March 2005
Vol. 2, No. 9
This newsletter is a basic HTML file, giving you control over the appearance through your Web browser's options and window controls.
From the Editor: Using sports teams as a model for business teams is overdone, but several factors came together this issue to make me want to mention the connection between the two:
This year's players are mostly the same set the team had last year, when the team's performance was disappointing. The difference, the sports analysts say, is that those players have "bought into the system" of their second-year coach, Roy Williams, a former UNC assistant who later built a great program at the Univ. of Kansas. This year the players pass more, help each other on defense, and excel in their roles, but also know when to take the lead if others are struggling. In short, they are "playing like a team." This is what comes when people dedicate themselves to each other within a proven teamwork system.
What also makes UNC stand out for me is not just an amazing string of statistics (making the national tournament for most of 30 years, three national championships, etc.) but the way they achieve high performance. Through four coaches since 1960, there has never been even a hint of intentional rule-breaking, and the players have graduated at higher rates than non-athletes, which is not the case at many colleges. Players and coaches maintain close relations decades after graduation. With one notable exception who did not last, the coaches have generally handled themselves with a decorum few less-successful coaches match. In short, Carolina doesn't just win. It wins with, or through, discipline and character. Would that all coaches—and politicians and CEOs—followed that path.
Are you in a professional, business, civic or volunteer group that invites in outside speakers for its meetings? I am available at no charge to provide passionate, fun, and practical talks on a range of topics related to people working in groups. Check out the "Speeches" page of the TeamTrainers site and forward the link to your group's meeting coordinator today.
Study: Three Canadian business professors have taken the first look at the connection between workplace safety and "high-performance work systems." These systems, also called "employee-centered management," are a set of management techniques that the professors say value employees as "a primary source of competitive advantage that is difficult for others to imitate…" It includes open-book management, employee participation in company decisions, rigorous hiring practices, high investment in training, pay for performance, and yes, independent empowered teams. In contrast, the authors write, most safety programs focus instead on "the use of rules to enforce behaviors and the use of punishment to increase rule compliance…"
Taken as a set, the employee-centered approach to management has been shown to cut costs and improve performance in many areas, and this study of 138 manufacturing firms found the same to be true for safety. The researchers surveyed the human resources directors at these firms to gauge their use of ten employee-centered techniques and compared those to the numbers of injuries serious enough to cause missed work time. After eliminating the effects of company size, sector (private vs. public), levels of unionization and other factors, the strength of high-performance systems explained eight percent of the differences in injury levels.
A second study looked into how this happened. The professors sent a survey to "196 employees at two Canadian organizations from the petroleum and telecommunications industries" in plant or field operations. It asked about management techniques and safety practices from the worker's standpoint. High-performance work systems turned out to be strongly linked to the workers' trust in management and sense of a climate of safety at the company. These, in turn, affected the "personal safety orientation" of the workers.
Application: If you're looking for more data to convince your managers it's time to shift their thinking about workers, here it is. A potential eight-percent reduction may not seem huge, but anyone in HR can tell you injuries that serious have cost impacts well beyond the lost productivity. Higher Workman's Compensation and insurance premiums are only the most obvious.
If managers don't feel trusted, the study's 77% connection between use of employee-centered management techniques and trust levels should say something as well. The ten elements are too much to go into in detail in a newsletter, so here's the bottom line for those at the top who want to reduce injuries: stop treating your employees as children who need rules and punishment to protect them from danger, and start treating them like adults who have good reasons for doing dangerous things—reasons your management priorities probably created. Now that's dangerous thinking, eh?
Source: Zacharatos, A., J. Barling, and R. Iverson (05), "High-Performance Work Systems and Occupational Safety," Journal of Applied Psychology 90(1):77.
Study: Guilt may be the secret ingredient in why groups are more likely to compete with each other than are individuals, according to a pair of studies from the University of North Carolina. The studies used variations of the "prisoner's dilemma game" to determine this and some other surprising results that show the danger of thinking of a group as a single person. (If you know or don't care about how the prisoner's dilemma works, you can skip the next paragraph.)
