Bachelor Thesis
The effect of gender diversity on team performance
AbstractThis thesis is an empirical study to the effect of gender diversity on team performance under different team environments. The goal is to find the best way to design, manage and motivate teams to generate the highest team performance. This is important because the work force is changing, since more and more females are entering labor market. An online experiment is generated consisting of 56 teams participating in a simple task. Only negligible effects were analyzed in the data from the experiment for the effect of gender diversity on team performance. Especially, it shows some suggestive evidence that the effect of gender diversity in team composition on team performance depends on the team environment. Those findings can be used as guidance for further research on team composition and team incentive research. Name: Lisanne Reizevoort Student number: 10581847 Program: Economics and Business - Finance and Organization Supervisor: Mr. Junze Sun Date: January 31st 2016
Statement of Originality This document is written by Lisanne Reizevoort who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Table of contents Abstract ... 1
1. Introduction ... 4
2. Literature review ... 6
2.1 Teams and rewards ... 6
2.2 Gender diversity and performance ... 6
2.3 Gender gap in different environments ... 8
3. Research method ... 10
3.1 Data collection and survey ... 10
3.2 Research setting ... 11
3.3 Descriptive statistics ... 12
3.4 Data variables ... 12
4. Results ... 13
4.1 Results on individual-level ... 13
4.2 Results on team-level ... 15
5. Conclusion and Discussion ... 18
5.1 Implications ... 19
5.2 Limitations and further research ... 19
References ... 21
Appendices ... 23
Appendix 1: Math skills ... 23
Appendix 2: Math grades in high school ... 23
Appendix 3: Survey ... 24
1. Introduction
Nowadays, teamwork is more common in all kind of organizations than several years ago. In businesses colleagues collaborate to find efficient ways to serve customers, at school students work together in teams to do a project or make an assignment, and at sport clubs people are together in a team to play the sport they like. All these different settings of teamwork include decision-making, risk-taking, supporting each other, and setting and achieving goals. These operations can have some negative outcomes and lead to conflicts and misperceptions among team members (Campion, Medsker & Higgs, 1993). Therefore, it is important that teams are managed well and stay motivated. One way to motivate team members is by giving rewards. A lot of previous research has been done to different team reward systems (Garbers & Konradt, 2014; McClurg, 2001; Shaw et al., 2001; Hamilton, 2003). Two common incentive schemes are revenue sharing and a bonus pay. With revenue-sharing the pay-off is based on total team performance and equally divided between team members. With bonus-pay the team that performs best compared to the other teams receives a bonus.
Organizations are facing more competitive and technological challenges. Teams are a solution to these challenges, because of diversity in the personality of team members (Neuman, Wagner & Christiansen, 1999). Besides, organizations see changes in workforce demographics and new organizational forms. This leads to an increase in the diversity of teams (Jackson, May & Whitney, 1995). Each member contributes unique attributes to the team and a rich pool of knowledge is created, so dealing with problems becomes easier and faster (Neuman et al, 1999). To obtain favorable outcomes, it is important to know which design of diversity leads to the highest team performance (Campion et al., 1993).
One part of the design of diversity is the gender composition of teams. Jackson et al. (1995) found that more and more females are taking part in professional teams, because they are entering the labor force in a growing number, and they have a positive effect on team performance in the higher-level functions (Adams & Ferreira, 2009). The changing labor force affects behavior of people and teams; this makes it important to know which composition of females and males is ‘the best’ nowadays. Therefore, this thesis focuses on gender diversity in teams.
A lot of papers have already studied the performance between gender in higher level teams, but only a few have looked into the differences in performance among different team compositions. Nevertheless, some studies found a positive effect of gender diversity on performance (Hoogendoorn et al., 2013), and others found a negative effect (Tsui, Egan & O’reilly, 1992). So, the results are mixed and there is no clear conclusion whether heterogeneous teams outperform homogeneous teams. Therefore, this thesis includes two different incentive schemes: revenue-sharing in cooperation and team competition in competition. Researchers found that males and females perform differently under cooperation and competition (Nowell & Tinkler,
1994; Gneezy, 2001; Healy & Pate, 2011). This makes it possible to investigate which team composition generates the largest team performance under which incentive scheme. The research question to be answered will be: Does gender diversity in teams lead to differences in team performance under different incentive schemes? This thesis will contribute to previous literature in finding the best way to design, manage and motivate teams to generate a large team performance by including incentive schemes. To achieve this, a small experiment is designed. Earlier studies often used board performance or firm value in a business game as performance measure. This experiment is mostly based on the experiment of Ivanova-Stenzel and Kübler (2011), and will use the amount of correctly solved problems as a performance measure. Instead of the memory games used in Ivanova-Stenzel and Kübler’s (2011) experiment, in this thesis arithmetic problems will be used. This because males and females perform differently in memory games (Tottenham, Saucier, Elias, & Gutwin, 2003), and they perform the same in arithmetic problems in a neutral situation (Hyde, Fennema, & Lamon, 1990). The data from this experiment can be used to measure the difference in performance between males and females, but also the differences in team performance. These results can be useful for leaders of the teams, when they have to decide which team composition and incentives scheme works best in their organization.
The analysis of the data provides suggestive evidence that differences in team performances can occur by different incentive schemes and team compositions. The results show that female teams are outperformed by male teams and mixed teams in both cooperation and competition. However, the performance of male teams and mixed teams is almost the same in competition, while the performance of male teams is much larger than mixed teams in cooperation. Overall, the average performance of single-sex teams was higher under competition than under cooperation, while the performance of mixed teams was smaller under competition than under cooperation. This resulted in a suggestive larger effect of team diversity on the total performance under cooperation than under competition. The remainder of this thesis consists of four sections. In section 2, an extensive literature review is provided related to this topic, including the hypothesis following from this literature. In section 3, the research methodology and variables are described in more detail. In section 4, the results of the experiment are presented. And in section 5, the results are summed up, and a discussion and some limitations are provided.
