“Psychological safety was by far the most important of the five key dynamics we found. It's the underpinning of the other four.”

—Julia Rozovsky, “The five keys to a successful Google team.”¹

The tiny newborn twins seemed healthy enough, but their early arrival at only 27 weeks' gestation meant they were considered “high risk.” Fortunately, the medical team at the busy urban hospital where the babies were delivered included staff from the Neonatal Intensive Care Unit (NICU): a young Neonatal Nurse Practitioner named Christina Price* and a silver-haired neonatologist named Dr. Drake. As Christina looked at the babies, she was concerned. Her recent training had included, as newly established best practice, administering a medicine that promoted lung development as soon as possible for a high-risk baby. Babies born very prematurely often arrive with lungs not quite ready for fully independent breathing outside the womb. But the neonatologist had not issued an order for the medicine, called a prophylactic surfactant. Christina stepped forward to remind Dr. Drake about the surfactant and then caught herself. Last week she'd overheard him publicly berate another nurse for questioning one of his orders. She told herself that the twins would probably be fine – after all, the doctor probably had a reason for avoiding the surfactant, still considered a judgment call – and she dismissed the idea of bringing it up. Besides, he'd already turned on his heel, off for his morning rounds, white coat billowing.

Unconscious Calculators

In hesitating and then choosing not to speak up, Christina was making a quick, not entirely conscious, risk calculation – the kind of micro-assessment most of us make numerous times a day. Most likely she was not even aware that she had weighed the risk of being belittled or berated against the risk that the babies might in fact need the medication to thrive. She told herself the doctor knew better than she did, and she was not confident he would welcome her input. Inadvertently, she had done something psychologists call discounting the future – underweighting the more important issue of the patients' health, which would take some time to play out, and overweighting the importance of the doctor's possible response, which would happen immediately. Our spontaneous tendency to discount the future explains the prevalence of many unhelpful or unhealthy behaviors – whether eating that extra piece of chocolate cake or procrastinating on a challenging assignment – and the failure to speak up at work is an important and often overlooked example of this problematic tendency.

Like most people, Christina was spontaneously managing her image at work. As noted sociologist Erving Goffman argued in his seminal 1957 book, The Presentation of the Self in Everyday Life, as humans, we are constantly attempting to influence others' perceptions of us by regulating and controlling information in social interactions.² We do this both consciously and subconsciously.

Put another way, no one wakes up in the morning excited to go to work and look ignorant, incompetent, or disruptive. These are called interpersonal risks, and they are what nearly everyone seeks to avoid, not always consciously.³ In fact, most of us want to look smart, capable, or helpful in the eyes of others. No matter what our line of work, status, or gender, all of us learn how to manage interpersonal risk relatively early in life. At some point during elementary school, children start to recognize that what others think of them matters, and they learn how to lower the risk of rejection or scorn. By the time we're adults, we're usually really good at it! So good, we do it without conscious thought. Don't want to look ignorant? Don't ask questions. Don't want to look incompetent? Don't admit to mistakes or weaknesses. Don't want to be called disruptive? Don't make suggestions. While it might be acceptable at a social event to privilege looking good over making a difference, at work this tendency can lead to significant problems – ranging from thwarted innovation to poor service to, at the extreme, loss of human life. Yet avoiding behaviors that might lead others to think less of us is pretty much second nature in most workplaces.

As influential management thinker Nilofer Merchant said about her early days as an administrator at Apple, “I used to go to meetings and see the problem so clearly, when others could not.” But worrying about being “wrong,” she “kept quiet and learned to sit on my hands lest they rise up and betray me. I would rather keep my job by staying within the lines than say something and risk looking stupid.”⁴ In one study investigating employee experiences with speaking up, 85% of respondents reported at least one occasion when they felt unable to raise a concern with their bosses, even though they believed the issue was important.⁵

If you think this behavior is limited to those lower in the organization, consider the chief financial officer recruited to join the senior team of a large electronics company. Despite grave reservations about a planned acquisition of another company, the new executive said nothing. His colleagues seemed uniformly enthusiastic, and he went along with the decision. Later, when the takeover had clearly failed, the executives gathered with a consultant for a post-mortem. Each was asked to reflect on what he or she might have done to contribute to or avert the failure. The CFO, now less of an outsider, shared his earlier concerns, acknowledging that he had let the team down by not speaking up. Openly apologetic and emotional, he lamented that the others' enthusiasm had left him afraid to be “the skunk at the picnic.”

