Psychological Safety Is Not a Shared Understanding

Transcript

0:00

Junior: The prevailing definition of psychological safety is wrong. If you go to Google, which I have open right now, and you search psychological safety, you’ll see this, and it’s wrong. Psychological safety is the shared belief that a team or workplace is safe for interpersonal risk-taking.

0:00

Tim Clark: And that has been the prevailing definition for quite a while, Junior.

0:00

Junior: Yeah, and I think for most people out there, if you would share that definition, you’d get a lot of nods. But today, we’re going to share 1.3 million data points from our proprietary instrument, PS Index, that are going to disagree with that definition.

0:01

Tim Clark: Not only disagree, we’re going to prove to you with empirical evidence and statistical analysis that that definition is wrong.

0:01

Junior: You can have two people on the same team in the same room with the same leader having the same conversation, and one of them is speaking up constantly and challenging the status quo, and the other has never asked a question ever. So does that team have psychological safety or not?

0:01

Tim Clark: Good question.

0:01

Junior: It’s a trick question. One person might, another person might not.

0:01

Tim Clark: How shared is that?

0:01

Junior: Exactly. Psychological safety, it’s not in our estimation. In the proof that we’ll get into the data, we’ve got a new data set, well, a new analysis for you today that I think you’re going to find quite interesting. And it supports the fact that psychological safety is an individual perception. It is not a shared team property. And the difference is not academic. It decides what you measure, what you’re going to fix, and who you’re going to fix it for. The distinction between shared understanding and individual perception is where this all hinges, and that’s what we’re going to be exploring today.

0:02

Tim Clark: Think about the implications, Jr. to the leader. If the leader understands that it’s based on individual perception and that it is not a shared condition or reality, it makes all the difference.

0:02

Junior: So we’re going to get to the invitation to stop behaving based on the average, and we’re going to start talking about the individual. So why this episode in the first place? We recently launched the four stages of psychological safety second edition, which has, I mean, it’s doubled in size. There’s a whole bunch of new data. So all of the data that we gathered through PS Index is represented to prove out the empirical pattern of stage one inclusion safety to learner, to contributor, to challenger.

0:03

Tim Clark: That’s right.

0:03

Junior: So that is no longer our idea, our theory, our framework that we thought would be really nice. It’s empirically proven with the data that’s shared in the four stages. Now, we’ve also taken that PS Index data and run additional analysis to see if there were any patterns in there that we might find interesting. During the launch webinar of the four stages of psychological safety second edition, we got this question many times.

0:03

Tim Clark: We’ve had it many times over the years, over the past five years.

0:03

Junior: How does your four stages model work with the prevailing definition of psychological safety? The answer is that it doesn’t, because the definition is wrong. The prevailing definition is wrong.

0:03

Tim Clark: That’s right.

0:03

Junior: So as Jillian said on Monday, a lot of people are asking about this, and we get it on LinkedIn as well. There’s a lot of chatter even in the last week. So the field’s most cited definition is that it’s a shared belief held by a team. It’s a team level climate that you either have or you don’t have. And we’re grateful for that definition. We’re grateful for all of the research that’s happened to this point because it’s helped bring the spotlight to psychological safety. And we’ve certainly seen over the last 12 months that this is not a concept that’s going anywhere. But the four stages definition is different. The four stages school is this psychological safety as an individual’s perception of the social risks, the risk of acts of vulnerability that are different person to person, that are different moment to moment, developing through four stages.

0:04

Tim Clark: And then most importantly, they are different stage to stage based on that linear progression that is empirically founded. So that’s what matters. We will still continue to measure psychological safety at a team level, at an organizational level. We will aggregate data. We will generate mean scores and standard deviations. And those are very important to help us understand relatively speaking where a team or an organization falls. Because as we know, psychological safety is not a binary proposition. It’s a matter of degree. And so we can measure the construct at a team level or at an organizational level, but it is not experienced at a team level or an organizational level.

0:05

Junior: I want to jump right to the data. Is that okay with you?

