The Safety of Work

Ep.27 What Makes Teams Effective?

Episode Summary

In this episode, we discuss what elements make an effective team. We define “teams” and explain how research on teams is performed.

Episode Notes

We use the paper, Embracing Complexity, to frame our discussion. Tune in to hear our chat about this important issue.





“A topic that comes up a lot in the research is virtual teams. Who would have guessed that teams meeting over Zoom was going to be a topical and relevant hot-button topic?”

“...The research suggests that functional diversity, as well as individual educational diversity have positive relationships with team performance.”

“There were some studies that said if there is a general climate in the organization around innovation, then the team will display more innovative characteristics and things like that.”



Mathieu, J. E., Gallagher, P. T., Domingo, M. A., & Klock, E. A. (2019). Embracing Complexity: Reviewing the past decade of team effectiveness research. Annual Review of Organizational Psychology and Organizational Behavior, 6, 17-46.

Episode Transcription

David: You're listening to the Safety of Work podcast, episode 27. Today, we're asking the question, what makes a team effective? Let's get started. Hi, everybody. My name's David Provan and I'm here with Drew Rae. We're from the Safety Science Innovation Lab at Griffith University. Welcome to the Safety of Work podcast. If this is your first time listening, then thanks for coming.

The podcast is produced every week and the show notes can be found at In each episode, we ask an important question in relation to the safety of work or the work of safety and we examine the evidence surrounding it. Drew, what's today's question?

Drew: David, our question for today is pretty simple. It's, what makes a team effective? Because the question is simple, I don't think we need a lot of background on the question itself. What we wanted to do with this episode, though, was find a paper that could summarize all of the recent research on a topic outside of safety, so that we could then look at what's the most interesting in the latest research out there on teams and then talk about how that matters for safety.

I thought there was no background at all needed. David, you've snuck into our notes that we need to define what a team is. When we talk about a team, we're not talking about individuals, we're talking about more than one individual. We're talking about people who interact socially at work, either face-to-face or virtually in some way, who have some common goal, and are brought together to work together on the accomplishment of tasks. We need to recognize that that happens inside a broader organizational system. If your team is your entire organization, then there's not much point in studying the team. You might as well just study the organization.

When it comes to researching teams, the main idea here is just what unit do we study. We could look at whole organizations, we could look at individuals, or we can look at this intermediate step where we have this small dynamic building block that we can study as a single thing. Recognizing that it's part of something bigger, then not focusing too much on that bigger context, focusing on the dynamics within the team.

The other question we need to quickly talk about is what does it mean for a team to be effective? We're actually taking this fairly straight from the paper we're going to study today. There are two sorts of outcomes that a team can have. There are very tangible outcomes. What are the deliverables that the team needs to produce? We can measure those things in terms of productivity, efficiency, quality of the outputs, or there are influences on the team members. How does it affect the relationships in the team or how does it individually affect the attitudes, reactions, learning, and behaviors of people with the team?

To bring all that back, we're asking what makes a team effective, we've defined what a team is, we've defined somebody with what we think effectiveness is. Today, do you want to tell us a little bit about the paper itself?

David: Yes. The paper that you found, Drew, was titled Embracing Complexity: Reviewing the Past Decade of Team Effectiveness Research. This paper was published in the Annual Review of Organizational Psychology and Organizational Behavior. It's a journal that pretty much only publishes review articles and it only publishes a few articles each year. Each article that it publishes aims to capture the state of knowledge on a particular topic. These papers tend to be very highly cited and they tend to be very highly regarded. 

Drew, when you found this paper, you did a bit of research into the authors. Do you want to tell our listeners a little bit about who the authors were? This was a large review to try and look at a decade of research into team effectiveness.

Drew: Yes. Since we've talked on the podcast before about why we talk about the authors, it's not because being the right author automatically means we believe everything you say. But there's a lot that we need to trust the authors about that they've selected the right research, that they've summarized it appropriately, and that they've given us a fair interpretation. In this particular case, the authors are all from the same department. It's the management department of the University of Connecticut. They're John Mathieu, Peter Gallagher, Monique Domingo, and Elizabeth Klock.

The first author, John Mathieu, seems a bit like the research equivalent of a rock star who's also anonymous. You expect when someone's researching leadership or teams that they're going to be doing TED talks, that they're going to be publishing books on how to make your team more effective, how to be a good team leader. As far as I can tell, this guy's done none of that stuff. But if you look at his academic work, it's received tens of thousands of citations. Maybe the people with the big exciting ideas give the TED talks and the real down and gritty experts prepare annual review articles.

The other authors, as far as I can tell, they appear to be either grad students or research assistants of the first author. This is like a team of junior people together with a leader putting together a summary of the literature.

David: I think the review started (from memory), with about 1500 articles—you'll talk a little bit about this style of review in a second—and then, they ended up cutting that down to 500 or 600 which they reviewed in detail. You're going to need a little bit of a team to read through all of those papers to actually pull out the information that you need to publish a comprehensive review like this one.

