Proven Tactics for Improving Teams’ Psychological Safety

As organizations increasingly move away from top-down management controls to more democratized leadership built around empowered, self-organized, highly agile teams, building a culture in which employees can contribute fully and honestly to constructive dialogue and decision-making is essential. It’s well known that high levels of psychological safety are required for that — but what is less obvious are the evidence-based interventions that leaders can implement to create such an environment.

At Sandoz, a Novartis division, we ran a robust randomized controlled trial that included more than 1,000 teams comprising over 7,000 individuals globally to empirically test what works, in collaboration with external academics and behavioral science consultants. While we know a lot about psychological safety and its association with desirable outcomes such as higher productivity, better performance, and increased speaking-up behaviors, there has been little causal evidence suggesting how to foster it in practice.

We were particularly interested in how psychological safety can give every employee a voice and increase the likelihood that potential ethical issues and misconduct will be reported and addressed appropriately. Our previous research found that psychological safety is one of two fundamental elements that affect whether employees feel comfortable bringing up those concerns; the other is an employee’s relationship with their front-line manager. We sought to investigate managers’ behaviors first to uncover what might be keeping people from speaking up.

The Experiment Design

Over a six-week period in late 2021, we tested the effectiveness of a minimally invasive intervention that guided managers on how to conduct one-on-one meetings with their respective team members.

Our study involved three experimental groups. The first was a control group of managers who were notified that the company was conducting a study on meeting habits but were given few other details. The second and third groups were treatment groups, where managers received emails encouraging them to hold regular one-on-one meetings with team members and to focus on psychological safety using one of two different mechanisms.

The first intervention group focused on individuation — treating employees as unique individuals. Managers were asked to encourage their employees to use their one-on-one meetings to express what was important to them and where they needed support. This treatment was based on earlier research that suggested a link between individuation and psychological safety.

In the second intervention group, managers were encouraged to use the one-on-one meetings to focus on removing blockers for their teams. Removing blockers is a process that aids prioritization by subtracting from, rather than adding to, existing practices and has been shown to increase productivity by freeing workers to focus on what matters most.

Across the six-week period, managers in the treatment groups received regular email correspondence about conducting these employee meetings. The email schedule and content were the same for both groups, with the only difference being the contents of an attachment that advised the managers on the specific treatment approach to use.

To measure the impact of the interventions, we selected two questions from the organization’s regular pulse survey that closely reflected questions Amy Edmondson has developed for assessing psychological safety: “Different perspectives are valued in my team” and “I feel safe sharing feedback with colleagues.” We measured teams’ psychological safety scores before and after the intervention period to test for an effect compared with the control group. The data was evaluated at a team level, with a minimum of five responses per team to suitably maintain anonymity.

The Results

We found that managers who treated team members as unique individuals significantly boosted team psychological safety compared with the managers of other groups. Over six weeks, psychological safety in the individuation group increased 12% more than in the control group. The group that focused on removing blockers improved too, but by a smaller margin — 6% more than the control group. Both intervention groups showed both practical and statistically significant increases at the end of the experiment.

However, upon delving deeper, we discovered nuance in our results. First, when we restricted our analysis to teams that had initially scored relatively lower on psychological safety, the effect of both treatments was especially pronounced. This suggested a diminishing returns effect when teams already felt particularly psychologically safe. The individuation treatment for teams in the lower psychological safety band improved their scores by 19% more than the control group, and the removing-blockers group improved their psychological safety by 14% compared with the control group. It was heartening to see such improvements among the employees who needed it most.

Managers who treated team members as unique individuals significantly boosted team psychological safety more than in any other group.

While it was not unexpected that different employee groups would benefit differently from the interventions, there was even more to our findings. Indeed, when focusing on teams that started out with medium to high levels of psychological safety, the removing-blockers approach was the more effective intervention. So for the teams that already had relatively high psychological safety, removing blockers improved psychological safety by 14% more compared with the control group, whereas the individuation teams improved their scores by 9% more than the control group.

Our interventions had a meaningful impact beyond increasing psychological safety. We found positive spillover effects: For the individuation group, perceptions of career development and progress improved by 21%, and perceptions of managers as role models improved by 15% compared with the control group. As expected, the interventions didn’t just improve psychological safety in isolation but positively affected multiple elements of the manager-employee relationship.

Overall, our findings demonstrate that there’s no one-size-fits-all best approach to increasing psychological safety. Rather, the most effective approach varies depending on the team’s starting point. If a team has relatively lower psychological safety, then the individual treatment appears to be more effective; but as a team’s psychological safety increases, a shift to focusing on removing blockers may be warranted. This provides good lessons more generally for organizations looking to adopt behavioral insights: Adopting generalized insights without first testing them is unlikely to result in success, and each team’s individual context matters.

Beyond the Experiment: Lessons Learned

We believe that there are additional lessons from our experiment that can be applied to better drive behavior change in organizations.

First, the experiment showed that organizations can achieve meaningful change with a simple, low-cost intervention that causes minimal disruption, as in the case of our email communication. Big, expensive awareness campaigns and workshops aren’t always needed, so if you find yourself planning a webinar on psychological safety, consider whether you could shift behaviors by leveraging existing platforms and ways of working.

Second, it showed that psychological safety doesn’t need to be promoted as a concept in order to be fostered among employees. In fact, the experiment made no explicit reference to psychological safety, and managers were not encouraged to talk about it. Rather, simply focusing on the behaviors that need to change may be sufficient; not everything is about winning hearts and minds. Our approach has the added benefit of not eating into the cognitive and attentional bandwidth of employees: They don’t have to think about psychological safety — and neither do your managers.

Third, we gained insight into how to measure employee perceptions. When comparing the responses to the pulse survey questions at the team level, we found meaningful change; however, questions about high-level, organization-wide perceptions showed little significant effect. This is a valuable lesson for corporations that rely on pulse surveys for employee insights: The differences between organization-level and team-level perceptions matter. And if you’re looking to effect change at a team level, make sure you’re measuring team-level perceptions as well.

Finally, if you want to measure how people feel in general in your organization, or gauge their levels of engagement, don’t ask “In general, do you feel that our organization…,” but do ask how they feel at the team or individual level and then aggregate the data. The results will be much more reflective of how people feel in their day-to-day work lives rather than how they feel when they pause and think about the organization in general.