2. Literature review 2.1 Teams and rewards Organizations acknowledge the additional value that teams add to an organization, and therefore there is a large increase in teamwork. This increase in teamwork could be explained by the fact that people are willing to exert more effort in teams than when they work individually on a task, whereby individual performance increases (Deutsch, 1949). In addition, Hertel (2011) found that in teams, people can bundle their know-how, skills, and creativity, what could lead to good decision-making and in the end to better team performances.
Because organizations are more interested in teams, teams-rewards are increasingly implemented in organizations (Kerrin & Oliver, 2002). Giving rewards are a strong method to increase individual motivation and to affect productivity and performance (Garbers & Konradt, 2013). The effect of rewards on productivity has been investigated by a lot of studies (Rosenbaum et al., 1980; Nalbantian & Schotter, 1997; Dohmen & Falk, 2011), but the effect on performance only by a few. The studies focusing on performance made a distinction between individual and team rewards, and found that team rewards have a larger positive effect on performance than individual rewards (Condly, Clark & Stolovitch, 2003; Garbers & Konradt, 2013). However, team-based rewards could also lead to the free-rider problem: when one team member exerts more effort, the incentives for the second team member to exert effort will be removed, and he will benefit from this (Klor et al., 2014). But, this problem will be mitigated when team members motivate each other to achieve a group norm (Kandel & Lazear, 1992).
This thesis will include team-based rewards, instead of individual pay, because this is more in the interest of organizations nowadays. In addition to Garbers and Konradt (2013) finding that the effect of team-based rewards on performance in heterogeneous team than in homogeneous teams, this thesis will make a distinction in the effect of two different team-based reward schemes in different team compositions. This can lead to new and specific insights which incentive schemes generate the best outcomes in heterogeneous teams and homogeneous teams, and which incentive scheme can best be applied to already selected team members, or to people who are already working for the organization. 2.2 Gender diversity and performance
Teams generate different outcomes, because of diversity among team members. Milliken and Martins (1996) stated that diversity can be categorized in two different types: observable and less visible diversity. Observable diversity, also known as demographic diversity, includes race, gender, ethnic background, and age. And less visible diversity, also known as cognitive diversity, includes personality characteristics, education, technical abilities, and values. Because of these
differences between team members, they may have different perspectives to key issues and problems, and different preferences in the way of interaction (Milliken and Martins, 1996). This thesis focuses on the effect of gender diversity in team composition. Previous studies showed that males and female behave differently to different team compositions (Tsui et al., 1992; Ivanova-Stenzel & Kübler, 2011). For example, Tsui et al. (1992) show that people working in heterogeneous teams are less attached to their organization, and are on average more absent. They also found that males performed worse than females when their gender is of minority in a team. So, in their experiment diverse teams have a negative effect on performance.
Ivanova-Stenzel and Kübler (2011) also studied the effect of gender diversity on performance of males and female, and included two environments: cooperation and competition. 240 students were randomly assigned to the revenue-sharing treatment or the tournament treatment, and randomly matched with another participant, resulting in three different team compositions. The experiment contained a real task in which individuals had to solve as many memory games as possible in 15 minutes. The results show that under the revenue-sharing treatment males and female perform differently in mixed teams, but not in single-sex teams. And, under the tournament treatment males and females perform differently when single-female have to compete with single-female teams, or single-male teams have to compete with single-male teams.
Although Ivanova-Stenzel and Kübler (2011) investigated individual differences in performance in detail, they did not focus on the differences in total team performance. This thesis will examine this. Although people perform individually differently, the particular team composition and incentive scheme could still generate the largest possible team performance and high-quality solutions. This is of large value for the overall performance of an organization (Watson, Kumar & Michaelsen, 1993). Various empirical studies have already investigated the effect of gender diversity on team performance (Hoogendoorn, Oosterbeek & Van Praag, 2013; Apesteguia et al., 2012; Tsui et al., 1992). Hoogendoorn et al. (2013) worked together with a leading entrepreneurship education program, in which teams of students had to start and run a small-sized company for an entire academic year. 550 students were assigned to 45 teams. Results were analyzed by observing the relation between sales and profits and the share of women in a team. Observation are graphed in an inverse U-shape what means that profits and sales are increasing until the point that the share of women reach 0.5, and profits become flat afterwards, while sales decreases. On average, mixed gender teams (a share of women between 0.4 and 0.6) generated higher profits than teams dominated by males or females (no single-sex teams included). Hoogendoorn et al. (2013) considered different mechanisms to find an explanation for their results. Mechanisms include learning from team members, and relationships, conflicts, decision-making and mutual
monitoring within teams. However, they did not find a significant relation between these mechanisms and gender composition in teams. Apesteguia et al. (2012) also studied a business game, the L’Oréal e-Strat Challenge. They included a prize of €10,000 for the best performing team. Undergraduate and MBA students were assigned to three-member teams, and acted like a general manager of a beauty-industry company. Teams were divided in four different composition categories, and had to compete amongst each other. They had to make real business decisions which affected the market value of the firm. In the end, the market value was used to measure the team performances. The main result of the experiment is that teams consisting of three women are outperformed by all other teams. This could be explained by gender differences in decision-making. In the game three-women teams were less aggressive in their pricing strategy, and invested less in research and development and more in social sustainability initiatives. 2.3 Gender gap in different environments
Besides a different behavior to various team compositions, males and females are different in several other ways. They have different social preferences, competitive preferences and risk preferences (Croson & Gneezy, 2009). For instance, Croson and Gneezy (2009) show that females are more risk averse than males, and that females’ behavior changes faster because they are more sensitive for social signs than males. Because males and females have different preferences, they also perform differently in teams (Nowell & Tinkler, 1994; Dufwenberg & Muren, 2006). In teams, differences in productivity and effort arise between males and females. This can be explained as the gender gap in performance.