Drew: It's not just for reading the individual papers, but also for knowing which papers to include and which papers to ignore. I'll describe this paper probably best as a non-systematic review but fairly thorough. In a totally systematic review, you're using the human equivalent of a computer algorithm to decide what's in and out. You're searching for particular keywords in particular places and with very strict rules. This one tends to lean more heavily on the expertise of the researchers to decide what's relevant and important. It explicitly focuses on a few key quality journals rather than searching big databases and it only incorporates stuff outside of those journals selectively.

When the research team just decided that it mattered, they knew that on this particular topic there was something interesting, published something else. There's a strong focus on finding meta-analysis articles. On working out what the questions are and then saying, has someone published a comprehensive answer to this question?

David: Drew, can you give us an overview of those four categories of things that affect team performance?

Drew: Sure. The first one is what they call compositional features. That's basically who is in the team; looking at things like the personality of team membership, looking at diversity within the team. The second one is structural features, which is how the team is linked, who talks to who, how they communicate, how they set up, how the leadership relates, whether it's one leader of the team or multiple leaders of the team.

The third one is mediating mechanisms. This is anything other than those first two things that affect the performance of the team. Things like leadership characteristics would be considered as a mediating mechanism or the use of particular technology might be considered a mediating mechanism. The final one is context, which is things outside the team that affect the way the team itself behaves and performs.

To give us some idea of why they picked those four things, it's worth pointing out a very common framework that gets used in team or leadership research. It's something that actually that this review ends up being quite critical of. They call it the IPO framework or input-process-outputs. That's the idea that you have your input—what goes into the team—your process—what the team does—and then your outputs—what the team produces. You can measure each of those things often by surveys. That's where you get really easy answers, so you input what you put into the team.

It'd be lovely to know that putting a diverse team is better than putting a homogeneous team. It'd be lovely to know that extroverts in a team are better than introverts. The process is about how you structure and train the team. It would be lovely to know that if you send people on a team-building exercise, they will be a better performing team. The outputs are straightforward measurements of what you get from those inputs and processes.

One of the lessons throughout this paper is that most research into teams follows this framework. As a result, the research is hugely contradictory about any question you could possibly ask about a team where almost any question you ask, there's a dozen papers that give different answers. I might just throw in a quote that comes near the end of the paper.

This is a preview before we get into it. It says, "Despite all the progress in recent years, we believe that workgroup research is poised to enter a new era. Widespread adoption of the IPO framework, the ease of survey data collection, and the scholars’ desire to conduct field investigations and employ sophisticated statistical techniques have combined to yield a prototypical research design where members' reports of team properties are associated with some index of their effectiveness. Rich observational studies are few and far between, field experiments are uncommon, and action research is all but absent in the team's literature."

David: That doesn't sound too inconsistent with our manifesto paper from episode 20, Drew.

Drew: I think you could just swap out teams, put in safety, and it's exactly it. People want simple models. They want to collect data using surveys and they want to put their sophistication into the statistical analysis of the surveys rather than into rigorous data collection. But I don't want to be pessimistic. We pick this paper because we think it actually has some useful things to say and talk about. I just wanted to give you that warning upfront that the useful and interesting things aren't going to be simple recipes.

They're not going to be simple, clear statements about this work for teams, this doesn't work for teams. But the picture of the overall research gives a lot of ideas of different things we can think about teams, different ways we can think about improving teams, different ways we can think about improving team research. We'll dive into each one in turn, starting with structural, which is mostly about how the team communicates. If we think of the team as a network, how does that network work?

David: Yeah, Drew. I think that's the way you described it and we'll start with the structural characteristics in a minute. At least when I read it, there were lots of things that I found to be quite practical and quite intriguing takeaways. I would just echo your comments there in almost all of the areas that we're going to talk about.

There was generally a discussion where something like this impacts this except in this case, or this impacts this to some degree but then there was another piece of research that said that it didn't. We'll point out a few of those areas just so that we can make it clear whether there's alignment among the research in a particular area or whether there’s not.

When we come into the structural design of teams, I thought of these as a little bit like the design of the teams, like how you set up a team, when we start thinking about the management of a team. One study found that really high-performance managerial practices—having good knowledge management systems so people can find the information that they need, good models of decentralized decision making so people are able to make decisions they needed to and good systems of work, good clear processes for the team to work through—were positively associated with effectiveness.

I think also in that, when we think about managerial practices, there was another interesting study, albeit from a sports team that found if a key player of a team was missing, that the other members of the team actually reduced their interactions and the remaining members didn't actually experiment with new approaches. When the key member of the team was gone, they almost lost their confidence to play their natural game.

Drew: I love the idea of using a hockey team, both as a team to research, but also as a metaphor for any other team. The beautiful thing about hockey or soccer, particularly at sports, is that you can clearly see all of the interactions between the team members because they're very obvious, particularly when the teams are passing the ball from one person to another. That's a really obvious interaction.