0:05

Tim Clark: Let’s do it.

0:05

Junior: So Leader Factor runs PS Index. It’s our assessment to measure psychological safety. And the data set that we looked at just this last week has over 1.3 million data points in it from 81,504 responses across 513 organizations.

0:06

Tim Clark: Pretty good data set.

0:06

Junior: It’s a pretty healthy data set. It’s more than healthy. It’s very robust.

0:06

Tim Clark: And statistically projectable to a much larger population.

0:06

Junior: Yeah. So what we were looking at in this specific analysis was intra-team variance.

0:06

Tim Clark: That’s right.

0:06

Junior: So I’m going to show you what kind of led to this episode. Because Jillian brought this question and said, “Hey, we keep getting this, right? What do we do?” And I started thinking, “Well, the only thing that has to be done is prove significant intra-team variance. That’s all we have to do. And we know that it’s in the data, so let’s just go pull the data.” But when we pulled the data, we were even—we were surprised.

0:06

Tim Clark: More surprised.

0:06

Junior: And let me show you what we were surprised by. Members of the same team rate the same environment differently. That’s fairly obvious, but here are the numbers. The teams—the teams cut of the data is 10,432 intact teams of three or more people.

0:07

Tim Clark: So over 10,000 teams.

0:07

Junior: 10,000 individual teams. On four out of five teams, the same question is answered nine or 10 by one member and lower than six by a teammate.

0:07

Tim Clark: Now, we’re using an 11-point scale that goes from zero to 10. So just for context.

0:07

Junior: The median team’s highest and lowest rater sit 2.5 points apart on that zero to 10 scale. So that’s the intra-team variance 2.5. An inner quartile range of 1.4 to 4.2. It’s a huge range. On one team in three, the gap exceeds a third of the entire scale. On one in six teams, it’s five plus points. So 17% of teams had a gap of five plus points in the responses from people that are answering the same question that are on the same team.

0:08

Junior: I know we’re going to be talking about a lot of numbers. We’re going to be sharing a whole bunch of data. This isn’t a downloadable, and this is going to be a really good one.

0:08

Tim Clark: Yeah, it is.

0:08

Junior: So check that out because if you can’t follow the exact train, you can look at that as reference. And then one team in four, 26.8%, contains both a blue member, nine or higher composite. So if you talk about red zone, blue zone, for those of you who are familiar with the instruments reporting, and a red member, six or lower, a promoter and a detractor in the very same room.

0:09

Tim Clark: Now that’s an amazing finding, Junior. And that happens three out of four times.

0:09

Tim Clark: You have a promoter. Now we’re thinking about net promoter score and that methodology. You have a promoter and a detractor on the same team based on their perception of the level of psychological safety.

0:09

Junior: Yep. So three out of four, you’re going to have that big variance. And one out of four, you’re going to have red zone and blue zone composite, net promoter, detractor, or promoter, as you said. So then we can talk a little bit about the intercorrelation coefficient. And this is the one that

0:09

Tim Clark: really jumped out at you.

0:09

Junior: Yeah. So tell me about this.

0:09

Tim Clark: So in statistics, one of the statistics that we, a statistical test that we perform is what we call an intra-class correlation coefficient or ICC for short. And what it does is it measures the level of correlation among the responses within a group, within a team. And how many teams are we looking at? 10,432. The intra-class correlation coefficient turns out to be 0.16. Now, for reference, this is a correlation coefficient whose value ranges from zero to one. So that’s what we’re dealing with, zero to one. That’s the scale. An ICC of 0.16 is on the floor. Anything below 0.5, Jr. is considered very low intra-class correlation. And this is 0.16. So it is incredibly low. And again, our data set is over 10,000 teams, 81,504 completed teams. And then four completed responses.

0:11

Junior: So practically speaking, we’re in the same room. We have the same manager. We’re doing the same work. You’re having one experience and I’m having an experience that’s entirely different. And all those variables combined give us the most similar experience we could engineer.