The problem with sitting on our hands and staying within the lines rather than speaking up is that although these behaviors keep us personally safe, they can make us underperform and become dissatisfied. They can also put the organization at risk. In the case of Christina and the newborns, fortunately, no immediate damage was done, but as we will see in later chapters, the fear of speaking up can lead to accidents that were in fact avoidable. Remaining silent due to fear of interpersonal risk can make the difference between life and death. Airplanes have crashed, financial institutions have fallen, and hospital patients have died unnecessarily because individuals were, for reasons having to do with the climate in which they worked, afraid to speak up. Fortunately, it doesn't have to happen.

Envisioning the Psychologically Safe Workplace

Had Christina worked in a hospital unit where she felt psychologically safe, she would not have hesitated to ask the neonatologist whether or not he thought treating the newborns with prophylactic lung medicine was warranted. Here too, she might not even be aware of making a conscious decision to speak up; it would simply seem natural to check. She would take for granted that her voice was appreciated, even if what she said didn't lead to a change in the patient's care. In a climate characterized by psychological safety – which blends trust and respect – the neonatologist might quickly agree with Christina and call the pharmacy to put in a request, or he might have explained why he thought it wasn't warranted in this case. Either way, the unit would be better off as a result. The patients would have received life-saving medication, or the team would have learned more about the subtleties of neonatal medicine. Before leaving the room, the doctor might thank Christina for her intervention. He'd be glad he could rely on her to speak up in case he slipped up, missed a detail, or was simply distracted.

Finally, as she gave the medicine to the babies, Christina might come up with the idea that the NICU could institute a protocol to make sure that that all babies who need a surfactant would get it. She might seek out her manager to make this suggestion during a break in the action. And because psychological safety exists in work groups, rather than between specific individuals (such as Christina and Dr. Drake), it's likely her nurse manager would be receptive to her suggestion.

Speaking up describes back-and-forth exchanges people have at work – from volunteering a concern in a meeting to giving feedback to a colleague. It also includes electronic communication (for example, sending an extra email to ask a coworker to clarify a particular point or seek help with a project). Valuable forms of speaking up include raising a different point of view in a conference call, asking a colleague for feedback on a report, admitting that a project is over budget or behind schedule, and so on – the myriad verbal interactions that make up the world of twenty-first century work.

There is, of course, a range of interpersonal riskiness involved in speaking up. Some cases of speaking up occur after significant trepidation; others feel reasonably straightforward and feasible. Still others simply don't occur – as in the case of Christina in the NICU – because one has weighed the risk (consciously or not) and come out on the side of silence. The free exchange of ideas, concerns or questions is routinely hindered by interpersonal fear far more often than most managers realize. This kind of fear cannot be directly seen. Silence – when voice was possible – rarely announces itself! The moment passes, and no one is the wiser except the person who held back.

I have defined psychological safety as the belief that the work environment is safe for interpersonal risk taking.⁶ The concept refers to the experience of feeling able to speak up with relevant ideas, questions, or concerns. Psychological safety is present when colleagues trust and respect each other and feel able – even obligated – to be candid.

In workplaces with psychological safety, the kinds of small and potentially consequential moments of silence experienced by Christina are far less likely. Speaking up occurs instead, facilitating the open and authentic communication that shines the light on problems, mistakes, and opportunities for improvement and increases the sharing of knowledge and ideas.

As you will see, our understanding of interpersonal risk management at work has advanced since Goffman studied the fascinating micro-dynamics of face-saving. We now know that psychological safety emerges as a property of a group, and that groups in organizations tend to have very interpersonal climates. Even in a company with a strong corporate culture, you will find pockets of both high and low psychological safety. Take, for instance, the hospital where Christina works. One patient care unit might be a place where nurses readily speak up to challenge or inquire about care decisions, while in another it feels downright impossible. These differences in workplace climate shape behavior in subtle but powerful ways.