In organizations, teams have to perform under different environments. The two main categories are the cooperative environment and competitive environment. In these environments, incentive schemes are often used to motivate people to cooperate or compete with team members or other teams.1 In the cooperative environment, teams have to generate a high performance by working together, helping each other, and making decisions together. To motivate these teams, people can be rewarded by a revenue-sharing incentive scheme. Literature have already studied the gender gap in performance in cooperation. Nowell and Tinkler (1994) found that females contribute more to cooperative teams than males, and that females are more helpful. In addition, Dufwenberg and Muren (2006) have investigated the gender gap in decision-making of teams. 168 participants took part in a dictator game in teams of three. There were four different gender compositions
1 The environments, cooperation and competition, and incentive schemes, revenue-sharing and team competition, are used interchangeably.
possible. Every team got 1000kr to divide between the team and an additional pre-selected person. The money assigned to the team would be equally divided. The results show that female majority teams are more generous in giving money to the pre-selected person and that they also choose equalitarian division more often than male-majority teams. In the competitive environment, individuals or teams have to compete among each other. To motivate these teams, people can be rewarded by a team competition incentive scheme, in which the pay-off is based on competition between teams. Empirical evidence shows a gender gap in performance in competitive environments. Gneezy (2001) found that men perform much better while women perform the same in individual competition in comparison to cooperation. An explanation for this is that men are more in favor of competition than women (Niederle & Vesterlund, 2007; Buser, Niederle & Oosterbeek, 2014; Healy and Pate, 2011). Niederle and Vesterlund’s (2007) experiment shows that 73 percent of men choose competition and only 35 percent of women, based on the fact that men are more overconfident than women, and women are more uncertain about their own qualities. Buser et al. (2014) conducted an experiment similar to Niederle and Vesterlund’s (2007), and found that 49 percent of males and only 23 percent of females choose to perform in competition. They also found that on average males performed slightly better than females in competition, so a small gender gap in performance, however this gender gap is not significant.
Healy and Pate (2011) have investigated if teams could help to close the gender gap in performance in competitive environments. They focused on the individual performance of people performing on their own or in a team, in the cooperative environment and in the competitive environment. They first found that with inclusion of teams the difference between males and females in their preference for competition can be reduced. By focusing on the environments, they found that in teams, males solved on average slightly more problems than females. So, the performance of males was slightly better than females in both cooperation and competition.
The gender gap in performance under different team environments will be will deepened in this thesis by focusing on the composition of teams. The small gender gap in performance in competition and the larger gender gap in cooperation can be explained in more detail by investigating the performance under different team compositions. In summary, the literature shows that males and females behave differently to different incentive schemes, but also to different team compositions. It also shows that different team compositions affect team performance. However, the conclusions are really inconsistent and there is not one big solution which team composition is the ‘best’ to generate the largest team performance. So, this thesis tries to find this solution by including all factors in one experiment. It also focuses on how males and females behave differently in the different environment and in different team
compositions, and on how team composition and incentive schemes affect the total team performance. In line with the literature above, the following hypotheses are stated: Hypothesis 1: In the cooperative environment, females perform overall better than males. Hypothesis 2: In the competitive environment, males perform overall better than females. Hypothesis 3: In the cooperative environment, gender diversity in team composition has a positive effect on team performance. Hypothesis 4: In the competitive environment, gender diversity in team composition has a positive effect on team performance. Hypothesis 5: Gender diversity in team composition has a larger effect on team performance under the competitive environment than under the cooperative environment. The fifth hypothesis has not been investigated yet, so this could generate new insights in what would be the best team environment to obtain the highest performance in organizations. 3. Research method 3.1 Data collection and survey
This thesis performs an empirical study to investigate the effect of team diversity on team performance. The data used in this research is obtained by an online questionnaire. Participants took part in this survey voluntary. The focus laid on subjects between an age of twenty and thirty to enhance generalizability of the results. The survey was provided in English, and consisted of two parts. The first part included a questionnaire and the second part a short experiment. The first part focused on four control variables. It included some general questions about gender, age, the level of schooling, the level of math skills, and the grade of mathematics in high school. Participants had to give an indication of their math skills through a five-point Likert-type scale, ranging from 1 (extremely bad) to 5 (extremely good). The level of schooling was the highest level obtained or the level still in, ranging from 1 (high school) to 3 (master degree). The average grade for mathematics obtained in high school was ranged from 1 (extremely bad) to 10 (excellent).
The second part consisted of a small experiment. An introductory page was included in the experiment to prime the participants with different treatments, so the differences in performance per treatment can be analyzed in the end. Participants were primed in two different ways. First, they were assigned to one of the two incentive schemes, revenue-sharing or team competition, which were used to prime respectively for cooperation or competition. Second, they were randomly matched with Lisa or Tim. These names were used to prime respectively for a female team partner or a male team partner. This results in three possible team compositions: female-female, male-male, and male-female/female-male. The partner treatment makes it possible to test for differences in individual performance between males and females among the different team compositions, but also to test the effect of various team compositions on the total team performance.