In the study of the hockey teams, they could see that there were a couple of key players, the playmakers, and everything that the team does revolve around interactions with these people. You've got a really highly skilled center, so you tend to pass the ball to them over and else run around and then they feed the ball out to the appropriate place. They found that if you remove that highly-skilled center, that doesn't change the strategy of the team.

They still keep trying the same strategy of your past to the center, run around, and then fade out again, except the strategy is no longer working very well and there are fewer passes going on because you lack that key player who was making that particular team strategy work. I think that's a great metaphor.

If you've got one person within your team who is handling the interactions, creating the interactions, making sure the team works and that person disappears, it's not just that the team can keep on doing what they were doing before. But teams will tend to just keep trying to keep the same strategy.

David: Yes. We'll try to hone in on these key takeaways because we are talking about a lot of different aspects of team performance. In those early structural ones, setting teams up for success with clear work processes, clear information, clear decision-making, knowing who the team players are, and making sure they’re present is going to improve your team effectiveness.

And then, there was a lot of discussion in the research about task, scope, and complexity. In this area, there was lots of conflicting research. Does a simple task make a team more effective or does a more complex task make it more effective? Does a team operate differently when it's a simple task or when it's a complex task? I must admit, I got a little bit confused in this research about task complexity. It seems that if you are researching teams, you really need to think about how complex the task is because it's going to change the way the team functions. Is that the way of thinking about it?

Drew: Yeah, I think the underlying question of most of this type of research is how do you set up the team leadership. They're talking about things, like in a surgical team, is it better to have the surgeon in a command and control mode? Or is it better to have joint leaders, delegated authority, or lots of the team members working autonomously? Where the task complexity comes in is that the answers to those questions change depending on what the task is.

With some tasks, it's really important to have a single leader. For other tasks, the team is much more effective with joint authority or very delegated authority across the team. But the research is conflicting because there's no simple answer like as the task gets more complicated, then you should move towards more delegated authority. It's like if the task is already very complicated, then a small change in complexity means you should move this way. If the task is already very simple, then it means the opposite. The clear takeaway is task complexity matters. Very unclear takeaway, what you do with that information?

David: I think at the end of the day, for a team, the task is the task. Let's use the research to understand how to set the team up and the individuals within that team up around the task that they're going to complete.

Drew: One of the things that is very useful out of this structured little research is it points out a lot of the features of a team that can be manipulated because researchers love to tinker with teams in order to research. You can look at all of the different variables that are worth studying and therefore also, presumably, are worth trying to manage in real-world situations.

There's a bit of a list they give. You can manipulate how the resources are parceled out at your time, money, access to other people. You can manipulate the workflows, you can manipulate how goals and rewards are assigned, whether they are set for individuals and then rewarded for individuals, or set for the team and rewarded for the team. Then, you could look at how things like team trust or teamwork processes influence those structural features.

David: We're going to talk later about trust and teamwork processes, team cognition and cohesion, and those types of things. I think the interesting part here structurally is we'll talk about how those old matters for team effectiveness later. But you can actually mediate for those things through these formal structural aspects like resources, workflows, goals, rewards. I think the teams, the right amount of resources are very clear workflow processes, clear goals and rewards, then you can still get an effective team even when you may not have the levels of trust, cognition, and cohesion that might otherwise make your team effective. It's good to know that you can actually look at a team and know based on how the team is performing, as we talk about later, whether structural features or some of the more internal features of the team are going to be better to improve performance.

Drew: Thanks, David. I think that's a really useful way of interpreting the research. The final one is actually the reverse side of that. A topic that comes up a lot in the research is virtual teams. Who would have guessed that teams meeting over Zoom was going to be topical, relevant, and a hot button topic? There, the research says the converse of what you're just saying, which is that where you don't have a lot of those normal structural features, then you can compensate for that with things like trust. If teams have a reduced ability to communicate because they're virtual, then things like trust, cohesion, and common understanding are going to be very important in the effectiveness of the team.

David: I think there'll be a lot of people around the world who are finding out just how much trust there is in their team at the moment by how well they're working together virtually.

Drew: I think you measure the trust in the team by how much your boss insists that the video camera is on to check that you're actually sitting at your desk rather than trying to juggle 15 other things while you handle the team meeting.

David: Yeah. Including homeschooling. There were structural features and let's move on to the compositional features because this aspect of the paper was based on more than 150 studies. This is about looking who is in the team and what are the characteristics of those individuals within the team. As you mentioned earlier, things like how diverse are the team and a little bit about how they work together. There are also some really interesting structural features that we'll talk about but these are tied into the compositional area for some reason. We'll talk about fault lines a little bit later, and I found this really interesting to think about teams within teams.