0:11

Tim Clark: Yeah.

0:11

Junior: In a professional setting, you can’t get more close than that. But if you’re a shared understanding, then what would we expect that ICC to be?

0:11

Tim Clark: At least above 0.5. At least to make any claim at all. And if it were like, for example, 0.55, that’s still pretty low. That’s a moderate intra-class correlation coefficient. So you would expect that it would be, I’d say minimum 0.6 or above. We are nowhere near that. And so it clearly disproves the prevailing definition. Now, as I said before, we’re going to go ahead and continue to measure psychological safety at the team level, to understand mean scores, to understand standard deviations. To preserve anonymity. Yeah. But to say that it is a shared understanding is absolutely not the case. That is not the construct.

0:12

Junior: Let’s talk about SD for a minute. Standard deviation. This is also really cool to see across the four stages and helped reinforce to me why it’s so important that there be four. Standard deviation increases as we go through the stages. So your SD gets higher as we approach challenger safety. It’s two at challenger. Our SD is two at challenger. So person level SD runs from 1.6 to about 1.7 across the first three stages. It’s in that range. And then it jumps to two at challenger,

0:13

Tim Clark: Yes.

0:13

Junior: which is really interesting. Why is that interesting to you? Well, why the jump?

0:13

Tim Clark: The jump, it corresponds with our interpretation that the nature of vulnerability shifts in a fundamental way. When you get to stage four, challenger safety. Why? Because we’re talking about acts of vulnerability that challenge the status quo. Engaging in divergent thinking. Offering up contrary points of view. Engaging in all kinds of divergent thinking that take the status quo to task. And so the nature of vulnerability would suggest that it’s going to be hard. It’s going to be more difficult to establish challenger safety. The perception that I can safely challenge the status quo and that doing so is a rational choice rather than an act of courage. And so we’re not surprised that there’s a step change in the standard deviation and that it’s going up to 2.01.

0:14

Junior: So what that means practically is that if you and I are on the same team, our disagreement in the perception of how psychologically safe the environment is will vary the most. When the risk is the highest.

0:14

Tim Clark: That’s right.

0:14

Junior: Okay. Let’s get into another piece of the data that we found to be pretty compelling. When you split the variance, only about 16 percent of the variation in psychological safety ratings is between teams. The other 84 percent lives inside teams.

0:14

Tim Clark: Amazing.

0:14

Junior: So we talked about the ICC, but when you take that variation and you pin it not intra-team but team to team, not a lot of variance is explained. They actually don’t move as much as we would expect.

0:15

Tim Clark: Right.

0:15

Junior: They’re mostly similar when you look at averages. And so when you peek inside the team, that’s where the variation spikes.

0:15

Tim Clark: Now, based on the prevailing definition that psychological safety is a shared understanding of interpersonal risk taking, you would think it would be the very opposite. Because if it’s shared, then most of the variance should be explained between teams, not within teams.

0:15

Junior: So in short, a team doesn’t have one psychological safety. It has one per person. That’s right. What we’re not saying teams do differ. So team level agreement in the data runs a median RWG of 0.87 to 0.95, which is above the point seven agreement bar. So on their own test, these teams agree. The claim isn’t about proportion, not absence. So we’re all in the same room. We’re experiencing the same thing or in

0:16

Tim Clark: terms of the reality of what’s happening. But the perception of what happening is completely different. And what it means to you and what you would be punished or rewarded for is very different.

0:16

Junior: And one of the principles we want to share is that you perceive before you understand and that that affects our behavior.

0:16

Tim Clark: That’s right.

0:16

Junior: So we were talking through this and you said, oh yeah, that’s effective primacy.

0:16

Tim Clark: What’s that? In psychology, we call it affective primacy. That means that often your emotional perceptions kick in first and then your cognition kicks in second. Does that make sense? So think about when you feel something before you think something and then the feeling informs the thinking. So the feeling brain is going first and then the thinking brain is going second. That often is the case.