An Accidental Discovery

As much as I'm passionate about the ideas in this book, I didn't set out to study psychological safety on purpose. As a first-year doctoral student in the process of clarifying my research interests for my eventual dissertation, I had been fortunate to join a large team studying medical error in several hospitals. This was a great way to gain research experience and to sharpen my general interest in how organizations can learn and succeed in an increasingly challenging, fast-paced world. I had long been interested in the idea of learning from mistakes for achieving excellence.

My role in the research team was to examine the effects of teamwork on medical error rates. The team had numerous experts, including physicians who could judge whether human error had occurred and trained nurse investigators who would review medical charts and interview frontline caregivers in patient care units in two hospitals to obtain error rates for each of these teams. These experts were, in effect, getting the data for what would be the dependent variable in my study – the team-level error rates. This was a great arrangement for me, for at least two reasons. First, I lacked the medical expertise to identify medical errors on my own. Second, from a research methods perspective, it meant that my survey measures of team effectiveness would not be subject to experimenter bias – the cognitive tendency for a researcher to see what she wants to see rather than what is actually there. So the independence of our data collection activities was an important strength of the study.⁷

The nurse investigators collected error data over a six-month period. During the first month, I distributed a validated instrument called the team diagnostic survey to everyone working in the study units – doctors, nurses, and clerks – slightly altering the language of the survey items to make sure they would make sense to people working in a hospital, and adding a few new items to assess people's views about making mistakes. I also spent time on the floor (in the patient care units) observing how each of the teams worked.

Going into the study, I hypothesized, not surprisingly, that the most effective teams would make the fewest errors. Of course, I had to wait six months for the data on the dependent variable (the error rates) to be fully collected. And here is where the story took an unexpected turn.

First, the good news (from a research perspective anyway). There was variance! Error rates across teams were strikingly different; indeed, there was a 10-fold difference in the number of human errors per thousand patient days (a standard measure) from the best to the worst unit on what I sincerely believed was an important performance measure. A wrong medicine dosage, for example, might be reported every three weeks on one ward but every other day on another. Likewise, the team survey data also showed significant variance. Some teams were much stronger – their members reported more mutual respect, more collaboration, more confidence in their ability to deliver great results, more satisfaction, and so on – than others.

When all of the error and survey data were compiled, I was at first thrilled. Running the statistical analysis, I immediately saw that there was a significant correlation between the independently collected error rates and the measures of team effectiveness from my survey. But then I looked closely and noticed something wrong. The direction of the correlation was exactly the opposite of what I had predicted. Better teams were apparently making more – not fewer – mistakes than less strong teams. Worse, the correlation was statistically significant. I briefly wondered how I could tell my dissertation chair the bad news. This was a problem.

No, it was a puzzle.

Did better teams really make more mistakes? I thought about the need for communication between doctors and nurses to produce safe, error-free care. The need to ask for help, to double-check each other's work to make sure, in this complex and customized work environment, that patients received the best care. I knew that great care meant that clinicians had to team up effectively. It just didn't make sense that good teamwork would lead to more errors. I wondered for a moment whether better teams got overconfident over time and then became sloppy. That might explain my perplexing result. But why else might better teams have higher error rates?

And then came the eureka moment. What if the better teams had a climate of openness that made it easier to report and discuss error? The good teams, I suddenly thought, don't make more mistakes; they report more. But having this insight was a far cry from proving it.

I decided to hire a research assistant to go out and study these patient care teams carefully, with no preconceptions. He didn't know which units had made more mistakes, or which ones scored better on the team survey. He didn't even know my new hypothesis. In research terms, he was “blind” to both the hypothesis and the previously collected data.⁸

Here is what he found. Through quiet observation and open-ended interviews about all aspects of the work environment, he discovered that the teams varied wildly in whether people felt able to talk about mistakes. And these differences were almost perfectly correlated with the detected error rates. In short, people in the better teams (as measured by my survey, but unbeknownst to the research assistant) talked openly about the risks of errors, often trying to find new ways to catch and prevent them. It would take another couple of years before I labeled this climate difference psychological safety. But the accidental finding set me off on a new and fruitful research direction: to find out how interpersonal climate might vary across groups in other workplaces, and whether it might matter for learning and speaking up in other industries – not just in healthcare.