Let's start with average member attributes. This is where we start looking at a team and going, on average, what does this team look like and how does it perform? Drew, surprisingly or unsurprisingly, teams that have a higher average level of expertise and a higher average level of mental or cognitive ability tend to perform better.

Drew: Yeah, I'd love to know who thought that that one was worth researching. Maybe the underlying theory was that teams destroy competence or something. If you're looking at the average composition of your team and having better people in the team results in a better team on average. I think the more interesting one—and this one did surprise me just how consistent this does come out—is that you can actually predict team performance based on average personality traits in the team.

Listeners may know that I'm fairly skeptical of research that makes use of personality traits to predict other things. But there's a fairly consistent finding that the average conscientiousness, agreeableness, and extroversion, which are three factors in the commonly accepted Big Five model of personality, predicts team performance. I remember that these are average so we're not saying you should aim for this in every team member. The way it goes is that, where on average, those three factors are high, you get a lot more cooperative, helping, and engagement behaviors within the team which in turn leads to better productivity and effectiveness on tasks.

David: Yeah. This seems a little bit obvious but I was also a little bit surprised at how consistently it was found in the research. When we think it's finally worth restating it for people as it's a takeaway right now. Conscientiousness, how well each member of the team on average is focused on the task and paying attention to the needs of the team. Agreeableness, how able is the chain to align, agree, and move forward with the task performing. Then, extroversion, which, in a team setting, the ability to make your contribution, lean in, and speak up. Those are three interesting traits if you're selecting team members. But like you said, it's not the traits himself, it's the behaviors that tend to be displayed by people with those traits that matter.

Drew: Yes. In fact, there is some research that speaks directly to this which says that directly measuring emotional intelligence, which I think can fairly be summarized as the way in which those personality traits manifest as interpersonal behaviors is a better predictor than personality. Rather than giving someone a straight personality test and putting in all of your extroverts, giving out all of your introverts, probably what you really want to do is to test based as much on behaviors rather than psychometrics as possible on who is displaying that cooperation, leading in, willing to seek agreement-type behaviors in the team.

David: Yeah. The next compositional fixture was diversity. There's a lot of general information about diversity and team effectiveness. I think in the general press, which generally says that diverse teams are better-performing teams, I think that would be a general conclusion of the popular press. But here, in this analysis, they split diversity into three different types, which they call surface diversity, deep diversity, and functional diversity. Let's go through each of these and then talk about what it means for team effectiveness. Drew, do you want to start with surface diversity?

Drew: Sure. Surface diversity is the idea of deliberately having people in your team with different ethnic backgrounds, different genders, or different ages. You readily measurable demographic attributes. The hypothesis is that having a stronger mix leads to a higher-performing team. With the caveat that we're talking here specifically about all else being equal, how well do they work as a team? That's not actually a clear finding on research at all. If anything, the research tends to point the other way. With lots of contradictory evidence, some of it suggests that the more diversity there is, the harder it is for the team to work together and harms performance.

David: Yeah, Drew. It wasn't mentioned in this paper, but I had heard of finding that diverse teams are less effective with simple and knowing tasks. Surface diversity only becomes potentially more useful in teams facing novel law or complex challenges and problems.

Drew: Yeah. The way they do this research, I think there's a lot of potential benefits of diversity that the research won't capture. I think it will tend to score a team that converges on a particular solution to a problem and gets that solution implemented quite highly, even though that solution might only work for a small percentage of the population and they've totally missed out on important ideas.

This isn't like a statement against diversity, but it is a clear finding not to be simplistic about how you think of diversity, and to recognize that deliberate attempts to introduce diversity can come at a cost of team efficiency and performance. There's a couple of studies and the research is fairly sparse, so don't take this too heavily. Try to explain why that's the case.

They suggest the leaders’ attitudes may be one of those things that even if you've got diversity, having a leader who sees that as diversity, who tends to categorize people based on external attributes, may increase the harmful effect of diversity. Whereas having a leader who just takes a different mix of people for granted and treats them all just as the same group of people may mean that diversity leads to a stronger, more highly-performing team. That's surface diversity. David, do you want to say anything more about that?

David: No. I'm happy to move on from that. I think there's a lot to say about that but you've described it well and it's such an important and topical issue for the business world to increase diversity. I think it's absolutely necessary for the more effective functioning of organizations. It’s just we need to think about diversity beyond just surface diversity and the two other ways that this paper talks about are (like I've said) deep level and functional level diversity.

The deep level diversity is looking in now at the individual psychological characteristics regardless of their surface, gender, or ethnicity. These characteristics are things like personality, values, and attitudes. There's a reinforcement here for what you said earlier, that a lot of the deep level diversity factors seem to produce mixed results, but emotional intelligence consistently seems to be related to improving team performance. If mixing up your surface level diversity keeps your people with a higher level of emotional intelligence, then that's going to be a good thing. But emotional intelligence seems to be consistently applied to improve team performance.

