Making sure psychological safety...stays safe

Laura de Moliere on why taking a behavioural science lens to psychological safety is critical for effective execution

In 2012, project Aristotle made a big impact in the HR world and beyond. Google had sought out to understand what makes a team effective. The answer? Psychological safety. A concept that would soon be a sought-after outcome in organisations.

Psychological safety describes the shared belief amongst individuals as to whether it is safe to engage in interpersonal risk-taking in the workplace. In other words, it describes the feeling of team members that they are able to be vulnerable in front of their colleagues, that they can fail and still be themselves without worrying about being judged or embarrassed.

Whilst many behavioural outcomes and impacts have been captured, psychological safety has been linked with people being more likely to actively seek feedback, voice concerns, and engage in open dialogue with their colleagues. Importantly, the original conceptualization of psychological safety also included a strong link with organizational and team learning, particularly from failure.

With such positive outcomes in sight, and with project Aristotle weighing in heavily it is no surprise that organisations have put psychological safety high on their agenda. But how easy is it to achieve psychological safety?

We take a behavioural science view to offer some best practice look-outs on how best to  operationalise psychological safety  in organisations: both to avoid possible negative consequences of well-meant interventions, and to point to the positive ways for behavioural scientists to start engaging with the concept. 

With this in mind, it surely goes without saying that organisations that aim to address low perceptions of psychological safety will have their unique way of doing so. However, three common challenges become apparent through applying a behavioural science lens.

1.  The problem diagnosis may not acknowledge system influences

As behavioural scientists, when starting to diagnose a problem, (such as “in our organization there seems to be very little open challenge and people are very defensive, what can we do about it?”), we would normally try to understand which behaviours we are currently seeing, and importantly, why we are seeing them (including understanding belief systems and what in turn is causing them). What exactly is driving behaviours and perceptions, what is the meaning of them? Who are the people involved in the problem, and how do they relate to one another?

One key issue with how psychological safety is often addressed in organisations is that whilst current behaviours might be described and witnessed, influences on these behaviours are not always well understood. It is perhaps all too easy to point to single out individuals that act in a manner that undermines others’ psychological safety perceptions for example, but this might come at the expense of understanding what causes their behaviours in the first place.

Psychological safety perceptions and associated behaviours across organisations are likely at least partially driven by systemic influences: be that the current performance management system, the way projects are managed, or the kind of leaders that are recruited into the business, amongst many others.

The risk is that if systemic drivers are poorly understood then interventions may not have their intended impact. Ultimately, if systems and processes in an organization do not support psychological safety (for example by ranking staff according to performance measured in outcomes) then any effort to increase safety will have a vastly reduced chance of working.  Indeed, research by Denise O‘Leary demonstrated that where there was a general lack of organizational norms on shared decision-making and a lack of stability in team membership, then psychological safety interventions had a decreased impact compared to teams where these factors were present.

In fact, interventions to achieve psychological safety are already very hard to locate. Most evidence relating to psychological safety is simply correlational and finding evidence for the causal positive impact of interventions is rare. 

In terms of designing interventions for psychological safety, while we may have concrete behaviours in mind that we can see cause low psychological safety, this alone does not address the real issue: that psychological safety is a concept that sits between individual behaviours and has many causes.

In addition, the frequently cited correlational evidence (such as project Aristotle) also means that we are trying to recreate the end states of these teams (i.e. that psychological safety is perceived to be high) rather than recreating how they got there and understanding the context in which this was achieved. While the importance of modelling leadership behaviours has long been understood, there will likely be other factors that are harder to measure: these can come from a set of actions, values and beliefs of a specific group of people rather than something any single individual can do or be on their own.

Thus, it is important to take an organization‘s particular context into account when thinking about how to improve psychological safety. Systems influences will be unique, as will be the individuals that are in the organization, but also the unwritten rules and values that dictate which behaviours associated with high psychological safety will actually lead to more effective teams (see work here on the boundary conditions of psychological safety).

2. Interventions that address individual behaviours

There is little doubt that the concept of psychological safety is appealing. The way it is often described in organizations is to focus on key behaviours that are readily observed (e.g. “there is often silence at the zoom call even though I can tell others don‘t like our proposals”).