Over the years, in studies in companies, hospitals, and even government agencies, my doctoral students and I have found that psychological safety does indeed vary, and that it matters very much for predicting both learning behavior and objective measures of performance. Today, researchers like me have conducted dozens of studies showing greater learning, performance, and even lower mortality as a result of psychological safety. In Chapter 2, I will tell you about some of the studies.

In that initial study over two decades ago, I learned that psychological safety varies across groups within hospitals. Since that time, I have replicated this finding in many industry settings. The data are consistent in this simple but interesting finding: psychological safety seems to “live” at the level of the group. In other words, in the organization where you work, it's likely that different groups have different interpersonal experiences; in some, it may be easy to speak up and bring your full self to work. In others, speaking up might be experienced as a last resort – as it did in some of the patient-care teams I studied. That's because psychological safety is very much shaped by local leaders. As I will elaborate later in this book, subsequent research has borne out my initial, accidental discovery.

Standing on Giants' Shoulders

I might have stumbled into psychological safety by accident, but understanding of its importance traces back to organizational change research in the early 1960s. Massachusetts Institute of Technology professors Edgar Schein and Warren Bennis wrote about the need for psychological safety to help people cope with the uncertainty and anxiety of organizational change in a 1965 book.⁹ Schein later noted that psychological safety was vital for helping people overcome the defensiveness and “learning anxiety” they face at work, especially when something doesn't go as they'd hoped or expected.¹⁰ Psychological safety, he argued, allows people to focus on achieving shared goals rather than on self-protection.

Later seminal work by Boston University professor William Kahn in 1990 showed how psychological safety fosters employee engagement.¹¹ Drawing from rich case studies of a summer camp and an architecture firm, Kahn explored the conditions in which people at work can engage and express themselves rather than disengage or defend themselves. Meaningfulness and psychological safety both mattered. But Kahn further noted that people are more likely to believe they'll be given the benefit of the doubt – a wonderful way to think about psychological safety – when they experience trust and respect at work.

Next, my dissertation introduced and tested the idea that psychological safety was a group-level phenomenon.¹² Building on the unexpected insights into interpersonal climate from the hospital error study, I studied 51 teams in a manufacturing company in the Midwest, measuring psychological safety on purpose this time. Published in 1999 in a leading academic journal, this research – which later influenced Google's celebrated Project Aristotle, discussed in Chapter 2 – showed that psychological safety differed substantially across teams in the company and that it enabled both team learning behaviors and team performance.¹³

A key insight from this work was that psychological safety is not a personality difference but rather a feature of the workplace that leaders can and must help create. More specifically, in every company or organization I've since studied, even some with famously strong corporate cultures, psychological safety has been found to differ substantially across groups. Nor was psychological safety the result of a random or elusive group chemistry. What was clear was that leaders in some groups had been able to effectively create the conditions for psychological safety while other leaders had not. This is true whether you're looking across floors in a hospital, teams in a factory, branches in a retail bank, or restaurants in a chain.

The results of my dissertation research bolstered my confidence that all of us are subject to subtle interpersonal risks at work that can be mitigated. Whether explicitly or implicitly, when you're at work, you're being evaluated. In a formal sense, someone higher up in the hierarchy is probably tasked with assessing your performance. But informally, peers and subordinates are sizing you up all the time. Our image is perpetually at risk. At any moment, we might come across as ignorant, incompetent, or intrusive, if we do such things as ask questions, admit mistakes, offer ideas, or criticize a plan. Unwillingness to take these small, insubstantial risks can destroy value (and often does, as you will see in Chapters 3 and 4). But they can also be overcome. People at work do not need to be crippled by interpersonal fear. It is possible to build environments, such as those showcased in Chapters 5 and 6, where people are more afraid of failing the customer than of looking bad in front of their colleagues.

Why Fear Is Not an Effective Motivator

Fear may have once acted to motivate assembly line workers on the factory floor or farm workers in the field – jobs that reward individual speed and accuracy in completing repetitive tasks. Most of us have been exposed to, and internalized, the figure of a villainous boss who rules by fear. Indeed, popular culture has exaggerated the stereotype to become comical, as in the animated Pixar film Ratatouille, where Remy the rat, the story's cartoon hero, must first overcome the tyrannical restaurant chef who rules the kitchen if he is to realize his dream of becoming a chef.

