Abstract
There are conflicting findings in team diversity research on whether it is better for an individual on a team to be similar to or different from the rest of the team. This lab study with undergraduates completing a critical thinking and decision-making task uses optimal distinctiveness theory to examine the idea that finding a balance between these two states for team member personality will result in positive perceptions of team process. Our results supported this such that participants had the most positive perceptions of team process when optimally distinct from the rest of the team in terms of personality.
Team diversity has been a heavily researched topic in the teamwork arena for many years. Previous studies have examined the influence of individual difference variables on important team outcomes such as performance, satisfaction, and viability (e.g., Barrick et al., 1998; Bell, 2007; Mohammed & Angell, 2003; Neuman & Wright, 1999; Peeters et al., 2006). Research in social categorization has looked more specifically at whether it is better for people in group contexts to be more similar or more different than other members of the group. One stream of research found that being more similar demographically 1 to members of a group was associated with positive job attitudes on the part of an employee (e.g., Tsui et al., 1992). Another stream of research found that groups with greater diversity in personality tend to experience greater cohesion and collaboration over time (Harrison et al., 1998, 2002).
These differing findings demonstrate that there are still unresolved questions regarding whether it is better for an individual team member to be similar to or different from the rest of the team. Previous research in this area has taken an either-or approach, where individuals are considered to be either similar or different from the rest of the group/team. However, individuals are rarely just one or the other; rather, it is likely that there is a continuum of diversity along which teams can fall. This is especially true when moving beyond a demographic conceptualization of team diversity (e.g., based on gender or race characteristics of the team members) to a dispositional conceptualization of team diversity that considers psychological differences such as the personality characteristics of the team members. Personality becomes more important for predicting work outcomes over time, at both the individual (Murphy, 1989) and at the team level (e.g., Harrison et al., 1998), and is therefore important to examine in the team diversity context where people are working together over time. Another literature gap relevant to this study is that most extant methods used to examine team personality (e.g., maximum score, minimum score, range of scores) do not include information from all members of the team, limiting the amount of information those methods provide about the team as a whole. To fully understand how individual group members are contributing to the functioning of the team, researchers need methods that incorporate information from every member of the team.
The purpose of the present study is to address these gaps in the literature by applying an extension of optimal distinctiveness theory (ODT; Brewer, 1991, 1993) in a laboratory setting via a novel statistical method for combining individuals’ personality variables within teams to predict important team process criteria. ODT has advantages compared to previous work because it posits that, rather than focusing on being either wholly similar to or wholly different from the rest of the group, there is an ideal balance between the two that will allow people in groups to fulfill both of the opposing needs of wanting to be unique and wanting to fit in. Extending this idea of being optimally distinct to a team context where we consider individual team member diversity in relation to the rest of the team allows us to reconcile the contradictory findings of the previous work on team diversity. Finally, we examine personality at the level of the individual’s reactions to the process of the team’s work, which is another topic with little research. Incorporating ODT into this literature allows for a more nuanced, theoretically-grounded examination of how personality relates to these team process outcomes. This is important because individual team members can react to the diversity of their team in different ways, depending on that team member’s personality traits and how they interact with the traits of the other members of the team. We examined five perceptions of team process that would reflect different aspects of the individuals’ experience of the team process: commitment, satisfaction, cooperation, justice perceptions, and team viability. In looking at how the individual’s reaction to the team process may depend on the individual’s experience with the diversity of the team, we bring together the research on team diversity, group differentiation, and dispositional diversity (i.e., personality). The current study operationalizes ODT in a novel way by measuring the Big Five personality traits (McCrae & Costa, 1997) of team members and examining how they relate to the overall group. Our method includes information from each member of the team and focuses on the individual personality level rather than the team personality level, as opposed to most of the team personality research (e.g., Barrick et al., 1998; Neuman et al., 1999; Prewett et al., 2009).