Drew: Yeah, we might touch on this more in the takeaways. But my interpretation of this research, to the extent that there is a consistent message, is that when we talk about trying to improve diversity, there are two separate things we want to improve. There's the diversity itself and then there's the diversity, competence—the ability of people to work with other people of diverse backgrounds, diverse thinking. You need both. If you just have the diversity without the diversity competence, it reduces performance. If you have both, it seems to be a stronger performance.

David: The third is functional diversity. I think it resonated most with me. This is having people with different backgrounds and people from different functional areas. What that might mean is putting a team together with someone from human resources, someone from finance, and someone from the safety department. I suppose the research suggests that functional diversity, as well as individual educational diversity, have positive relationships with team performance.

When they talk about functional diversity, they separate this out into the person's dominant function but then they also look at this aspect of the number of their diverse functional experience. Their dominant function might be there in the engineering function or in the safety function. But there are also some studies that looked at their career functional experience.

There's a saying that I use sometimes in business that two people can have 20 years of experience. One can actually have 20 years of diverse experience and one can have 1 year of experience repeated 20 times. In this research, it suggests that functional diversity in terms of the number of different, diverse, functional experiences a person had in their career and bringing those people together is better for team performance.

Drew: Yeah. There aren't a lot of papers surveyed in this part of the study, but at least they seem to be fairly consistently pointing to the same thing, which is that people who themselves have had diverse backgrounds tend to be able to work and play better with others in a team. 

David: Diversity matters for team performance which struck three levels: surface, deep, and functional. The obvious one that we worked towards (surface diversity) seems to have the least consistent impact on team effectiveness. We'd probably be concluding here and encouraging people to think broader about diversity, particularly in terms of team makeup and just understand this deep level diversity around emotional intelligence and understands people's functional diversity throughout their career and how they might bring that into the team.

Drew: The final compositional factor they look at is the idea of fault lines. This seems to be an increasing trend in team research to move beyond what's the average properties of the team or what's the distribution of the team diversity to look at is there some other internal structure. Can we split the team in half and see where one-half of the team is more similar and the other half of the team is not? The real-world application of that is the idea of having a clique in the group. There's one group of people who are very similar, who talk to each other and ignore the other members of the team. The general idea here is this idea that having cliques or faultlines is going to have a negative effect on performance outcomes.

David: I think generally we would assume, as the researchers for a long time have assumed, that these fault lines are bad for teams. Any divergence or separation between the members of a team isn't great. They found that this can be moderated by friendship and it can also be enhanced by animosity. If there are fault lines in a team, if there are subgroups, if there is some reason that half the team is at one location and half the team is at another location, then if there's actually conflict among the groups across that fault line, then it will absolutely worsen performance where there's some friendship and social cohesion across the group, then it will moderate the negative effects of those fault lines within the group.

This research is really unreliable, even though there's something like 25 of the 150 papers in this part of the review had mentioned fault lines. There's actually no real agreement on how to identify and how to measure what fault lines actually are. Therefore, this research is fairly sporadic. But I think the practical take-out for me (at least) would be that if there is some separation within a group or a team, just try to understand what the impact of that might be on performance. If there is an impact on performance, then when we talk a bit later about things like social cohesion, there are ways of mediating that impact.

Drew: Yeah. I was honestly unclear how to interpret this research, mainly because the way they do it is very statistical. It's giving surveys to the members of the team and trying to identify the fault lines based on the survey responses and the statistical analysis, rather than actually talking to the members of the team about who they liked and didn't like. Yeah, it's an interesting topic. Perhaps the statistics don't enlighten nearly as much as trying to understand what's actually going on within the team.

David: Drew, once we go past structural and compositional, we've talked about the structural aspects of a team in terms of how they're designed and how they're managed. Then, we talked about the compositional factors about the people who make up the teams and the characteristics of those people.

Now, there's a list of mediating mechanisms. These are the things maybe in between the structural and the compositional factors that play a role in the link between those structural, compositional factors, and outcomes of the team and many of these are emergent. You put the structural and compositional factors together and certain characteristics of a team emerge. They're really interesting to study because they're the things that then create that link between the structural factor, the mediating mechanism, and then the performance.

Drew: The way I like to think of these is they're the type of things that we might want or not want in the team. They give us some suggestions of the type of things that we might use, either as formal or informal measurements of whether a team is working well or not, or whether we have the team that we want. David, I thought it might be worthwhile, just like really quickly going through and just giving these as a list of emergent states to get our listeners to think about whether they are things that they strive for in teams and how important they are.

David: Okay. Do you want me to rip through them, Drew?

Drew: Yeah. You want to read through each one and explain what it is.

David: Oh, wow. Explain them as we go.

Drew: Well, you just have one sentence here. What is team cohesion?

David: Let's start. There are team processes. These are the actual processes that the team is following to accomplish their work. There's information sharing which is quite obvious, which is how the team shares information among members. There's team cohesion, which is how aligned is the team around the task and how well did I get on. There's the inter-team trust which is pretty obvious, how well do members trust each other. There's team potency, which is what's the belief of the team that they can be successful in the task they've been given.