Worse, many managers – both consciously and not – still believe in the power of fear to motivate. They assume that people who are afraid (of management or of the consequences of underperforming) will work hard to avoid unpleasant consequences, and good things will happen. This might make sense if the work is straightforward and the worker is unlikely to run into any problems or have any ideas for improvement. But for jobs where learning or collaboration is required for success, fear is not an effective motivator.

Brain science has amply demonstrated that fear inhibits learning and cooperation. Early twentieth century behavioral scientist Ivan Pavlov, who housed dozens of dogs in his laboratory, found their ability to learn behavioral tasks was inhibited after they'd been frightened in the Leningrad flood of 1924. The lab workers who swam in to rescue the animals reported that water had filled the cage, with only the dogs' noses visible above water.¹⁴ Since then, neuroscientists have discovered that fear activates the amygdala, the section of the brain that is responsible for detecting threats. If you've ever felt your heart pound your palms sweat before making an important presentation, that's due to the automatic responses of your amygdala.

Fear inhibits learning. Research in neuroscience shows that fear consumes physiologic resources, diverting them from parts of the brain that manage working memory and process new information. This impairs analytic thinking, creative insight, and problem solving.¹⁵ This is why it's hard for people to do their best work when they are afraid. As a result, how psychologically safe a person feels strongly shapes the propensity to engage in learning behaviors, such as information sharing, asking for help, or experimenting. It also affects employee satisfaction. Hierarchy (or, more specifically, the fear it creates when not handled well) reduces psychological safety. Research shows that lower-status team members generally feel less safe than higher-status members. Research also shows that we are constantly assessing our relative status, monitoring how we stack up against others, again mostly subconsciously. Further, those lower in the status hierarchy experience stress in the presence of those with higher status.¹⁶

Psychological safety describes a belief that neither the formal nor informal consequences of interpersonal risks, like asking for help or admitting a failure, will be punitive. In psychologically safe environments, people believe that if they make a mistake or ask for help, others will not react badly. Instead, candor is both allowed and expected. Psychological safety exists when people feel their workplace is an environment where they can speak up, offer ideas, and ask questions without fear of being punished or embarrassed. Is this a place where new ideas are welcomed and built upon? Or picked apart and ridiculed? Will your colleagues embarrass or punish you for offering a different point of view? Will they think less of you for admitting you don't understand something?

What Psychological Safety Is Not

As more and more consultants, managers, and other observers of organizational life are talking about psychological safety, the risk of misunderstanding what the concept is all about has intensified. Here are some common misconceptions, along with clarifications.

Psychological Safety Is Not About Lowering Performance Standards

Psychological safety is not an “anything goes” environment where people are not expected to adhere to high standards or meet deadlines. It is not about becoming “comfortable” at work. This is particularly important to understand because many managers appreciate the appeal of error-reporting, help-seeking, and other proactive behavior to help their organizations learn. At the same time, they implicitly equate psychological safety with relaxing performance standards – that is, with an inability to, in their words, “hold people accountable.” This conveys a misunderstanding of the nature of the phenomenon. Psychological safety enables candor and openness and, as such, thrives in an environment of mutual respect. It means that people believe they can – and must – be forthcoming at work. In fact, psychological safety is conducive to setting ambitious goals and working toward them together. Psychological safety sets the stage for a more honest, more challenging, more collaborative, and thus also more effective work environment. As Chapter 2 will explain, researchers around the world have found that psychological safety promotes high performance in a wide range of work environments and industries. In short, as depicted in Figure 1.1, psychological safety and performance standards are two separate, equally important dimensions – both of which affect team and organizational performance in a complex interdependent environment.

When both psychological safety and performance standards are low (lower left), the workplace becomes a kind of “apathy zone.” People show up at work, but their hearts and minds are elsewhere. They choose self-protection over exertion every time. Discretionary effort might be spent perusing social media or on making each other's lives miserable.

Next, in workplaces with high psychological safety but low performance standards (upper left), people generally enjoy working with one another; they are open and collegial but not challenged by the work. Let's call this the “comfort zone.” Today, fewer workplaces around the world than ever fall into this quadrant, and it's just as well. When employees are comfortable being themselves but don't see a compelling reason to seek additional challenge, there won't be much learning or innovation – nor will there be much engagement or fulfillment.