0:17

Junior: If you look at shared understanding, is more like us speaking it into existence than it actually being there. So we could ask what would a team have to do to share quote unquote an understanding of risk? There’s a whole bunch of stuff that would have to happen cognitively to even arrive at that point of understanding. Right.

0:17

Tim Clark: Yeah.

0:17

Junior: And so if we go into affective primacy, we know, well, a whole bunch of behavior probably already happened because of our emotional perception before we could possibly arrive at a shared understanding. There’s too much work that would have to be done to get to shared understanding. We would lose before we got behavior that was already out there. We can have it.

0:17

Tim Clark: I don’t think we’ll ever get there, Junior, because all of that has to come through different lenses and different filters that we each have. And they’re never going to come into alignment ever, ever. They may come into semi alignment, but never, never total alignment to where it is actually shared.

0:18

Junior: So the unit of analysis has to be a person. And so broadly speaking, if we don’t have variance information, we’re going to be missing a big part of the picture. And so this has spurred some interesting insight for us about how we might change some of the reporting features in PS Index, even as it stands, to display some of the ICC, to display some of these measurements of variance that would be interesting and maybe actionable for us. So why is this the case? Why could it be that Tim and I are in the same room, yet we perceive the psychological safety to be different? First thing is hierarchy, right? So if you sit on a different place in the hierarchy than I do, that’s going to be a big factor. We have relationship asymmetry. So the calibrating that each person is doing against another person relative to maybe who’s presenting or who asked the question or who spoke up, that’s happening all the time. And as you said, when is that going to be in perfect alignment across an entire team?

0:19

Tim Clark: Never. Right. Never. So there’s asymmetry in the relationship. That means there’s a power gradient. That means that based on status and role, we are not equal. Yeah. Does that bear on perception? Absolutely.

0:19

Junior: And then we have the ladder of vulnerability data, which I think plays into this, because we know that there’s different levels of carried risk for every act of vulnerability, depending on who you are. And that’s a function of a whole host of things. Yeah. We don’t know the exact math. But if we look at the ladder, we know that let’s say we take an act of vulnerability that is asking for more resources. Right. Maybe that is a really easy thing for me to do. Maybe that’s a really difficult thing for you to do.

0:20

Tim Clark: That’s right.

0:20

Junior: And so if there’s a question about that happening inside of a meeting, are we going to be perceiving the risk the same way?

0:20

Tim Clark: No.

0:20

Junior: Is showing up different for you than it is for me? Is asking a question different? Is making a mistake? Voicing a concern? All of those things carry different levels of risk. And that’s borne out in the ladder of vulnerability data. We know that for sure.

0:20

Tim Clark: That’s right.

0:20

Junior: It’s not the same across the human race.

0:20

Tim Clark: It’s not the same. No. The carried risk of each of those acts of vulnerability. There’s 20, right? January, we have 20. The 20 most common in the ladder of vulnerability. Those are the ones that we measure. Yeah. And it’s absolutely amazing to see the variation in the carried risk that one person has versus another in each of those.

0:20

Junior: Let’s take another three. So tenure. Is tenure going to affect the level of risk that we feel in a meeting?

0:21

Tim Clark: Absolutely.

0:21

Junior: Absolutely. Okay. Let’s take stakes. So let’s say we’re debating an issue and it’s highly impactful to my function and not so much to yours. Are we going to perceive risk the same way?

0:21

Tim Clark: Yeah.

0:21

Junior: Yeah. No, of course we’re not. What about stress?

0:21

Tim Clark: Maybe I just had a bad day. Yeah.

0:21

Junior: Maybe you had a great day and we’re in the same place in the same situation. Are we going to perceive things the exact same way?

0:21

Tim Clark: Yeah. No. What about role? What about personality? What about socialization? What about upbringing? What about your demographics? What about gender? What about educational level? What about all of these things they bear on your perception?

0:21

Junior: Another way of saying it is that all of those unique variables would have to be precisely the same in order for our perception of an environment to be precisely the same.