Specifically, our study builds on the extant research in the following ways: (1) It applies ODT to the examination of intra-team dynamics (rather than just postulating about its relevance) in order to move beyond the dichotomy of being either similar or different currently present in the social categorization literature, (2) it extends ODT to dispositional facets of diversity rather than demographic facets, (3) it examines additional outcomes that are indicative of the team process that the individual team members experienced, (4) it considers individual team member personality in relation to the rest of the team rather than personality at the team level, and (5) it uses a more innovative method to compare individual team members to the rest of the team in terms of personality that can be applied across studies. This method combines all of the traits into an overall “distinctiveness” score, rather than looking at individual traits. This combination of traits gives a better sense of how distinct a person is overall from other team members in terms of global personality as opposed to looking at distinctiveness at the individual trait level. What follows is a more thorough discussion of the relevant literature that inform the purpose and hypotheses of the study.
Relational Demography
The investigation of the relationship between individual-level traits, perceptions of team similarity (or dissimilarity), and how this relates to team outcomes has historically been investigated under the lens of relational demography (Tsui & O’Reilly, 1989). Relational demography generally refers to individual level differences between a team member and their team; these differences can be at the demographic level (e.g., gender, race), dispositional level (e.g., attitude, personality), or some combination of the two. Relational demography research relies on the social categorization approach which proposes that if people categorize each other into different social categories, they devalue each other which produces intergroup bias.
This theoretical approach found support in early studies focused on demographic differences. For example, Tsui et al. (1992) found that group homogeneity was positively related to several job attitudes like job satisfaction and psychological commitment. In a similar study examining supervisor-subordinate dyads, Tsui and O’Reilly (1989) found that dissimilarity in demographic characteristics resulted in lower perceived effectiveness for subordinates as well as increased role ambiguity for subordinates.
Subsequent research has produced results inconsistent with the social categorization approach. For instance, Harrison and colleagues conducted two longitudinal studies that simultaneously examined the impact of demographic and dispositional diversity on team performance (Harrison et al., 1998, 2002). Both studies found that dispositional diversity (e.g., personality) was associated with positive outcomes, such as increased cohesiveness and collaboration, as team members work together more frequently. These studies expanded the examination of relational demography by including dispositional diversity and a longitudinal research design to show that having similar group members is not always better in terms of individual and group outcomes.
As a further extension, empirical and theoretical work by Chattopadhyay and colleagues (see Chattopadhyay et al., 2004) added social identity theory (Tajfel & Turner, 1986) into work on relational demography as a potential moderator variable for the general finding that similar teams, in terms of demographic characteristics, are more harmonious. The combination of self-categorization theory and social identity theory postulates that individuals compare themselves to others in their team, make judgments on their level of similarity/dissimilarity to the rest of the team, and that their reaction to dissimilarity is influenced by the salience of the characteristics that are different. Using this framework, Chattopadhyay et al. (2008) found that sex dissimilarity was associated with increased task and emotional conflict for women while men actually reported lower levels of task conflict. The organization examined in the study was male-dominated and men were seen as having a higher status in the organization. Therefore, being different, at least for men, was a positive attribute because they were in a higher status group. The opposite effect was true for women.
The studies by Harrison et al. and Chattopadhyay et al. show that being similar does not always result in improved group processes. However, in general, the common wisdom continues to be that, all other things being equal, it is better to be similar. Indeed, in a review of four decades of research, Williams and O’Reilly (1998, p. 120) state that “Consistent with social categorization and similarity/attraction theories, the preponderance of empirical evidence suggests that diversity is most likely to impede group functioning.”
Implicit in all the relational demography research and theory is the notion that someone can be similar to or different from others, but not both (e.g., Chattopadhyay et al., 2004). However, ODT is a theory that suggests that group members can find a “sweet spot” where they are simultaneously similar to and different from other members of their group. Thus, whether it is better for someone to be similar or different from others is a false dichotomy according to ODT. Therefore, the previous relational demography research may be lacking in its ability to fully explore how individuals react to other members of the group. The current study answers the question of whether it is better to be similar or different from other group members by suggesting that, consistent with ODT, some group members can simultaneously be both and that these group members are the ones with the most positive perceptions of team process.
Optimal Distinctiveness Theory
Two of the most basic human social needs are the desire to be unique (Snyder & Fromkin, 1980) and the desire to belong (Baumeister & Leary, 1995). Per Brewer’s (1991, p. 477) original formulation of ODT, “social identity derives from a fundamental tension between human needs for validation and similarity to others (on the one hand) and a countervailing need for uniqueness and individualism (on the other).” Theoretically, when a person has found the perfect balance between uniqueness and belonging within a group context, they are “optimally distinct.” Brewer contends that “[s]ocial identity can be viewed as a compromise between assimilation and differentiation from others, where the need for deindividuation is satisfied within in-groups, while the need for distinctiveness is met through inter-group comparisons” (Brewer, 1991, p. 477).