In the revenue-sharing treatment, used as prime for cooperation, participants were told that the total performance would be measured by the total number of correctly answered questions by oneself and its team partner. Every correct answer was worth €0.25, and the total revenue would be equally split between team members. In the team competition treatment, used as prime for competition, participants were told that the total performance would be measured by the total number of correctly answered questions by oneself and its team partner and that each correct answer was worth €0.25, but also that the total team performance would be compared with the total team performance of another randomly chosen team. The team with the higher performance will receive the payoff, which will be equally split between team members. After reading one of these texts, participants had to do a simple task in which they had to solve as many arithmetic problems as possible in a time limit of 3 minutes. All participants made the task on their own, so they chose their effort level without knowledge of the choice made by the team partner. The arithmetic problems include multiplying, dividing, subtracting or summing 1-digit, 2-digit, or 3-digit numbers. A mathematical task is included because on average males and females perform quite the same on mathematical tasks in a neutral situation (Hyde, Fennema, and Lamon, 1990), so there are no influences from the particular task in showing gender differences in performance. In addition, this task requires full concentration like a real-world case, but does not cause any differences in interpretation or perception that could affect the outcomes. In the final part, participants were asked to fill out their name and e-mail address if they want to have a chance to win the monetary payoff. Two randomly chosen teams are selected from the sample that filled out their names to be paid out. 3.2 Research setting A quantitative analysis is used to test the theoretical hypotheses. One advantage of quantitative research is that numerical data is collected and can be used for statistical analysis. The design of
this empirical research is mostly based on an earlier research done by Ivanova-Stenzel and Kübler (2011). The data for this analysis is obtained by a survey generated via Qualtrics, an online survey software. Via social media and e-mail individuals between an age of twenty and thirty were approached to participate in this experiment. All individuals got the same information about the study. The survey was open for a time period of one week. 3.3 Descriptive statistics The data originally consisted of 117 participants. Qualtrics randomly assigned people to one of the treatments, resulting in 56 participants in the revenue-sharing treatment and 61 participants in the team competition treatment. The sample has a male fraction of 45% and a female fraction of 55%. The average age of the participants is 22 years and the majority is still in or has already obtained the bachelor degree.
The sample is divided over 56 teams of two individuals. Five participants are dropped because of disability to form teams, due to different treatments and an unequal division of male and female. 30% of the teams are female teams, 25% male teams, and 45% mixed teams. 3.4 Data variables The variables used in this thesis are used at individual-level and at team-level. Performance. Performance is the dependent variable in both models. It is measured in the amount of correct answered arithmetic problems. Every team member obtains its own amount and by counting the individual performance of the two team members the total team performance is measured. Table 1 Distribution of male and female, and teams per treatment
Incentive scheme Team composition Females Males Participants Dropped
females Teams Revenue-sharing Female teams 18 0 18 9 Male teams 0 14 14 7 Mixed teams 14 10 24 4 10 All 32 24 56 4 26 Team competition Female teams 16 0 16 8 Male teams 0 14 14 7 Mixed teams 16 15 31 1 15 All 32 29 61 1 30 All 64 53 117 5 56
Gender. This variable is the first independent variable. It is a dummy variable and gets
value 1 for males and value 0 for females. It can help to find the overall gender gap in performance on individual-level. Therefore, this variable is not included in the regression on team-level.
Treatment. This is the second independent variable. There are two different treatments
included in the experiment: revenue-sharing for the cooperative environment and team competition for competitive environment. A dummy variable is included to show if there is a difference in performance between the two incentive schemes on individual-level and on team-level. The variable gets a value 1 for team competition and value 0 for revenue-sharing.
Team composition. This is the third independent variable, and also a dummy variable.
Individuals are randomly assigned to a male or female partner, resulting in three different team compositions. The variable team composition is divided into mixed teams, value 1, and single-sex teams, value 0. This variable can demonstrate a potential effect of team composition on individual and team performance. 4. Results
Table 2 and 4 show the mean performances per team composition and incentive scheme. In particular, table 2 shows the mean performance on individual-level, so the performance of males and females can be compared. And table 4 shows the mean performance on team-level, so the team performances can be compared. To test these mean differences in performance, the Welch’s t-test is used. In comparison to the Student’s t-test, this test includes unequal variances sample means. The variance-comparison test has shown that the variances across the teams are not equal. 4.1 Results on individual-level
Hypothesis 1 predicted that females perform better than males in both single-sex and mixed teams under cooperation. Table 2 shows that in mixed teams males performed on average worse than females (x#$%&'= 13.70 < x/&#$%&'= 15.20: p = 0.596). However, the results are only partially in line with hypothesis 1, because in single-sex teams males performed on average better than females (x#$%&'= 14.43 > x/&#$%&'= 12.22: p = 0.368). These differences in performance have led to a large difference in the gender gap between single-sex and mixed teams GG'<=>%&=
2.21 > GG#<?&@= −1.50 . An explanation for these results could be that people cooperate more with males than with females (Andreoni & Petrie, 2008). Table 2 shows a similar result, because males and females performed both better when their team partner was a male.
Hypothesis 2 predicted that males perform better than females in both single-sex and mixed teams under competition. Table 2 shows that males performed on average better than females in both single-sex as mixed teams (single: x#= 15.07 > x/= 12.81: p = 0.366; mixed: x#=
2. Table 2 also shows that both males and females performed better in single-sex teams than in mixed teams in team competition. An explanation for this could be that males like competition more than females, and therefor are more inclined to compete (Gneezy, Niederle & Rustichini, 2003). Females in single-sex teams are surrounded by other females, who are less inclined to compete, therefor they feel more comfortable and they generate a higher productivity than in mixed teams.
