The notion of psychological safety is an idea first developed in the 1950s that has been extended in important and careful ways in recent years, especially by Amy Edmondson of Harvard Business School. The core of the idea is that a context that is psychologically safe is one where people feel they won’t experience interpersonal harm — being ridiculed or otherwise personally attacked — if they try to speak up, make mistakes, take risks, or ask for help. There is a lot of very strong evidence that creativity, learning, and exploration are better where psychological safety is higher.
But is psychological safety always good?
One issue concerns jobs that consist of repetitive and standardized tasks where we want employees to follow rules and not take risks, especially where we cannot afford mistakes. For example, there are studies showing that if you want nurses to help figure out how to improve practices, psychological safety matters a lot because it can be risky to put forward new ideas. But once nurses go out on the hospital floor to take care of patients in their regular job, we do not want them taking trying to innovate and take risks. The limited research on psychological safety in contexts where taking risks is not part of the job shows inconsistent results.
The second issue is, how much psychological safety is useful? This had not been studied and was the focus of research that we conducted with Michael Hodor of Tel Aviv University. Previous studies had focused only on the average effects, which, of course, are made up of high scores and low scores. But psychological safety is not an either/or outcome; it is a question of degree. Situations where it may be very useful can mask other situations where it may actually be harmful if we are only looking at averages.
We addressed both issues — what does psychological safety look like in typical jobs and how much is useful — in five independent settings. The jobs spanned frontline positions in retail, nursing, finance, software, pharmaceuticals, and the start-up world. The good news for fans of psychological safety is that we find that even doing routine work that is not creative, job performance improves when you move from up from low levels of psychological safety — what you might think of as unsafe situations — to average levels. That should not be too surprising because jobs that are really unsafe — where you feel you cannot ask for help, express your opinion, and so forth — are likely to have a lot of things wrong that affect job performance.
But when you move from average to high levels of psychological safety, we found that performance in these typical jobs actually declined. We might think that too much of anything has diminishing returns, but in this case, the returns actually go negative: required job performance gets worse.
Why would that be? As we suggest, it could be that an environment where bosses push the “no bad ideas here!” view encourages people to take risks when they shouldn’t. But there is a more fundamental issue linked to the aspect of psychological safety that says, “Mistakes will not be held against me.” While it makes perfect sense to be tolerant of mistakes when we want people to take risks, that does not make sense in most jobs most of the time where we want people to follow rules, conform to standards, and not take risks. There should be consequences for poor performance, as psychological safety experts have long noted. But that nuance is all too often overlooked: Consultants and managers push psychological safety way too broadly — i.e., for all kinds of jobs, including ones consisting of routine tasks — and then frontline supervisors, reacting to what they are being told by their managers, end up being too forgiving when workers make mistakes that should not be tolerated.
Suppose you had a team whose members reported that there was almost zero chance that they would be held accountable for making mistakes. Psychological safety would be rated as very high there. That is fine if your job requires taking risk; it’s not fine if mistakes matter.
Evidence supporting the view that high levels of psychological safety are associated with contexts where being held accountable for mistakes is low comes from another workplace characteristic: collective accountability, where employees feel they are together as a group responsible for performance even if they are not held accountable as individuals. When that collective accountability is very strong, it can offset the negative effects we otherwise see on job performance from high levels of psychological safety. But achieving high levels of collective accountability is just as much of a challenge — and perhaps an even bigger one — than getting high levels of psychological safety.
A practical problem for employers that is hard to duck is that supervisors at all levels are constantly being pressed by their leaders to be tougher about job-performance problems and to hold employees accountable for them. Pressing them at the same time to create psychologically safe environments where employees feel mistakes will not be held against them is very hard to pull off. Pushing them to drive psychological safety ever higher — past an average level — pulls them in two difficult directions.
What do we take away from this? First, a reminder that psychological safety is a concept focused on situations where we want to encourage taking actions that involve some risk of failure, such as learning a new procedure or brainstorming new ideas.
Second, even in regular jobs, situations where psychological safety is very low is a bad thing. We should identify those situations and figure out how to fix them, recognizing that this may not be an easy task because many different practices and behaviors can cause them. Some aspects of psychological safety may be more important to an organization than others and can be tracked separately — such as tolerance of individuals’ differences. It may be easier to move the needle on those aspects by tracking them separately as goals to improve.
Third, there is also something wrong with situations where psychological safety in jobs consisting of routines, or standardized tasks, is very high. Again, that’s because high levels of psychological safety hurt the performance of such jobs.
Fourth, although collective accountability can serve as a safeguard against the dangers from very high levels of psychological safety, achieving a high level of the former is difficult. It is a simpler option to not have a very high level of the latter.
More psychological safety in the workplace, unfortunately, may not always make everything better.
by Peter Cappelli and Liat Eldor
Harvard Business Review- January 2024