Although Brewer initially conceptualized ODT to help understand intergroup dynamics among social groups, later researchers began to extend this conceptualization to intra-group comparisons. Brewer and her coauthors stated in a 2010 chapter that “Optimal distinctiveness theory was originally intended to apply exclusively to the collective social self. However, the basic underlying tension between inclusion and differentiation most likely plays itself out in other aspects of self-construal as well. . .At the individual level, the needs are expressed in the opposition between the desire for similarity on the one hand and the need for uniqueness on the other” (Leonardelli et al., 2010, pp. 102–103). Additionally, Meyer (2017) noted that ODT “assumes that team members seek an optimal balance of inclusion and distinctiveness within and between social groups” (p. 21). Thus, individuals within a team context who are too similar to other team members (in terms of either demographic or dispositional characteristics) are redundant and are not able to fulfill their need to be unique. Contrastingly, individuals who are too distinct may have difficulty fitting in and fulfilling their need to belong. Pickett et al. (2002) found some empirical support for the theory in discovering that individuals prefer to identify with groups in which both their needs for belongingness and uniqueness are met.
Several team theories mention the relevance of ODT in passing. One such theory is group faultlines, mentioned in the preceding paragraph and posited by Thatcher and Patel (2012). Faultlines are “hypothetical dividing lines that split a group or team into two or more subgroups based on one or more individual attributes” (p. 970). The more intense the faultline or division between these subgroups, the greater levels of conflict and lower levels of group performance and satisfaction exist within the team. Thatcher and Patel (2012) noted that ODT explains intra-group dynamics in relation to faultlines, in addition to inter-group dynamics. Therefore, an optimally-distinct team member could serve as the mediator or middle ground between the two (or more) subgroups created by the faultline in the team, thereby easing the conflict and facilitating team effectiveness and team member satisfaction.
Another team theory that mentions ODT is the bi-theoretical approach to team diversity (Meyer, 2017). The bi-theoretical approach, stemming from the 40-year review of team diversity literature quoted earlier by Williams and O’Reilly (1998), refers to the fact that there are two prevailing theories in the team diversity literature with contradictory predictions. The social categorization/similarity attraction perspective combines a few different theories, but ultimately predicts better working relationships between people that are more similar. This perspective is foundational to the relational demography researched mentioned previously. On the other hand, the information/decision making perspective proposes that individual differences among people within a team can prevent premature team consensus and increase the likelihood that someone in the team will find the solution to a problem, due to the fact that people are coming to the team with different perspectives and knowledge bases. Meyer discusses four streams of research that have sought to address this contradiction, but taking an intra-team approach to ODT provides an additional possibility. Someone who is optimally distinct within a team could benefit from both being similar and distinct to the other members of the team, in line with the predictions made by the two classes of theories making up the bi-theoretical approach. Despite noting the relevance of ODT to intra-group dynamics, neither the faultlines literature nor research on the bi-theoretical approach have empirically tested the effects of ODT within teams. This study will take the further step to empirically examine how optimal distinctiveness manifests within teams.
Organizational researchers have occasionally called for the application of ODT to help understand group and team functions at work. Specifically, in a narrative review of the inclusion and diversity literature, Shore et al. (2011) suggested ODT as a framework to determine the best way to integrate minority workers into work teams. They focused on demographic-level diversity (e.g., race, gender), but we propose that the theory can be extended to other, dispositional aspects of the individual, including personality. This follows the trend in the past couple of decades of team diversity research that also is moving from examining demographic characteristics to dispositional characteristics like personality (e.g., Mohammed & Angell, 2004).
Personality and Team Process
Personality influences how people complete tasks, interact with others, manage stress, and how open they are to others’ ideas, among other behaviors. All these behaviors will come into play when interacting with a group to complete a common goal. Team members who have too similar of a personality may not provide any added benefit of a unique perspective to the team, whereas team members who have too different of a personality may clash too much with the other team members and impede the team’s functioning.