There's conflict, which is, as you might expect, there's conflict on over tasks, conflict over relationships or relationship conflict, and conflict over the work process. There's team empowerment, which is, again, obviously how empowered the team members and a team collectively feel to do what it needs to do to get its task done. There's shared leadership, which is to what extent that leadership roles and behaviors within the team are shared between members. And then, psychological safety, which is quite relevant and topical now within research and industry, which is really about how safe people feel what they want to say.

Drew: What I found really interesting was that pretty much that whole list, except for conflict, there's weak evidence that those things are good for teams. The trust, cohesion, team empowerment, shared leadership, psychological safety are good for teams, which is surprising, given that you'd think that those things would all be unambiguously good. The fact that it's not yet unambiguous rah-rah in favor of each of these things, I think is either a limit of how well we understand teams or a limit of how well we research teams.

David: Yeah. Teams are (by nature) very, very complex. As soon as you put a group of people together, it becomes complex really, really, really quickly. There's so much context, situational dependencies, and interactions that occur within a team. I get that it's hard to understand, hard to research, and I suppose hard to research consistently if cognitive psychology had a replication crisis in the recent decade or so about trying to even just retest individuals and get consistent results. Trying to retest teams and get consistent results is those results are even harder to replicate.

Drew: There are some studies in there that hint at some really exciting things that are worth trying. For example, a couple of interesting studies about shared leadership that say that there's a positive relationship between shared leadership and performance. That includes things like rotating the leadership role or deliberately splitting different leadership responsibilities and having different members in the team have different leadership functions.

There are multiple studies that have given different benefits of applying those shared leadership strategies. I found myself reflecting on teams that I've been part of myself. Very often, this isn't something that happens deliberately but where it does happen organically can be (at least in my experience) very helpful in getting the team stuff done just because very seldom the official leader has a desire, an interest, or capability to do everything that you'd expect of a leader.

David: Yeah. I think I also found team empowerment interesting because team empowerment seems to be quite closely related to the leadership of a team. They called it an exemplary, longitudinal experimental study. What this study did was manipulate the leadership style between a directive style of leadership and an empowering style of leadership.

What they found was that teams with a directive leader performed better initially, which means that if a leader just stands up and gives instructions, the team will get to work and do stuff faster. Whereas the empowering leader of teams, these teams with an empowering leader ended up with a higher overall performance, which means they got off to a slow start while they were trying to format may be how they were going to sort out their shared leadership, and how they were going to align around that task in the process. But once they did do that, the teams with an empowering leader that let the members don't get on and do their work ended up being high-performing.

Drew: I'd be interested in your own experiences with this one, David. It certainly resonates with me that I always have this tension between, I know I could do the job better myself if I just did it or if I told someone exactly what I expected of them. But the next time the capability hasn't increased at all and I'd have to do that same thing again. In the longer term, often means making a short-term sacrifice in letting people do things their own way, do them not necessarily in the way I would have done them. Over time, that gives us far more capacity as a team to do things.

David: Yeah, Drew. I think that I find myself agreeing with that observation. I think that's why we say different leadership styles and different types of operations. In a construction and contracting environment where teams generally have a short lifespan and they get mixed up and churned up a lot, then we see a lot of directive leadership styles to just get the job done. When we look at more stable and longer-term environments and I think, in here, it was really clear in the research that inter-team trust grows over time. It's something that is very hard to start off with a level of trust. It's something that you just have to give a team time to form it and the same with things like cohesion as well.

In the long run, I think I would say that a less directive style of leadership is going to build a stronger team. I think the research supports that. But what I just shared earlier is you may have to be willing to sacrifice the initial performance of the team to achieve that in the long run.

Drew: I think this is really interesting when it comes to safety and thinking about things like organizational climate that often our ability as teams is strongly shaped by what the organization allows. If the organization isn't going to allow a team to stay together, to be measured on their long-term performance, then the leader, for the good of themselves and the team, is going to be required to take a strategy which is optimal in the short-term, but not good for either the team, the individuals or the organization in the longer term.

David: I think where we say this, where an organization might have a sudden serious incident, a sudden rise in the number of lower severity incidents, and all of a sudden the organization feels there's a crisis and then leaders need to step in. We see that sometimes having a short-term impact, that we rarely see that stepping in as having any sustainable improvement in safety in the organization.

Drew: Moving on, we mentioned originally four categories of things and we've covered three of those categories. We've covered the composition of the team, the structure of the team, the mediating factors. The fourth one was factors outside of the team. I've noticed that neither David nor I have taken any notes at all in this section because there was nothing interesting. A little bit of leadership, a little bit of organizational climate, very similar to a lot of the safety science literature in the same space.

