacy does not show the same relationship to learning once psychological safety is included.
- Mediation tests indicate that learning behavior mediates the effects of psychological safety on performance. Psychological safety is also found to mediate the effects of leader coaching and context support on learning behavior.
5. Why this matters for the endoStrategy Collective #
This article is the foundational reference for psychological safety at the team level.
For the endoStrategy Collective it provides:
- A precise, research backed definition of psychological safety, which you can use consistently in the book, workshops, and tools.
- A mechanism for how the “corporate immune system” works at team level. When safety is low, people avoid interpersonal risk, including sharing unusual ideas or disclosing struggles. When safety is high, those same inputs become raw material for learning and improvement.
- Evidence that leader behavior and local context are practical levers. This helps shift the conversation from “we need braver individuals” to “we need teams and leaders that make it safer to take risks.”
- A performance argument: psychological safety is directly tied to the learning behaviors that drive outcomes. It is not only a “nice to have” culture factor.
6. How the Collective will use this source #
- As the canonical definition of team psychological safety in research briefs, tools, and chapters.
- As a theoretical backbone when exploring masking, Rejection Sensitive Dysphoria, error reporting, and speaking up in technical or high stakes environments.
- As a key citation when stating that team climate, not individual courage, often determines whether people raise concerns, disclose mistakes, or offer unpolished ideas.
- As a bridge between neurodiversity and leadership work, because it shows how climate design can either unlock or suppress contributions from endoStrategists and other spiky profile people.
7. EndoStrategist takeaways #
- Psychological safety is a shared belief that the team is safe for interpersonal risk taking. It is about the local climate in a specific team, not general trust in the company.
- Learning behaviors such as asking for help, admitting mistakes, and trying new approaches are the pathway through which psychological safety improves team performance.
- Team efficacy is not enough. Teams can be confident and still suppress the learning conversations that would improve their work if the climate feels unsafe.
- Leader coaching and supportive context are concrete design levers. Local managers can raise or lower psychological safety through how they frame work, respond to problems, and invite or shut down input.
- For endoStrategists and neurodivergent employees, psychological safety often determines whether they contribute their full pattern spotting and integrative thinking or mask, withdraw, and let the system run on autopilot.