The topic of individual-level personality and team process has not received a great deal of attention in the literature, and the attention it has received lacks a unified approach to conceptualizing personality in the team context. Indeed, a review of 31 studies on group personality and group effectiveness found a “surprisingly low number of relationship criteria” (Halfhill et al., 2005, p. 93) among these studies as opposed to task criteria. The relationship criteria that were present in the literature included variables such as group cooperation and group viability. One such study that examined varying levels of diversity in teams found that extraversion and team process (operationalized as a combination of communication, cooperation, and leadership) were moderately correlated (r = .30, Mohammed & Angell, 2004). Another found that extraversion and emotional stability were also related to team viability (Barrick et al., 1998). Finally, a more recent study found that openness to experience and emotional stability moderated the relationship between team conflict and team performance, such that conflict had a positive effect on team performance when the team was high in these traits, and a negative effect when the team was low in these traits (Bradley et al., 2013).
In this study, we examined five variables reflecting team members’ perceptions of the process of the team’s work. The first perception we examined was commitment, which represented the extent to which the person felt a part of the team. The second perception we examined was satisfaction, which reflected an affective response to the teamwork experience. The third perception we examined was cooperation, which reflected one’s feelings about how well they got along with their teammates. The fourth perception we examined was justice perceptions, which represented how fairly the individual felt the other team members treated him/her during the teamwork experience. Finally, we examined team viability, which reflected the participant’s interest in doing subsequent activities with his/her team.
These outcomes are important because they represent specific aspects of the teamwork process to which a person could have a positive or negative reaction, and thereby influence the likelihood of a person being willing to engage with those team members in the future. Although there is scant previous research on how personality might interact with the team process, some hypotheses can still be generated based on how ODT postulates group diversity impacts outcomes. Based on ODT, we expected that moderate overall distinctiveness scores, which will be discussed in greater detail in the method section, may produce the most positive perceptions of the team process because they are different enough to provide unique contributions to the team, yet similar enough to effectively work with the team. Such curvilinear effects on team process variables have been posited before for team member diversity in cognitive ability (e.g., satisfaction; Basadur & Head, 2001), but not for personality.
H1: Distinctiveness and commitment will be related in a quadratic fashion (i.e., inverted-U shape function) in which moderate levels of distinctiveness will produce the highest levels of commitment.
H2: Distinctiveness and satisfaction will be related in a quadratic fashion (i.e., inverted-U shape function) in which moderate levels of distinctiveness will produce the highest levels of satisfaction.
H3: Distinctiveness and cooperation will be related in a quadratic fashion (i.e., inverted-U shape function) in which moderate levels of distinctiveness will produce the highest levels of cooperation.
H4: Distinctiveness and justice perceptions will be related in a quadratic fashion (i.e., inverted-U shape function) in which moderate levels of distinctiveness will produce the highest levels of justice perceptions.
H5: Distinctiveness and team viability will be related in a quadratic fashion (i.e., inverted-U shape function) in which moderate levels of distinctiveness will produce the highest levels of viability.
Method
This study used an experimental design in an undergraduate educational context with teams performing a critical thinking and decision-making task.
Participants
Participants were 128 undergraduate students at a public university in the southeastern United States, recruited from a participant pool whose average age was 21.9 years (SD = 4.2). Seventy-three (57%) of the participants were female, 55 (43%) were male. Seventy-seven (60%) participants were White, 23 (18%) were Hispanic, 11 (9%) were Black, 8 (6%) were Asian, 8 (6%) identified as “other,” and 1 (1%) was Indian. Participants completed the study as part of 3- to 4-person teams resulting in a total of 33 teams. Participants were awarded course credit for participation and were also entered in a raffle for $200.
Measures
All the measures used in this study employed a 5-pont Likert-type scale (1 = strongly disagree, 5 = strongly agree). The perception of team process outcomes were measured at the individual level.