Used for decades to test human relationships, the prisoner's dilemma is based on two criminal suspects being questioned in separate rooms. If they both say nothing to the police—that is, they cooperate with each other—they are better off than if they compete to get better treatment in court by snitching on each other. But if one cooperates (with his partner) and the other competes, the snitch has the best possible result. After all, if both confess, they both get the worst sentences; if both cooperate, they're still in trouble, but not as much; and if one testifies against the other, he usually gets a light sentence while his partner gets the worst possible. But what happens if the pair are career criminals who face the scenario over and over? By attaching numerical values to the possible outcomes, in the original prisoner's dilemma research, scientists learned that the highest possible value over all repetitions comes to the the prisoner who cooperates with his also-cooperative partner every time except one: the last time they commit a crime. If he then testifies against his partner while the partner again keeps his mouth shut, the betrayer will have gotten moderate results every previous time, plus the maximum value that last time.
In most human interactions in the real world you can compete, cooperate, or just withdraw, neither competing nor cooperating. The basic prisoner's dilemma study design is that two people or groups are separated, made to believe something about the other side, then put in a situation in which they have to choose between those three options. In the UNC studies, the subjects were tested either one-on-one or in groups of three-on-three. (Unlike the classic design, they were told there would only be one interaction, so they didn't have to worry about the future.) In the first study each side was tricked into thinking the other was either very trustworthy or not, and that the other side was given no information about their own trustworthiness. In other words, the other side was supposedly at a disadvantage due to having less information. The authors report, "Consistent with repeated past findings, groups competed more, withdrew more, and cooperated less than individuals." If one side thought the other side was trustworthy, both individuals and groups were more likely to cooperate. This "may be due to the questionable fairness of competing with a trustworthy, but vulnerable, other," the authors write.
In the second study, the sides were tricked into thinking they preferred the art of the same or a different painter as did their opponents. Personality tests in combination with the dilemma choices revealed a big difference between groups and individuals. People who easily felt guilty were less likely to compete with their partner than those who didn't. But teams with guilt-prone members were more likely to compete with other teams. One-on-one the guilt-avoidance was directed at the other individual. In teams, it was directed at fellow team members. To avoid feeling guilty toward them, guilt-ridden people apparently took a chance on the biggest possible payoff (which required competing with the other team).
Application: To look for applications on the job, let's tackle what the researchers considered to be the main findings:
Source: Insko, et al. (05), "Interindividual-Intergroup Discontinuity as a Function of Trust and Categorization: The Paradox of Expected Cooperation," Journal of Personality and Social Psychology 88(2):365.
Study: Another element in getting team members to help each other is accounting for their skill differences (by type, not level), according to a study out of the Netherlands. Two management professors at the University of Groningen followed the progress of 120 students in a grueling management game. Regular classes stopped for three weeks and student teams played the role of corporate top management teams in various industries. In each of four decision periods, "the teams had to make decisions on a total of about 30 strategy issues that covered all important business areas such as investment in production capacity and quality, marketing," and personnel issues, the study says. To determine performance, their decisions were plugged into numerical models from the real world.
After both the first and second weeks, the researchers also tested the students' beliefs about the following factors (shown with some of the questions asked):
Note that, as I mentioned earlier, the skill measures were about different kinds of skills, not different levels of skills. Also, the researchers had team members rate each other on helping behavior and compared those results to the "self-ratings."
As found in previous studies, teams whose members actually helped each other more performed better than less helpful teams. Equally typical was the finding that higher reliance on others to finish one's tasks at the end of Week 1 was linked to higher helping (both self- and other-reported) during the following week. The perception of different skills did not have an effect overall, but it did on some teams. People who didn't feel they needed others to get their work done and thought team members had different skills were less likely to help each other.
Application: If you have a cross-functional team whose members aren't helping each other, this study gives you another tool. Look into how much they interact to get their individual jobs done. If the answer is "not much," changing that may reduce the problem. As the authors suggest, "This can be realized by redesigning individual jobs, or adapting the rules regarding how task-related decisions are made, in such a way that subtasks can no longer be accomplished by a single individual."
One way provides extra benefits. Have people cross-train to fill in for people in different skill areas. Even if the skills sets are very different, there have to be some low-level tasks in each function that the other person can do. Find those, no matter how trivial they may be. Another approach is to find opportunities to make people share duties. For example, a bookkeeper who handles payroll and an HR coordinator may rarely interact in their daily duties and consider themselves to have different skills, yet in many companies both would be on a "Support Services" team. Instead of having the coordinator talk about paychecks during new-employee orientations, require the two to work together. If you also train them to do at least the basic tasks of each other and have them serve as helpers when one is overloaded, the research literature says you will gain three major elements for improving team performance: lower perceived differences in skills, higher reliance on each other to complete their work, and more helping behaviors.
