The relationship between psychological safety and creative output is not primarily about whether people feel comfortable. It is about which ideas survive the pre-filtering process that operates between conception and expression. Genuinely original ideas, by definition, are uncertain, non-obvious, and potentially wrong. The implicit voice mechanisms that Detert and Edmondson documented predict that precisely this category of idea is the one most likely to be suppressed before it reaches the group.

Edmondson’s foundational research: the shared belief that interpersonal risk is safe

Edmondson’s (1999) Administrative Science Quarterly research established psychological safety as the shared belief that the team is safe for interpersonal risk-taking, and documented that it was the strongest predictor of team learning behaviour. Learning behaviour includes proposing novel ideas, acknowledging uncertainty, and raising concerns: exactly the behavioural repertoire that creative and original thinking requires. Teams with high psychological safety do more of all of these. Teams with low psychological safety suppress them in proportion to the perceived interpersonal cost.

The hospital teams research provided the most counterintuitive confirmation of the mechanism. Teams with higher psychological safety reported more medication errors than teams with lower psychological safety, not because they were making more errors but because they were reporting them. The low psychological safety teams were not safer. They were more silent. The same dynamic operates in creative work: psychological safety does not produce more original ideas but does produce more of the original ideas that exist being made available for evaluation rather than being pre-filtered before the meeting begins.

Detert and Edmondson’s implicit voice theories: the pre-filtering mechanism

Detert and Edmondson’s (2011) implicit voice theories research identified the specific psychological mechanism through which creative self-censorship operates. The implicit voice theory is the largely untested belief that speaking up with something non-obvious, uncertain, or potentially wrong carries real interpersonal risk: it will damage the relationship with a supervisor, reduce the evaluation received, mark the speaker as someone who does not understand how things work, or expose them to the embarrassment of being visibly wrong in front of peers.

These beliefs are rarely articulated explicitly. They operate as the automatic assessment that precedes the decision to speak, filtering the content before it reaches the point of expression. The ideas that survive this filter are disproportionately the ones that are safe to propose: conventional, incremental improvements on existing approaches, ideas that align with what the leader has already signalled they favour, or proposals that can be attributed to external evidence rather than to the speaker’s own uncertain judgment. Genuinely original ideas, which are by definition uncertain and non-obvious, are the category most vulnerable to the implicit voice suppression precisely because they are the category that carries the most interpersonal risk.

The post-meeting disclosure is the most widely documented evidence for this mechanism: the team member who shares after the meeting the idea they did not voice during it. The idea was present. The psychological safety to express it was not. Research consistently documents that the most novel and commercially valuable ideas are disproportionately in this category, which means that the implicit voice suppression is not filtering out bad ideas but removing the most original ones.

Amabile’s evaluation apprehension: anticipated negative evaluation destroys creative motivation

Amabile’s (1983) evaluation apprehension research established the specific cognitive mechanism through which low psychological safety impairs creative output. Anticipated negative evaluation activates the performance orientation that narrows cognitive focus from broad exploratory search toward narrow performance-relevant search. The person who is managing the risk of being evaluated negatively for their ideas is not cognitively available for the broad associative exploration that genuinely original thinking requires. They are managing a social threat.

The Dweck (2006) fixed versus growth mindset research adds the individual difference variable that interacts with the environmental psychological safety level. In a fixed mindset environment, where mistakes are treated as evidence of stable inadequacy rather than as information about the problem, the evaluation apprehension is maximised: a wrong idea is not just an incorrect solution but a signal about the person’s capability. In a growth mindset environment, where mistakes are treated as learning opportunities, the evaluation apprehension is reduced: a wrong idea is a useful data point rather than a self-revealing failure. The growth mindset culture is, among other things, a psychological safety culture.