But it's not the comfort or apathy zones that worry me most. What keeps me up at night is the lower right-hand quadrant. When performance standards are high but psychological safety is low – a situation far too common in today's workplace – employees are anxious about speaking up, and both work quality and workplace safety suffer. In Chapters 3 and 4, you will see many such workplaces. Managers in these organizations have unfortunately confused setting high standards with good management. High standards in a context where there is uncertainty or interdependence (or both) combined with a lack of psychological safety comprise a recipe for suboptimal performance. And sometimes, as you will see in the chapters ahead, it's a recipe for disaster. I call this the “anxiety zone.” Here I'm not referring to anxiety about being able to accomplish a demanding goal or about the competitive business environment but rather to interpersonal anxiety. The experience of having a question or an idea but not feeling able to share it can be deeply unsatisfying at work. And it is a serious risk factor in any company facing volatility, uncertainty, complexity, and ambiguity, or VUCA – the acronym introduced by the U.S. Army War College and widely used in the business world today.¹⁹

Finally, when standards and psychological safety are both high (upper right in Figure 1.1), I call this the learning zone. If the work is uncertain, interdependent, or both, this is also the high-performance zone. Here, people can collaborate, learn from each other, and get complex, innovative work done. In a VUCA world, high performance occurs when people are actively learning as they go.

Measuring Psychological Safety

Researchers and managers have useful tools at their disposal to measure psychological safety, and these are in the public domain. Surveys are certainly the most popular of these, and Figure 1.2 presents seven survey items, introduced in my dissertation and widely used in the research community ever since. I use a seven-point Likert scale (from strongly agree to strongly disagree) to obtain responses, but a five-point scale works as well. Note that three of the seven items are expressed positively, such that agreement indicates greater psychological safety, and three are expressed negatively (represented in Figure 1.2 with an “R” for reverse), such that disagreement is consistent with higher psychological safety. In analyzing the data, therefore, it is important to “reverse score” data from the negatively worded items, where a 1 in the data set is converted to a 7, a 7 to a 1, a 2 to a 6, and so on.

Fortunately, the psychological safety measure has proven to be robust despite variations in both the number and the wording of the items used. By robust, I mean that the collected data demonstrate the necessary statistical properties, such as inter-item reliability as measured by Chronbach's alpha and predictive validity, as measured by correlations with other variables of interest. The appendix at the back of the book shows some of the survey item variations of which I am aware. The measure has also been translated into numerous other languages, including German, Spanish, Russian, Japanese, Chinese, and Korean, all of which have yielded robust research findings.

In purely qualitative case-study research, interview data can be coded to detect the presence or absence of psychological safety. Several examples of research where this approach has been taken are found in Chapter 2. Another fruitful approach is to provide interviewees with hypothetical scenarios that fall into gray areas at work and ask them what they or their colleagues might do in that situation. When people trust that their answers will be kept confidential, they will be quite open in reporting that they would hold back unless they were extremely confident that what they want to say will be well received. Well-designed vignettes, with questions asking about how people would respond, can also be used to collect data from a larger number of employees than individual interviews will allow. I will mention examples of both approaches in Chapter 2.

Psychological Safety Is Not Enough

I do not mean to imply that psychological safety is all you need for high performance. Not even close. I like to say that psychological safety takes off the brakes that keep people from achieving what's possible. But it's not the fuel that powers the car. In any challenging industry setting, leaders have two vital tasks. One, they must build psychological safety to spur learning and avoid preventable failures; two, they must set high standards and inspire and enable people to reach them. Setting high standards remains a crucial management task. So does sharing, sharpening, and continually emphasizing a worthy purpose.

The key insight to take away from this chapter is that in most workplaces today it's simply not possible to ensure excellence by inspecting proverbial widgets. In knowledge work, excellence cannot be measured easily and simply along the way. More to the point, it's almost impossible to determine whether people have failed to hit the highest possible standards. It takes time for the results of uncertain programs to become clear, and reliably measuring good process is difficult. In other words, today's leaders must motivate people to do their very best work by inspiring them, coaching them, providing feedback, and making excellence a rewarding experience. Motivating and coaching both receive substantial attention already. What I hope you will take away from this chapter is that making the environment safe for open communication about challenges, concerns, and opportunities is one of the most important leadership responsibilities in the twenty-first century.