Source: Van der Vegt, G., and E. Van de Vliert (05), "Effects of Perceived Skill Dissimilarity and Task Interdependence on Helping in Work Teams," Journal of Management 31(1):73.
Study: One nagging problem in research about teams is the fact that in most workplace studies, not every team member is willing to participate. This is more of a problem for me than for you, since I try to filter out "bad science" for you. But a clever look into the issue by a decision sciences professor does have something to say to team leaders and HR departments.
Thomas Timmerman dug into sports statistics to find out what happens to data about a team when you subtract the data provided by less-cooperative members of that team. People who are less cooperative with teammates are also less likely to fill out the surveys or give the interviews that are a big source of our information about teamwork. So if their answers would be a lot different from the answers a researcher does get back, then the researcher cannot honestly claim the returned answers accurately show what the team is like. (One reason the major political polls are on average pretty accurate in predicting election results is that they adjust their numbers for the likely responses that "nonrespondents" would have given. Differences in those adjustments are part of why polls give somewhat different numbers.)
So Timmerman looked at the statistics from every professional basketball player and team in the U.S. (the National Basketball Association) from 1983 through 1998 (6,207 players). The key number was for "assists"—passes whose receivers then scored. He used this as a model for cooperation that lead to positive results for the team. He compared the number of assists and the average experience of the players with the teams' winning percentages. For further comparison, because baseball players do not have to interact at the same level of intensity as basketball players, Timmerman also looked at the team experience issue using Major League Baseball players from 1902 to 2002.
First, his number-crunching confirmed that basketball teams with more assists (as a percentage of total points scored) had the best winning records. Also not surprising was that teams with more experienced team members did better. Timmerman worked the numbers so that individual performance results were ignored, since, obviously, high individual performance would increase team performance. Even after thus "controlling" for individual performance, teamwork experience by itself was a positive factor in team performance, he says.
Anyone worried about the quality of employee survey data should pay special attention here. Within each team he dropped out the results from 10 percent of the players, then 20 percent, and so on, both randomly and after ranking the players by various categories. It turned out that he could cut up to 30 percent of the uncooperative players and have the team statistics for the rest of the team be basically the same as the "true" number that included every player. Beyond that, averaging the remaining players' statistics did not provide an accurate picture of the team.
Application: Cooperation was as helpful to team performance on the court as in the conference room. In the classic basketball movie "Hoosiers," based on a true story, the high school coach played by Gene Hackman orders his small-town team to make at least four passes before someone takes a shot. Egos on the team were bruised, but the value of this initially forced cooperation proved itself and was part of what led the team to the state championship. Pull out the workflows for your team (since I know you have them handy!) and with the team, see if there are places where handing off work instead of having one role do all of them could speed up the overall process.
Timmerman says the findings that team experience helped team performance even when you factored out individual performance "is consistent with the idea that team experience leads to knowledge, skills, and abilities about teamwork that may lead to better team performance…" In short, when putting together a project team, consider not only a mix of backgrounds and work skills but also of experience with working in teams.
As an ex-journalist reporting the team research, I have always been concerned about studies in which only half the members of a team helped out, and this study shows why. If you conduct surveys in your company, this study should scare you, too. Even if all of your statistical controls make for a good, scientifically valid study (and most of the employee or customer surveys I have seen do not), you can still end up with bogus results if you assume the results from half the team members accurately reflect the team as a whole. Look into ways to increase participation as high as possible, and compare the team results to other metrics. For example, if you do an employee satisfaction study that includes team averages, compare the results to other data like what percentage of team members quit every year ("turnover") or are sick ("absenteeism").
Source: Timmerman, T. (05), "Missing Persons in Studies," Journal of Organizational Behavior 26:21.
Study: You probably share with your close friends a similar view of how the world works. Scientists call this a "mental model." Studies have confirmed that team members who share similar mental models are able to work well together. But of course, some mental models more accurately reflect reality than others (a bigot has a mental model, after all). Organizational psychologists with two American universities and the U.S. Army teamed up to look at how having shared models is affected by the accuracy of those models in affecting team performance.