These differences in individual performance can also be caused by differences in math grade in high school, math skills and level of schooling. Table 3 shows ordinary least squares regressions on individual-level. In model 6 control variables are added to the basic regression of model 4. Model 6 shows that level of schooling and level of skills have a significant effect on team performance. When controlling for these variables, the sign of the mixed team coefficient changed2. This implies that there is a possibility that the differences in individual performance between mixed teams and single-sex teams is caused by the level of schooling and the level of skills. Table 2 Mean individual performance by gender Incentive scheme Team composition Male Female Gender gap Revenue-sharing Single-sex teams 14.43 12.22 2.21 (7.122) (6.292) Mixed teams 13.70 15.20 -1.50 (7.689) (4.185) All 14.13 13.29 0.84 (7.207) (5.734) Team competition Single-sex teams 15.07 12.81 2.26 (7.820) (5.115) Mixed teams 14.73 12.27 2.46 (6.204) (5.378) All 14.9 12.55 2.35 (6.904) (5.163) All 14.55 12.90 1.65 (6.985) (5.407) NOTES: The value between parenthesis indicates the standard deviations. The weighted average is used to calculate the mean for the ‘all’ groups.
2
p = 0.838
Table 3 Regression on individual performance Independent variables (1) (2) (3) (4) (5) (6) (7) Male 1.649 1.636 1.519 1.738 0.818 (1.272) (1.286) (2.216) (1.320) (2.032) Team competition 0.010 -0.059 -0.870 -0.180 -1.435 (1.175) (1.168) (1.426) (1.097) (1.411) Mixed teams 0.364 0.292 1.053 -0.211 0.224 (1.148) (1.154) (1.391) (1.661) (1.374) Male x team competition 1.514 (3.313) 2.376 (2.943) Male x mixed teams -1.781 (3.451) -1.229 (3.152) Male x team competition x mixed teams 0.390 (4.153) 0.599 (3.712) Bachelor degree 2.761* 2.545 (1.661) (1.632) Master degree 4.878** 4.817** (2.125) (2.177) Math grade 0.784 0.716 (0.748) (0.729) Level of math skills 2.271** 2.429*** (0.894) (0.876) Constant 12.898*** 13.673*** 13.516*** 12.806*** 12.910*** -3.414 -2.817 (0.711) (0.843) (0.876) (1.100) (1.361) (4.874) (4.827) R2 0.018 0.000 0.001 0.019 0.026 0.208 0.220 F-statistic 1.68 0.00 0.10 0.60 0.52 3.37 2.97 NOTE: The regression includes clusters for teams. The values between parenthesis show the standard robust errors. *** significance at 1 percent level / ** significance at 5 percent level / * significance at 10 percent level. 4.2 Results on team-level Hypothesis 3 predicted that gender diversity in team composition has a positive effect on team performance under cooperation. This implies that the team performance should be larger in mixed teams than in single-sex teams in cooperation. Table 4 shows that the team performance of mixed teams is on average larger than the team performance in single-sex teams (xmixed= 28.90 > xsingle = 26.38). The single-sex teams are divided into female teams and male teams. Results show that female teams are outperformed by both male teams and mixed teams under cooperation
(x/&#$%&'= 24.44 < x#$%&'= 28.86: p = 0.379; (x/&#$%&'= 24.44 < x#<?&@= 28.90: p = 0.273), and that the difference between diverse teams and male teams is nullifying (xmixed= 28.90 > xmales =
28.86: p = 0.992). So, these results are not in line with hypothesis 3. Apesteguia et al. (2012) also found that female teams were outperformed by all other teams in competition. They explained this by the fact that female teams made decisions which lead to lower market value than the decisions of the other teams. In this case, a possible explanation of the differences in team performance are differences in individual performance between gender. Table 2 shows that females performed much better in mixed teams than in single-sex teams (x#<?&@= 15.20 > x'<=>%&=
teams (x#<?&@= 13.70 < x'<=>%&= 14.43: p = 0.816). Therefor the average individual performance in mixed teams (x#<?&@= 14.45) is much better than the average individual performance in female teams and slightly better than the average individual performance in male teams. These differences in individual performance directly affect the total team performances. In addition, individual math skills3 can be used as proxy for individual performance. The math skills of males are only slightly better than the math skills of females in the cooperation treatment. So, nothing can be concluded from differences in math skills. Hypothesis 4 predicted that gender diversity in team composition has a positive effect on team performance under competition. This prediction is not supported by the results, because table 4 shows that male teams outperformed both female teams and mixed teams under team competition (x#$%&'= 30.14 > x#<?&@= 27.00 > x/&#$%&'= 25.63). This can also be explained by differences in individual performance. Table 2 shows that females and males performed individually better in single-sex teams than in mixed teams under competition. However, in mixed teams, the individual male performance is larger (x#$%&'= 14.73) than the individual female
performance (x/&#$%&'= 12.27). Therefore, the average total team performance of mixed teams
increased and became larger than the average total team performance of female teams. In addition, males performed on average better in single-sex teams than in mixed teams, therefore the total team performance in single male teams is larger than in mixed teams. In addition, table 5 shows the results of the ordinary least squares regression on team-level. The regressions include control variable for math skills, level of education and math grade, which can be used as proxies for individual ability. These control variables make the effects more accurate. Table 5 model 5 include an interaction variable for mixed teams and team competition. It shows a negative sign in front of the coefficient (p = 0.489). This gives suggestive evidence that the effect of team competition on team performance is smaller in mixed teams than in single-sex teams. This is the opposite of what was predicted by hypothesis 4. In model 7 of table 5 is controlled for math grade, that has a significant positive effect on team performance. However, by controlling for math grade, the signs of the coefficients did not change and they are still insignificant. So, math grade as proxy of individual ability is not the main driver of the differences in team performance. The differences in individual performance in both cooperation and competition can also be explained by the fact that females are more risk averse than males. For instance, Gneezy et al. (2003) found that females and males behave differently to uncertainty in payments when effort is costly. In this experiment, payments are uncertain, because in revenue-sharing participants’ payments are depending on the performance of the team partner, and in team competition they
3
See Appendix 1: Math skills
are depending on team partner and on the performance of the team they compete against. However, risk aversion is not measured, so could not be included and tested in the regressions. Hypothesis 5 predicted that gender diversity in team composition has a larger effect on team performance in competition than in cooperation. Table 4 shows the average differences in team performance between single-sex teams and mixed teams. In cooperation, team performance increased by 2.53 with more diversity in teams, while in competition team performance decreased by 0.73. In addition, in regression 5 and 7 in table 5 the interaction variable mixed teams x team competition shows how team diversity is dependent on team composition. The coefficient of this variable has a negative sign (p = 0.489). These two results show suggestive evidence that team
diversity effect is smaller in competition than in cooperation, and that hypothesis 5 is not supported.