Big 5
We examine the distinctiveness of individual team members using the Big 5 personality traits: (1) openness to experience, (2) conscientiousness, (3) extraversion, (4) agreeableness, and (5) emotional stability. The Big 5 is the most popular conceptualization of personality and its factor structure holds across cultures and languages (McCrae & Costa, 1997). All personality constructs were measured using 10-item scales for each trait from the International Personality Item Pool (IPIP; Goldberg et al., 2006). Higher scores on each item indicate greater prevalence of the personality construct. Overall, the internal consistency for each scale ranged from α = .74 for openness to α = .91 for extraversion (see all values in Table 1). Example items from each scale are: “I make people feel at ease” (Agreeableness), “I am always prepared” (Conscientiousness), “I don’t mind being the center of attention” (Extraversion), “I have frequent mood swings” (Neuroticism), and “I have a vivid imagination” (Openness).
Means, Standard Deviations, Reliabilities, and Intercorrelations Between Study Variables.
Note. N = 128.
p < .05. **p < .01.
Team commitment
We adapted a 5-item scale from Meyer and Allen’s (1997) affective commitment scale to reflect commitment to the team. An example item is “I really felt as though the team’s problems were my own.” The internal consistency for this scale was α = .74.
Team cooperation
We created a 5-item scale for the purposes of the current study to measure team cooperation. An example item is “My team is cooperative.” The internal consistency for this scale was α = .70.
Team satisfaction
We created a 3-item scale for the purposes of this study to measure team satisfaction. An example item is “I enjoyed working with this team.” The internal consistency for this scale was α = .87.
Team viability
We used a 3-item scale of team viability from Sinclair (2003). An example item is “I feel that this group of individuals would work well together on another task.” The internal consistency for this scale was α = .93.
Justice perceptions
We adapted a 9-item scale from previous work by Colquitt (2001) to focus more on the specific team context, rather than a general job context. In the current study, we treated justice as a unidimensional construct. An example item is “I had influence over the decisions arrived at by my team.” The internal consistency for this scale was α = .87.
Dependent variable dimensionality
Since we developed a few of our dependent variable scales for the purposes of this study, we sought to examine the discriminant validity and factor structure of our five dependent measures. To accomplish this, we ran several confirmatory factor analyses examining a five-factor model of perceptions of team process versus a single factor model and several others that combined two or three measures together based on moderately high scale correlations (such as satisfaction and viability; see Table 1). The results of these analyses (see Table 2) indicated that the five-factor model provided the best fit to the data, χ2 = 671.71, df = 265, p < .001; CFI = 0.84; RMSEA = 0.11; SRMR = 0.10. The next best fitting model was a four-factor model that combined satisfaction and viability, but the five-factor model provided a better fit to the data, Δχ2 = 120.97, df = 4, p < .001. The one-factor model did not compare well to the five-factor model, Δχ2 = 400.35, df = 10, p < .001. Thus, the five-factor model appeared to be the best representation of the dependent variables.
DV CFA Fit Statistics.
Note. N = 128.
SRMR = standardized root-mean-square residual; TLI = Tucker-Lewis Index; RMSEA = root-mean-square error of approximation; CFI = Comparative Fit Index.
Parentheses in RMSEA column indicate the 90% CI for the statistic.
p < .001.
Considering some of the high correlations between several of the dependent variable scales mentioned previously, we also ran a discriminant validity analysis specified by Fornell and Larcker (1981). Evidence for discriminant validity is present for a pair of latent variable constructs when each construct’s average variance captured is greater than the square of the correlation between the two constructs/latent factors. In our case, all the dependent variable pairwise comparisons met this criterion. Based on these analyses, we kept the five dependent variables separate in the main study analyses.
Procedure
Following informed consent, researchers randomly assigned participants to teams, with a goal of having two male and two female participants in each team where possible. Four of the 33 teams only had three participants. Participants first completed the personality and demographic measures independently in small break out rooms. Participants were then assembled in one common room to complete a group task. We used the Desert Survival Simulation (Johnson & Johnson, 1994) as the team exercise. In this task, teams needed to reach a consensus on the relative importance of 15 items available to them in a disaster survival scenario. This task is meant to create a realistic simulation of intergroup processes—such as turn-taking, speaking up, negotiating task approaches, and reaching sufficient consensus—that are major aspects of teamwork across many different specific contexts. Tasks such as this are common in lab-based team research (e.g., see Bluedorn et al., 1999). After the task was complete, participants completed the dependent variable measures individually.