Undergraduates were recruited to play a warplane simulation game in pairs, with one person piloting and the other controlling weapons. Two trained observers watched videotapes of the pairs and rated each on leadership, assertiveness, decision-making, communication, and other teamwork skills. Team members were asked to record their mental models of the game skills, such as diving/climbing, selecting/shooting weapons, and escaping the enemy. To do so, they rated how much they felt each of the nine skills was related to the other eight. (Presumably, for example, most thought the ability to turn the plane was closely related to escaping the enemy but not to choosing a weapon.) They also filled out a similar scale for the teamwork skills the observers used. In an intriguing twist, the researchers had previously gathered the mental models of experts across the country for comparison: "25 flight simulation enthusiasts" for the game skills, and teamwork researchers for team skills. Just as there can be more than one "right" answer to many business problems, the experts in the two skill groups displayed several mental models for each.
Student pairs with shared mental models about the game skills displayed better observed teamwork and did better at the game. But shared views about teamwork only helped if those views were accurate according to experts. If a pair's models were valid according to at least some experts, but different from each other, team process ratings were low. Even sharing a bad teamwork model was better than having different valid models. And as you'd expect, good team processes were strongly related to better team performance. In sum, if both people came up with equally useful but different answers, teamwork and thus team performance was harmed.
Application: Bear in mind, that these pairs did not have time to work out their different views of the job. Regardless, although the question of shared versus high-quality views is interesting, for practical application I'm always interested in what optimizes team performance. This study was really about pairs of people, not teams, but we know performance is best when people share both an accurate view of their task and of how to accomplish it. That means having your teams take time to A) work on their goals and internal processes, and B) use a project plan approach to problem-solving. Where people are competent but not "in sync" with their teammates, the researchers write, "it may be advisable to use team-training methods to ensure not only that high-quality models are developed, but also that shared models are formed."
Last month I mentioned giving talks at a University of Washington branch. I asked individuals what they did to plan for major papers in the class. Each talked about outlining the basic steps, then setting deadlines for each step to be sure they could complete the work on time. Then I asked if their class teams had done the same thing for their team projects. The general answer was "no." They immediately saw the disconnection. I'm willing to bet the same thing happens in your organization. When I asked a team in each class about its goals, one or two members said it was to get a top grade, but obviously the teams had not discussed those goals. I asked what would happen if a member of the team was married, had kids and full-time job, and only wanted to do enough to get a passing grade. Both views of the importance of grades would be equally correct for the individuals who held them. But clearly, not knowing about those different goals would cause conflicts.
The researchers have an important warning, however: "one needs to be concerned about movement between teams." If you allow teams to develop their views about how things are done in your company in isolation from each other, this "does not ensure that synthesis will exist if members participate in multiple teams simultaneously or if they rotate to a new team…" If you have influence over your organization's general approach to training, this warning was written for you!
Source: Mathieu, et al. (05), "Scaling the Quality of Teammates' Mental Models: Equifinity and Normative Comparisons," Journal of Organizational Behavior 26:37.
TeamResearch News summarizes the latest information from studies or articles on business teams, along with guidance on how to apply that research in your workplace. It is published the first full weekend of each month as a free service from TeamTrainersTM Consulting (www.suddenteams.com). Learn how to subscribe below and see the newsletter Web page for details about the newsletter, cautions about studies, and our privacy policy.
You may reprint any part of this newsletter at no charge in any electronic or print publication for which authors do not receive pay for individual articles, subject to the conditions below. This includes but is not limited to company and nonprofit organization newsletters, Web sites, intranets, etc. If authors are sometimes paid on a per-word or per-article basis, contact me to arrange reprint rights. The conditions for free publication are:
The right to reprint does not imply the granting of any other rights, and all copyrights remain the property of Jim Morgan.
Plain-text e-mail announcements are sent at no charge to subscribers whenever a new issue is posted, containing a list of that month's studies and articles and a link to the newsletter. To:
Your questions and suggestions are always welcome. Contact:
Jim Morgan
Head Coach, TeamTrainers Consulting
(425)
770-8595
jim@suddenteams.com
www.suddenteams.com
All content in this newsletter, including the title, is Copyright 2003-5 by Jim Morgan dba TeamTrainers Consulting. All rights reserved.
"SuddenTeams" is a registered trademark (US Trademark #2,456,849) and "TeamTrainers" and "The Science of Teams" are trademarks of Jim Morgan dba TeamTrainers Consulting.
TeamTrainers Consulting makes no guarantee or warranty regarding the use of information in this newsletter by individuals not employed by or under contract to TeamTrainers Consulting and performing official TeamTrainers business.