However, the concept of psychological safety does not actually describe individual behaviours, but a shared belief that is the property of a team (i.e. that this is a safe environment to speak out and take risks). Specific and individual behaviours might not in fact be the most useful unit of analysis here, or at least they have to be brought into relation with one another to understand the bigger picture of how safety is achieved.

The varied nature of behaviours associated with psychological safety is illustrated by observational methods to assess psychological safety developed by O’Donovan, van Dunn & McAuliffe. They captured observations, for example relating to voice behaviours (defensive or silence), behaviours expressing support of one another or behaviours associated with learning and an improvement orientation. In addition, they captured these observations dependent on who people interacted with (e.g. their leader or a peer), which brings yet more nuance in understanding group dynamics.

The risk is that when the focus is on single behaviours (e.g. “we want to speak our mind when we are worried something isn‘t quite right!”) then in the absence of also addressing broader beliefs about psychological safety, this can have a detrimental impact on team atmosphere. This might result in a culture of all too quickly criticizing each other as it can quickly turn into “you are not allowing me to speak our minds when we are concerned”.

On the flip side, merely focusing on behaviours related to being supportive of one another runs the risk to create “toxic positivity”, where negative emotions such as anger are not voiced and that in turn can also have negative impacts.

Thus, it remains key to think of psychological safety as a shared belief, rather than as an individual behaviour. With this in mind it is important to understand how individual behaviours relate to psychological safety: keeping this in mind as the key outcome is important to avoid the risk of unintended behavioural consequences.

3. Power dynamics are not sensitively handled

The drive to address psychological safety is often implemented from the top – where leaders in organisations either realise that there is an issue to be addressed, or are looking for ways to improve the effectiveness of their teams. Frequently, those in leadership positions will – almost by definition – have higher psychological safety perceptions in interactions with more junior team members than vice versa (see also Katherine Grailey‘s 2020 work showing that this is a relational feature, whereby also relatively senior members can experience low psychological safety given the levels of seniority at executive levels they may need to interact with) .

This power imbalance can lead to a multitude of complications: for one, in an attempt to understand why employees currently don‘t feel psychologically safe, organisations often run focus groups. However, many times the data collected here will lack accuracy precisely due to a lack of psychological safety.

In addition, there can sometimes be an expectation that to “take part and fix things” team members should start engaging in behaviours that are related to high psychological safety: speaking their mind more freely and participating for example. However, those in lower levels of the organization have a lot more to lose and less to gain from engaging in these behaviours than those that have initiated them.

The risk when ignoring power dynamics is that these well-meant interventions can backfire, and risk employee wellbeing as a result. All too often, people have a shared history. Where a team member was once bullied or dismissed by a leader, it might be perceived as disingenuous to focus on psychological safety without acknowledging past structures and patterns. There is no blank slate from which to start psychological safety interventions.

As well as looking backward, this also shows the importance of monitoring the impacts of any interventions over the long-term, as negative unintended consequences might take longer to develop. For instance, having interventions targeted at team members to share failures more openly, might turn into problems at performance management reviews if these in turn are looking at outcomes (where the frequent sharing of failures might be a sign of a poor performance) rather than at decision-making processes and valuing learning (where sharing failures might be a sign of a positive performance).

Conclusions: What can behavioural science do?

Strangely, the concept of psychological safety can give a...false sense of safety. When low psychological safety becomes a catch-all for potential issues that are raised by employees, organisations might well believe that they have “done enough” by addressing psychological safety perceptions, whilst the real causes remain firmly in place!

This means that working on psychological safety requires us to go back to the roots of the discipline, to fully appreciate the concept to ensure it does not get watered down in a way that can potentially backfire.

This requires us as behavioural scientists to look beyond individual behaviours in order to drive outcomes that were originally associated with it – such as organizational learning.

The topic is therefore an interesting opportunity to broaden our frameworks to take account of interactions, shared beliefs and value systems. This means getting deep into the nitty-gritty of power dynamics and systems thinking to understand the impact of single interventions, including potential unintended behavioural consequences to do psychological safety justice.