Although the differences in team performances are not that large between the two incentive schemes, from these results can be concluded that in mixed teams a revenue-sharing incentive scheme generates the largest total team performance, in male teams a team competition incentive scheme and in female teams also a team competition incentive scheme.
Overall, table 4 shows that the performance of mixed teams is slightly larger than the performance of single-sex teams. From this can be concluded that team diversity has a small, not significant, positive effect on team performance (x#<?&@= 27.76 > x'<=>%&= 27.03).
Table 4
Mean team performance
Incentive scheme Team composition Mean team performance Weighted average single-sex teams Difference between single-sex and mixed teams Revenue-sharing Female teams 24.44 26.37 +2.53 (9.684) Male teams 28.86 (9.634) Mixed teams 28.90 (7.047) All 27.35 (8.648) Team competition Female teams 25.63 27.73 -0.73 (6.844) Male teams 30.14 (13.359) Mixed teams 27.00 (7.874) All 27.37 (8.996) All Single-sex teams 27.03 +0.73 (9.786) Mixed teams 27.76 (7.463)
Table 5 Regression on team performance Variables (1) (2) (3) (4) (5) (6) (7) Team competition 0.021 -0.064 1.358 1.853 3.290 (2.361) (2.340) (3.589) (2.219) (3.21) Mixed teams 0.728 0.735 2.525 -0.417 1.390 (2.306) (2.344) (3.261) (2.163) (3.226) Math grade 5.701*** 6.031*** 6.033*** (1.647) (1.670) (1.726) Mixed teams x team competition -3.258 (4.674) -3.290 (4.343) Constant 27.346*** 27.032*** -14.605 27.063*** 26.375*** -17.839 -18.551 (1.694) (1.761) (11.741) (2.070) (2.413) (12.073) (12.777) R2 0.002 0.002 0.120 0.002 0.010 0.21 0.22 F-statistic 0.10 0.09 11.98*** 0.05 0.23 4.40*** 3.16** NOTE: The value between parenthesis indicates the standard robust errors. *** significance at 1 percent level / ** significance at 5 percent level / * significance at 10 percent level. 5. Conclusion and Discussion The purpose of this thesis was to answer the question how team diversity could affect the total team performance under different incentive schemes. Team performance was measured by the amount of correctly solved arithmetic problems in the online experiment. Teams performed under two different incentive schemes: revenue-sharing or team competition. And, the gender composition of the teams was an indication of team diversity, including female teams, male teams and mixed teams. In total, data from 56 teams of two individuals was collected, and analyzed on both individual, and team-level.
Some conclusions can be drawn from these analyses. However, most of the results are not significant, so all evidence is suggestive. The main finding is that gender diversity can both increase and decrease team performance. The effect depends on the incentive scheme by which teams are paid out. Overall, the results show a not significant positive effect of gender diversity on team performance. In the cooperation environment where teams are paid by a revenue-sharing incentive scheme, results show suggestive evidence for a positive effect of gender diversity. Some previous studies show the same result, for instance Lee and Fahr (2004) and Hoogendoorn et al. (2013). However, within the single-sex teams, male teams performed ways better than female teams. These results are affected by individual performance. The analysis on individual level shows evidence that people cooperate more when they are in a team with males than with females.
The result from the competition environment in which teams are paid by a team competition incentive scheme show that team diversity has a small negative effect on team performance. However, male teams performed much better than both female teams and mixed
teams. This could be explained by the result that males individually performed much better than females in competition. This can be caused by the fact that males like to compete and females do not like to compete with males (Gneezy et al., 2003). However, preferences are not taken into account during the experiment, so cannot be analyzed. 5.1 Implications The finding of these thesis can be used in all kind of organizations, especially in organizations in which incentive schemes are used. Depending on the team composition, an employer can choose to implement the revenue-sharing or the team competition incentive schemes. The findings can also be used in the selection process of people in organizations. Organizations can hire people who suit best with the already implemented incentive scheme. This makes is possible to create, motivate and keep effective teams. 5.2 Limitations and further research Even though this thesis shows some useful results, the included experiment has some limitations. One of the limitation is that the sample of the experiment only includes people between an age of 20 and 30, and mostly still in their bachelor or master. Because organizations hire people between an age of 16 and 68 and also people less educated than a bachelor or master degree, this decreases the generalizability of the results, and endangers the external validity.
There are also some limitations that affect the internal validity. The first one is the exclusion of control groups in the experiment. Now, it is not totally clear if the individual performance is affected by the prime for Lisa or Tim as team partner. The performance could also be affected by the time limit of 3 minutes, because people felt pressure to solve as many problems as possible. One control group could have shown the general performance per incentive scheme, and another the general performance per team composition. In addition, it is also not clear if the incentive scheme treatments really worked. It has been shown that people better react to reward systems when they get feedback on their previous performance. So, by adding an additional task to the experiment before the treatment the general performance can be measured, and feedback on this task can be provided to the participants. Another limitation is size of the sample. Only 56 teams were used. Although the amount of single-sex and mixed teams is almost the same, in the competition treatment there is a large gap in the amount of male or female teams and mixed teams. This made it difficult to make reliable estimations, and could have influence the significance of the results. A third limitation is that teams consisted of two individuals, whereby only three different team composition were possible. This limits the applicability of the results in real-life, because most teams consist of more than two persons. In further research, more people should be assigned
to one team, and the share of female or males can be used to measure team diversity. This will make it possible to say something about performance when a gender in the minority or the majority of a team, and it will make the results more applicable.