David: I think I just got it right, no really interesting findings here. There were a couple of pages in the study about these contextual features of teams. If you've got a team within an organization, what's the impact of the broader organizational climate on that team? There were some studies that said if there's a general climate in the organization around innovation, then the team will display more innovative characteristics and things like that. But it wasn't actually very many studies.

Of those studies, there was nothing really compelling there in relation to the senior leadership of the organization or organizational climate. That's not to say it doesn't matter and doesn't have an impact. It might be more to say that when people research teams, they research what's going on within the team and there hasn't. They may not have been great researchers when they looked at the team situated within the broader organization. I don't think it will be that hard to study to look at different teams operating within the same organization, trying to understand how the contextual features, and then how the compositional and structural features of a team worked within the same contextual climate.

Drew: Yeah, I think that's the holy grail of this research, is to have a study that's robust enough that you know what is the organization and what is the team. The survey concludes with a bit of a talk about intervention research. I found this eerily familiar when looking at teams versus safety science in the research that actually involved trying to tinker with the variables and improve the performance of the teams with very, very sparse and with very little interesting to say.

There are some of the things that you'd expect, things like team planning and having team charters, those team preparation tools seem to have some positive benefits. Team building has some effectiveness so long as the team is together long enough. Team training seems to be effective, although there's nothing that tells you what type of team training is better than any other team training.

David: I think you're linking some of those preparation tools back to—you mentioned team planning and team charters—some of those emergent properties and things we didn't talk about. Task cohesion has some relationship between team performance. I can see that if a team plans its work, it's going to have alignment around the task it needs to accomplish, which is linked to performance.

Also, what I talked about was team potency, whether the team believes that it can actually do the task. I think in team charter and things like that can actually create that belief. Whether or not these things themselves, processes and artifacts are the things that matter or what they then lead to as a result of going through the processes of doing those things.

Drew: Yeah. Those findings are pretty much the same as the state-of-the-art in team research 10 years ago. We knew those things then and we haven't learned a lot about them since. What I found interesting is some of the most recent studies are starting to put new methods for studying teams that I think are going to give us some pretty exciting results within the next few years.

Those are things like using big data to study team communication, specifically looking at how the team works as a network based on who talks to whom, how they communicate, who sends emails, not just looking at the raw fact of the communication but also studying things like the content of the communication, how ideas, common language, common goals, common preoccupations spread throughout the team. That's particularly interesting, given that we're about to have a massive archive of team meeting Zoom calls where it's every interaction the team has been explicitly recorded, either the content or the fact of it.

David: Yeah. I am aware of a study that was done a few years ago that just took the organization's email records and did a big network diagram or nodal analysis that basically every email—who is the sender and who is the receiver—and are really trying to look at who the key players were, if you like, in the team, who were the people that people went to most often, obviously to communicate with. Obviously, there are some limitations like that if you don't think about phone calls and other types of interaction, but for that organization, it was really telling about where the bottlenecks were in the organization and who were the people who were dealing with a lot of traffic, if you like, held a lot of knowledge potentially about the organization.

Drew: I think that sort of thing is interesting and promising. The other thing that's interesting is the use of wearable sensors and using personal electronic devices to study how people move and interact in teams. This is something that people have wanted to do for a long time. I remember studies coming out of MIT, Media Lab, 20 or 30 years ago but we've actually got the capability now not to need very specialized technology to do it, use off-the-shelf stuff.

David: Yeah. Necessity is the mother of innovation. I was actually talking to some research and professional colleagues in the US on the weekend. They're now wondering how they can go out, observe teams, and collect data with social distancing. We're brainstorming around how to do that.

What we ended up designing and agreeing with them is in each member of the crew—these are field crews that work with teams of about four or five—they're all going to have GoPros on their hardhats. It's like when an observer, when a researcher is out in the field, they're standing there and getting one perspective. But now they can have this movie of all of the individual perspectives about who's talking to who and what they're talking about with sound and audio, what they're looking at, what they can see, how they move around the worksite, and then expectedly get a whole lot more richer data than putting a field researcher in the field.

Drew: Yes. Even if not for the richness of the data, something they said in this study, which really resonated with me is they said, "The reason why we don't have lots more of these observational studies is not that people don't understand that to see how a team works, you have to watch how the team works. It's just that putting an observer to watch all of those daily interactions is a really, really hard and expensive way to do research." The use of technology can both replace that need. But also, as you say, David, massively even get a better view than having one researcher. You can have the point of view of lots of people. 

David: Drew, shall we move on to practical takeaways now?

Drew: Absolutely. You've got a list of 10 very practical takeaways. I'm going to sneak in first with my one, which is about the research side of things. If we're going to research teams, we need to study team processes over time, not just isolated states of teams. I think that applies to organizations managing teams as well. We're not interested in the snapshot. We're interested in how things operate over time. It's those processes rather than the personalities or the metrics that we have the biggest control over. I think that's a great thing to be thinking about right now.