The leader fallibility modelling mechanism: descriptive norms override injunctive declarations

Edmondson’s research consistently documented that leader behaviour is the strongest predictor of team psychological safety, and specifically that the willingness to publicly acknowledge uncertainty and mistakes is the leader behaviour with the largest effect. The mechanism is the Cialdini descriptive norm: what the leader actually does in the presence of failure and uncertainty is the social proof that communicates what is actually safe in this environment, regardless of what the leader says about psychological safety in the abstract.

The leader who says they want creative risk-taking while visibly responding negatively to failed ideas has installed a descriptive norm that overrules the injunctive declaration. The implicit voice theories will accurately read the descriptive norm as indicating that creative risk is not actually safe here, regardless of the stated commitment to it. The leader who publicly acknowledges their own failed ideas, expresses genuine curiosity about wrong turns, and responds to poor ideas with exploration rather than dismissal has installed the opposite descriptive norm, which the implicit voice theories will accurately read as indicating that creative risk is survivable here.

The Amazon written memo format is a partial structural solution to the same problem. By separating idea generation from social evaluation, the memo format reduces the implicit voice suppression that operates most powerfully in the real-time social context where interpersonal risk is immediately present. The writer alone with the document is not managing the social threat of being seen to be wrong in real time. The idea can be developed to the point where it is defensible before it enters the evaluative social context. This structural separation does not replace psychological safety but it does reduce the dependence on it for idea generation.

Google’s Project Aristotle: the commercial confirmation

Google’s Project Aristotle (2016) examined the performance of hundreds of internal teams using a range of variables including team composition, individual talent, management quality, and team norms. Psychological safety emerged as the single strongest predictor of team performance across all variables examined. The finding directly confirms Edmondson’s research at commercial scale in a high-performance knowledge work environment where the teams were composed of technically sophisticated, highly selected individuals for whom individual capability might have been expected to dominate group process variables.

The finding’s commercial significance is that psychological safety is not a wellbeing supplement to the real performance drivers. It is the primary performance driver in the knowledge work environment, and it is a designed condition that leader behaviour creates or destroys through the descriptive norms that behaviour establishes.

Books worth reading on this

The Fearless Organization by Amy Edmondson is the most research-grounded available account of psychological safety as a team-level condition, covering the empirical evidence from Edmondson’s hospital teams and subsequent commercial research, the specific leader behaviours that create and destroy it, the structural interventions that support it, and the specific mechanisms through which it predicts creative and learning outcomes. Edmondson’s account moves from the foundational research through the commercial applications with the precision that comes from being the primary researcher in the field, and her specific account of what the leader who wants to build psychological safety actually needs to do differently, rather than merely commit to in principle, is the most directly actionable available treatment of the conditions this article identifies as the prerequisite for creative risk.

If the dynamics described here are significantly affecting your wellbeing, speaking with a psychologist is the right next step. UK: Samaritans (116 123, free, 24/7). Mind (0300 123 3393). BACP: bacp.co.uk/search/Therapists. Crisis Text Line — text HOME to 741741 (US, UK, Canada, Ireland). International: internationaltherapistdirectory.com.

This article is for educational and informational purposes only. Sources: Edmondson, A.C. (1999), Psychological Safety and Learning Behavior in Work Teams, Administrative Science Quarterly, 44(2), 350-383. Detert, J.R. & Edmondson, A.C. (2011), Implicit Voice Theories: Taken-for-Granted Rules of Self-Censorship at Work, Academy of Management Journal, 54(3), 461-488. Amabile, T.M. (1983), The Social Psychology of Creativity, Springer. Dweck, C.S. (2006), Mindset, Random House. Cialdini, R.B., Reno, R.R. & Kallgren, C.A. (1990), A Focus Theory of Normative Conduct, Journal of Personality and Social Psychology, 58(6), 1015-1026. Dickerson, S.S. & Kemeny, M.E. (2004), Acute Stressors and Cortisol Responses, Psychological Bulletin, 130(3), 355-391. Edmondson, A.C. (2018), The Fearless Organization, Wiley. Syed, M. (2019), Rebel Ideas, John Murray.