Calculating Optimal Distinctiveness
Examining optimal distinctiveness involves testing nonlinear relationships which involves use of a quadratic term within multiple regression. More specifically, optimal distinctiveness theory suggests a quadratic relationship (inverted-U) in predicting outcomes of interest. To compute the quadratic terms, for each member in a team, we calculated a distinctiveness variable. This involved calculating the individual’s average score on each dimension of the Big 5. Next, we calculated the team average on each dimension (e.g., extraversion) excluding that member of the team. Then we computed a difference score by subtracting the individual score from the average of the other team members’ scores, resulting in a set of five difference scores per individual, representing the Big 5 dimensions. Next, we squared each of these difference scores to avoid negative numbers. To test H1a-1e, we then summed up the squared scores (five in total representing each of the Big 5). Lastly, we took the square root of this sum to convert the scores back into their original metric. 2 This sum is the score we used to measure distinctiveness, and reflects the distinctiveness of each individual in terms of their set of personality dimensions as compared to the rest of the team as a whole. For the individual trait analyses, we took the square root of the squared scores for each individual trait to produce distinctiveness scores for each trait.
Low scores of distinctiveness indicate that the individual is highly similar to the average of the other members of his/her team in terms of the Big 5 (i.e., not distinct). High scores of distinctiveness indicate that the individual is highly dissimilar from the average of the other members of his/her team in terms of the Big 5. Individuals with moderate scores of distinctiveness are the closest to being “optimally distinct” because they have the best balance of group similarity and individualism. In the following section, we discuss how we used these distinctiveness scores for hypothesis testing.
Data Analysis
The distinctiveness scores discussed above were entered as a first step into the hierarchical regression equations. We then squared each distinctiveness score to examine the quadratic functions proposed in our hypotheses. Using hierarchical regression, we examined the incremental validity of this squared term over and above the linear term for predicting each of our five outcomes. When a significant effect occurred, we graphed our results to see if they were in the expected direction. Graphs were created using the ggplot2 package in R (Wickham, 2009).
Although participants were randomly assigned to teams and the study aimed to investigate with individual-level distinctiveness, it is possible there were some team-level effects on the dependent variables. This would violate the independence assumption in the analyses, as there would be variance accounted for by team membership in the dependent variables. To examine the independence of the teams, we ran several tests. The first was a one-way ANOVA on each of the outcome variables using team as the independent variable. None of the five ANOVA tests produced significant results, all ps > .05. The second was a similar series of one-way ANOVAs of the residuals resulting from the regression analyses used to test Hypothesis 1 on the five dependent variables. If there are no team-level effects, then the residuals should be random and the ANOVAs should show no significant results. This turned out to be the case, with none of the five ANOVA tests being significant, all ps > .05.
Finally, we conducted an ICC analysis. To examine the independence of the teams, we calculated intraclass correlations (ICCs) for each of the dependent variables, following the definition of ICC provided by Kenny and Kashy (2014). Their definition aligns with the Case 1 ICC(1,1) model (Shrout & Fleiss, 1979). Using this model, the ICC for each dependent variable was as follows: commitment, 0.00, p = .9996; cooperation, 0.05, p = .9978; satisfaction, 0.10, p = .9894; viability, 0.07, p = .9954; justice, 0.05, p = .9973. In this analysis, the ICCs represent the amount of variance in the dependent variables due to team membership, or differences between the teams (Kenny & Kashy, 2014), and the values here indicate little to no variance is due to differences between the teams. To evaluate the significance of the ICCs, we followed the procedure given by McGraw and Wong (1996), where the hypothesis test was whether the ICC was significantly different from 0. Thus, the failures to reject the null hypotheses at the .05 level indicate that there is insufficient evidence to conclude that the true values of the ICCs were greater than 0.
The results of these analyses indicate that focusing on the individual level effects in this study is appropriate, as the vast majority of variance in the dependent variables is due to individual, rather than team, variation. The results of the main study analyses are described below.
Results
Means, standard deviations, internal consistency reliabilities, and correlations of the primary variables used in the study can be seen in Table 1.