That's one of the genuine opportunities of something like the COVID-19 isolation is where having to be much more deliberate at the moment about who we meet, when we meet with them, how we communicate within our teams. As a result, lots of organizations are actively improving the extent to which their teens identify each other and deliberately make opportunities to get together and communicate. I'd love to see that carry over to have organizations just focused on. Not having meetings for the sake of meetings or having meetings just for tasks, but working out where we need a team to function well and what processes do we need to have in place to help that team function?

David: Yeah. I think that's really important, Drew. What I did try to do, because this was a meta-analysis in the organizational study space, it's outside the safety space and it's a review of more than 600 papers over the last decade. I wanted to come up with a list of 10 practical takeaways, happen to be a list of 10, or might have massaged into a list of 10, but there's 10 anyway. I might run through each of these. 

Number one, the task is the task. Don't worry too much about the scope and complexity of what task giving people. A task is a task.

Number two is to pay attention to structural factors. There are things you can do in terms of providing the resources, the workflows and processes, the goals and the rewards, and that these will influence the emergent characteristics of the team. A team that doesn't know how it needs to do its work, what resources are available, or what the end goal is isn't going to be as effective.

Number three, team potency matters. Does the team believe that they can accomplish the task? Belief, I suppose, precedes success. That's really important for team effectiveness.

Number four, obviously, teams with a higher average mental ability, experience, conscientiousness, agreeableness, and extroversion will perform better. But remember, it's the behaviors that you want to see as opposed to just the traits in the individuals.

Number five, emotional intelligence and functional diversity are important. The more emotional intelligence you can cram into a team and the more experience in terms of career and functional experience that you can cram into your team, the better.

Number six, interpersonal processes and capabilities like conflict management and motivation are important for team performance. Drew, we didn't talk about this in detail, but you're going to have conflict, you're going to have issues, you're going to have challenges that the teams have to face. The individual abilities of the team members to manage those conflicts and motivate themselves and others are really, really important. This is another call that we've made in (I think) multiple podcasts for non-technical skills training.

Number seven, teams that share information freely perform better.

Number eight, team cohesion about the task is more important than social cohesion. Another thing that we didn't quite talk about and the main thing is that you can have people that don't necessarily like each other. But as long as they agree on what the task is they're doing, they'll be more effective. Where you've got people that like each other but don't necessarily agree on what the task is, they'll be less effective. Regardless of how you're supposed to take it out there, regardless of whether people in a team like each other or not, make sure they're aligned on what the task is.

Number nine, team trust and psychological safety increase cohesion and performance. This trust is really important to resolve the social and oppressive process conflicts which come up in any team. I think one of the takes here is these things take time. Don't expect to put your team together on day one and have automatically high levels of trust and high levels of psychological safety. I suspect true unless those aspects were already contextual factors within the organization. But even then, that may not be consistent among team members of a new team.

Number 10, shared leadership and team empowerment are both related and both contribute to improving team performance. Finding ways to share the leadership among the team and having the leadership empower the members of the team are going to give you a more effective team. That's it, Drew. There's 10. What do you think of them?

Drew: I think that's a pretty great list of 10, David. The only one I'm going to modify is that one about team cohesion being more important than social cohesion because this one surprised me. This isn't from this paper. This is something I went to look at a couple of years ago, trying to find evidence to disprove the idea of silly team-building exercises. What surprised me was that the evidence shows pretty strongly the team cohesion and social cohesion are correlated.

Two people who talk about football will also talk more about the job and to people who talk more about the job will also talk more about football. Sending a team out on some nonsense exercise out in the wilderness where they're getting no benefit related to the task at all, can actually be good for them just socially relating and talking to each other more. It doesn't increase team efficiency because people waste a lot of time talking about things that aren’t relevant, but it does improve team effectiveness.

David: Very good. There you go. Social cohesion and task cohesion are both important for team performance. Drew, usually we end with invitations to the listeners of things we'd like to know. As I said, this is a really big meta-analysis but doesn't mean it contains everything that's important for team effectiveness. I thought I'd ask our listeners, in their own experience, which things stand out for them as being really important to team effectiveness and why? There may be things that we haven't mentioned today and there are likely things we haven't mentioned and things that aren't into the literature, but if listeners want to share their experience with effective teams and what they feel made it effective, I'd love to hear them.

Drew: That's just about it for this week. Our goal in this episode was to try to do something that we have said before we think it's a good thing, which is stepping outside the safety science research to understand what's going on in adjacent fields. The people who did this study did their homework, so we didn't have to and we've done our homework so you don't have to.

Having a look at what's currently going on in the team research space, is there anything big that you're missing out on? I think the answer to that question is mostly no. A few useful takeaways but don't be afraid that there are some magical team results that you should be applying that you haven't heard of because it's been kept secret, buried within the research literature.

We hope you found this episode thought-provoking and hopefully a bit useful in shaping the safety work in your own organization. Send any comments, questions, or ideas for future episodes to