These limitations provide some new direction for further research. For instance, other measures of demographic and cognitive diversity can be used. Race, culture and personal characteristics, like agreeableness and extroversion, are also of great importance in the selection process of organization, in the creation of teams, but also important drivers of gender diversity. Moreover, another research setting can make the results more generalizable. For example, teams of real organizations from several branches can be used to investigate how they perform under different incentive schemes. They are used to work under a particular incentive scheme, so future research can investigate how another incentive scheme affects their performances. Concluding, this thesis is a step closer to a better understanding how different incentive schemes and different gender compositions could create and motivate teams in organizations to be as effective as possible.
References
Andreoni, J., & Petrie, R. (2008). Beauty, gender and stereotypes: Evidence from laboratory experiments. Journal of Economic Psychology, Vol. 29, 73–93.
Apesteguia, J., Azmat, G., & Iriberri, N., (2012). The Impact of Gender Composition on Team Performance and Decision Making: Evidence from the Field. Management Science, Vol. 58, 78-93. Buser, T., Niederle, M., & Oosterbeek, H. (2014). Gender, competitiveness, and career choices. The Quarterly Journal of Economics, Vol. 129(3), 1409-1447. Campion, M. A., Medsker, G. J. & Higgs, A. C. (1993). Relations between work group characteristics and effectiveness: Implications for designing effective work groups. Personnel Psychology, Vol. 46, 832-850.
Condly, S. J., Clark, R. E., & Stolovitch, H. D. (2003). The Effects of Incentives on Workplace Performance: A Meta-Analytic Review of Research Studies 1. Performance Improvement Quarterly, Vol. 16(3), 46-63.
Croson, C., & Gneezy, U. (2009). Gender differences in preferences. Journal of Economic Literature, Vol. 47(2), 448-474.
Dohmen, T., & Falk, A. (2011). Performance pay and multidimensional sorting: Productivity, preferences, and gender.The American Economic Review, Vol. 101(2), 556-590.
Dufwenberg, M., & Muren, A. (2006). Gender composition in teams. Journal of Economic Behavior
& Organization, Vol. 61, 50–54.
Garbers, Y., & Konradt, U. (2014). The effect of financial incentives on performance: A quantitative review of individual and team-based financial incentives. Journal of Occupational and
Organizational Psychology, Vol. 87, 102-137.
Gneezy, U., Niederle, M., & Rustichini, A. (2003). Performance in competitive environments: gender differences. Quarterly Journal of Economics, Vol. 118, 1049–1074
Hamilton, B.H., Nickerson, J.A. & Owan, I. (2003). Team incentives and worker heterogeneity: an empirical analysis of the impact of teams on productivity and participation. Journal of
Political Economy, Vol. 111, 465–97.
Healy, A., & Pate, J. (2011). Can teams help to close the gender competition gap? The Economic
Journal, Vol. 121(555), 1192-1204.
Hertel, G. (2011). Synergetic effects in working teams. Journal of Managerial Psychology, Vol. 26(3), 176 – 184. Hoogendoorn, S., Oosterbeek, H., & Van Praag, M. (2013). The Impact of Gender Diversity on the Performance of Business Teams: Evidence from a Field Experiment. Management Science, Vol. 59 (7), 1514-1528. Hyde, J., Fennema, E., & Lamon, S. (1990). Gender differences in mathematics performance: A meta-analysis. Psychological Bulletin, Vol. 107(2), 139-155. Ivanova-Stenzel, R., & Kübler, D. (2011). Gender differences in team work and team competition. Journal of Economic Psychology, Vol 32(5), 797-808.
Jackson, S. E., May, K. E., & Whitney, K. (1995). Understanding the dynamics of diversity in decision-making. Team effectiveness and decision making in organizations, 204-261. Kandel, E., & Lazear, E.P. (1992). Peer Pressure and Partnerships. Journal of Political Economy, Vol. 100(4), 801–17. Kerrin, M., & Oliver, N. (2002). Collective and individual improvement activities: the role of reward systems. Personnel Review, Vol. 31, 320-37. Klor, E. F., Kube, S., Winter, E., & Zultan, R. (2014). Can higher rewards to less effort? Incentive reversal in teams. Journal of Economic Behavior & Organization, Vol. 97, 72-83.
Lee, C., & Fahr, J. (2004). Joint Effects of Group Efficacy and Gender Diversity on Group Cohesion and Performance. Applied Psychology, Vol. 53, 136-154. McClurg, L. N. (2001). Team rewards: How far have we come? Human Resource Management, Vol 40, 73-86. Milliken, F. J., & Martins, L. L. (1996). Searching for common threads: Understanding the multiple effects of diversity in organizational groups. Academy of management review, Vol. 21(2), 402-433.
Nalbantian, H. R., & Schotter, A. (1997). Productivity under group incentives: An experimental study.The American Economic Review, Vol. 87(3), 314-341.