0:21

Tim Clark: But even if it’s not precisely the same junior, again, if we go back to the ICC, a correlation coefficient, if it were like 0.6, okay, then we concede and we’re going to acknowledge that it may not be a perfectly shared understanding of interpersonal risk taking, but it’s largely shared. It’s not even that.

0:22

Junior: Yeah. I’m glad you called that out. So what do we do with this? How do we operationalize this finding? If I’m a leader and I’m listening to these numbers, cool. It makes sense. It’s logical, but what do I go do differently tomorrow? One of the things that I’ve been rolling around in my brain over the last couple of days is this idea of reading a room or managing a room and managing a team. And we often use those words singularly. We’ll talk about a room as if it’s one thing. Yeah. A room is like eight things if there are eight people in it. And so when you’re reading a room, yeah, of course, you’re going to get a general feel for what’s going on, but it has to be

0:23

Tim Clark: Sure.

0:23

Junior: very specific in order for us to manage to the individual. And so if I know going in that there are eight different experiences being had, if I say the exact same words, I might add something here and take something away here. That’s a good question to this person in particular, because I know X, Y, or Z. It becomes much more specific. We’re not managing to the average. Let’s say that I go into an organization that has a team that is BlueZone. And so I read their PS index scores and I think, well, this team is great. This is fantastic. Everybody’s good. But there’s an outlier. There’s one where the spread

0:23

Tim Clark: was huge and we have a two. You may look at those average scores, but you got to look at the standard deviation. And that’s going to tell you.

0:23

Junior: And if you have a detractor like we talked about in the data, then I have to go in understanding that there’s somebody here who’s having an experience that’s less than ideal and I need to pay attention.

0:24

Tim Clark: Here’s the big epiphany for, I hope, for the average leader when they understand that psychological safety is individually perceived and it’s not a shared understanding of interpersonal risk taking. The big epiphany is that the imperative for you as a leader is to increase your powers of observation and focus on the individual and see the individuals in the room, see the individual experiences that are taking place in the room and respond to people that way. As individuals, you’ll become much more effective as you are focusing on each individual one at a time.

0:24

Junior: We tried to synthesize the real practical piece of the episode and this is what we came up with. You don’t install a shared climate. You lower the felt risk of vulnerability person by person until speaking up is the rational choice.

0:25

Tim Clark: That’s right.

0:25

Junior: That’s how we would summarize all of the operational piece of this conversation. So you’re not going in and as a unit installing a shared climate. That’s not how it works because we know that it’s individual. So you have to lower the felt risk, the perceived risk of vulnerability across a whole variety of acts of vulnerability for each person.

0:25

Tim Clark: Each person. So you’re going to ask the question, how do I lower the felt and carried risk of vulnerability for this person and this person and the calculation, the conclusions that you come to and the implications that are going to be different for each person, each member of your team. You have to ask it at an individual level.

0:26

Junior: What are some of the important takeaways that you would share with people as we wrap up the conversation and point them to the downloadable? What do you hope they walk away with?

0:26

Tim Clark: I hope that this is a massive aha moment for you as a leader and that you recognize that even though you have a team, each member of your team is having their own experience. They perceive psychological safety differently. They bring their own filters. They bring their own lenses. They bring their own experiences to your team. You need to take all of that into consideration and think about how you can reduce the carried risk of vulnerability for that individual. And then the next one, it changes your paradigm as a leader. That’s what it does. Yep. Couldn’t agree more. It was really fascinating to look at the data. We’re going to package it, present it to you in a format that I hope is really useful. So I would encourage you to go and look at the actual data in the downloadable that we’ll be putting out there. I really believe that you’ll find it valuable. It struck us. I mean, we’re in this every single day.

0:27

Junior: We’ve run, well, I can tell you how many surveys. A lot of that is represented in this dataset. And yet this was truly insightful. And so we’re grateful for the analysis that we have access to and hope that you enjoy it as well. And with that, we will see you in the next episode. Like and subscribe. We’ll see in the next one. Bye bye.