Neuman, G. A., Wagner, S. H., & Christiansen, N. D. (1999). The Relationship Between Work-Team Personality Composition and the Job Performance of Teams. Group and Organization
Management, 24(1), 28-45. Niederle, M., & Vesterlund, L. (2007). Do women shy away from competition? Do men compete to much? The Quarterly Journal of Economics, Vol. 122(3), 1067-1101. Nowell, C., & Tinkler, S. (1994). The influence of gender on the provision of a public good. Journal of Economic Behavior and Organization, Vol. 25, 25–36. Rosenbaum, M. E., Moore, D. L., Cotton, J. L., Cook, M. S., Hieser, R. A., Shovar, M. N., & Gray, M. J. (1980). Group productivity and process: Pure and mixed reward structures and task interdependence.Journal of Personality and Social Psychology, Vol. 39(4), 626-642.
Shaw, J. D., Duffy, M. K., & Stark, E. M. (2001). Team Reward Attitude: Construct Development and Initial Validation. Journal of Organizational Behavior, Vol. 22-8, 903-917.
Tottenham, L. R., Saucier, D., Elias, L., & Gutwin, C. (2003). Female advantage for spatial location memory in both static and dynamic environments. Brain and Cognition, Vol. 53, 381–383.
Tsui, A. S., Egan, T. D., & O'Reilly, C. A. (1992). Being different: Relational demography and
organizational attachment. Administrative Science Quarterly, Vol. 37, 549-579. Watson, W. E., Kumar, K., & Michaelsen, L. K. (1993). Cultural diversity's impact on interaction process and performance: Comparing homogeneous and diverse task groups. Academy of Management Journal, Vol. 36, 590-602.
Appendices Appendix 1: Math skills Appendix 2: Math grades in high school
Appendix 3: Survey
Dear participant,
Thanks for taking part in the survey for my bachelor thesis in the field of Organizational Economics. The survey consists of some general questions and a simple task. It will take no longer than 10 minutes. All responses will be kept strictly confidential. Kind regards, Lisanne Reizevoort Q1: Your gender o Male o Female Q2: Your age _____ Q3: What is highest level of school you have completed / are you in? o Less than High school o High school o Bachelor degree o Master degree Q4: What was your average grade for mathematics in high school? 0---10 Q5: What level are your math skills? o Extremely bad o Bad o On average o Good o Extremely good Q6: Randomly assigned to one of the following treatments:
In the following part, you are invited to complete a simple task in a team together with Tim, another participant in this survey. During the task, you and Tim will get 3 minutes to solve as many arithmetic problems as possible, both on your own. These arithmetic problems include multiplying, dividing, subtracting or summing 1-digit, 2-digit, or 3-digit numbers. In the end, the total team performance will be measured by the total number of correctly answered questions by you and Tim. Every correct answer is worth €0.25, the total revenue will be equally split between team members. Click on the button below, and the 3-minute task will start immediately. You are not allowed to use a calculator.
In the following part, you are invited to complete a simple task in a team together with Lisa, another participant in this survey. During the task, you and Lisa will get 3 minutes to solve as many arithmetic problems as possible, both on your own. These arithmetic problems include multiplying, dividing, subtracting or summing 1-digit, 2-digit, or 3-digit numbers. In the end, the total team performance will be measured by the total number of correctly answered questions by you and Lisa. Every correct answer is worth €0.25; the total revenue will be equally split between team members. Click on the button below, and the 3-minute task will start immediately. You are not allowed to use a calculator.
In the following part, you are invited to complete a simple task in a team together with Tim, another participant in this survey. During the task, you and Tim will get 3 minutes to solve as many arithmetic problems as possible, both on your own. These arithmetic problems include multiplying, dividing, subtracting or summing 1-digit, 2-digit, or 3-digit numbers. In the end, the total team performance will be measured by the total number of correctly answered questions by you and Tim. Every correct answer is worth €0.25. Your total team performance will be compared with the total team performance of another randomly chosen team. The team with the higher performance will receive the payoff, which will be equally split between team members. Click on the button below, and the 3-minute task will start immediately. You are not allowed to use a calculator.
In the following part, you are invited to complete a simple task in a team together with Lisa, another participant in this survey. During the task, you and Lisa will get 3 minutes to solve as many arithmetic problems as possible, both on your own. These arithmetic problems include multiplying, dividing, subtracting or summing 1-digit, 2-digit, or 3-digit numbers. In the end, the total team performance will be measured by the total number of correctly answered questions by you and Lisa. Every correct answer is worth €0.25. Your total team performance will be compared with the total team performance of another randomly chosen team. The team with the higher performance will receive the payoff, which will be equally split between team members. Click on the button below, and the 3-minute task will start immediately. You are not allowed to use a calculator. Q7: Task 80 x 30 = ... 1050 / 50 = … 73 - ... = 18 26 + ... = 62 135 - ... = 54 75 x 3 = ... 0.66 / ... = 0.6 3.45 + 7.55 = ... 92 / 4 = ... 56 x 50 = ... 210 / 7 = ... - 2.5 + 21.5 = ... 14 + 63 = ... 0.8 / 4 = ... 33 x 3 = ...
125 / 5 = ... ... / 18 = 0.5 0.3 x 0.21 = ... 12.5 x 4 = ... 196 - ... = 75 0.9 x 8 = ... 46 / ... = 2 386 - 91 = ... 4 x 45 = ... 35 x 9 = ... 23 + 165 = ... 0.4 x ... = 0.032 365 - 84 = ... 21 x 11 = ... 0.46 + ... = 2.37 135 / 5 = ... 576 x 2 = ... 5 / 0.25 = ... 45 x 20 = ... 0.4 x 0.8 = ... 178 - 86 = ... - 3.8 + ... = 15 0.2 x 0.36 = ... 35 / 0.5 = ... 76 + 85 = ... 231 / 7 = ... 336 + 64 = ... 18 + 47 = ... 99 / 3 = ... Thanks for participating in this survey. At the end of January, two teams will be randomly selected and paid out. If you want to have a chance to win the monetary payoff, please leave your name and e-mail address below. Name __________________________________________ E-mail address ________________________________