Overall Distinctiveness Effects
The results of H1 to H5 are displayed in Table 3 and Figure 1. H1, which proposed distinctiveness predicting commitment, was not supported, as the regression model including the quadratic term (i.e., optimal distinctiveness) was not significant for predicting commitment, although the relationship was in the expected direction, ΔR2 = .03, F(2,125) = 2.95, p = .056. H2, which proposed distinctiveness predicting satisfaction, was supported as the regression model including the quadratic term reached significance for predicting satisfaction and the relationship was in the expected direction, ΔR2 = .13, F(2,125) = 11.11, p < .001. H3, which proposed distinctiveness predicting cooperation, was also supported as the regression model including the quadratic term reached significance in predicting cooperation and the relationship was in the expected direction, ΔR2 = .11, F(2,125) = 10.18, p < .001. H4, which proposed distinctiveness predicting justice perceptions, was supported as the regression model including the quadratic term reached significance in predicting justice perceptions and the relationship was in the expected direction, ΔR2 = .11, F(2,125) = 8.94, p < .001. Lastly, H5, which proposed distinctiveness predicting team viability, was supported as the regression model including the quadratic term reached significance in predicting team viability and the relationship was in the expected direction, ΔR2 = .13, F(2,125) = 12.91, p < .001.
Regression Results for Overall Personality Distinctiveness.
Note. Places where p = .00 indicate places where exact p < .001.
p < .05. **p < .01.
Curvilinear effects of distinctiveness on dependent variables.
As a post hoc analysis, we examined optimal distinctiveness in terms of the original personality variables. To do this, we took the first derivative of the quadratic regression equations with respect to the distinctiveness scores for each outcome with a significant regression model containing a quadratic effect (i.e., all except commitment) and set the outcome equal to zero. Solving for the distinctiveness score provided the score where the slope of the curve was equal to zero, which will indicate the maximum height, or optimal level of distinctiveness, of each significant curvilinear effect. This resulted in the following distinctiveness scores for each significant outcome: satisfaction D = .72; cooperation D = .70; justice D =.74; viability D = .70. By design, these distinctiveness scores are in the units of the original 1 to 5 Likert response scale used for the personality items, and therefore indicate that the optimum difference between one person and the rest of the team in terms of combined personality scores is a little less than one on a 5-point scale. All the values are close to the distinctiveness mean of 0.71, SD = 0.25 and slightly above the median of 0.68. An important note is that these are point estimates of the true distinctiveness score, given that they come from the regression models.
Discussion
Overall, the hypotheses in our study were largely supported, and the significant results were consistent with the primary tenets of ODT. Specifically, these tenets state that those individuals in a team setting who are neither too similar nor too distinct, but struck an ideal balance between the two, will have more positive perceptions of their team process than those that do not. In the case of global personality, we found that this optimal distinctiveness level was related to all our dependent variables (i.e., commitment, satisfaction, cooperation, justice perceptions, team viability), as indicated by the significant standardized beta weights in Table 3, even if the overall regression model for commitment that included the optimal distinctiveness term was not significant.
Implications of Findings
The results of our study suggest important implications for both theory and practice. Theoretically, our results suggest that the debate between heterogeneity versus homogeneity in forming groups and teams, as noted in the research on the bi-theoretical approach and relational demography, may have focused on a false dichotomy. Our results indicate that it may not be about being only similar or only distinct but being in a sweet spot between the two extremes. Being similar to and different from other members of your team both have advantages, and those advantages can be realized at the same time by forming groups based on an optimal balance between the two. Ultimately, ODT is a parsimonious theory with one core tenet that is consistent with our results. Applying it to the team diversity literature helps to simplify the myriad of moderator effects and complex theoretical propositions found in the relational demography literature. As Albert Bandura noted, “The goal in theory building is to identify a small number of explanatory principles that can account for a wide range of phenomena (Bandura, 2005, p. 21).”
A further theoretical implication of our work is that more variables could be used to determine distinctiveness. For instance, individual difference variables such as cognitive ability and locus of control (see Ng et al., 2006) could also be used in calculating distinctiveness. Using more variables should theoretically increase the precision of the distinctiveness construct we developed in the present study. It is also somewhat surprising that personality distinctiveness had significant effects given the relatively short time the team spent working on the task and the constrained nature of the task, which speaks to the importance that team member personality has on how team members perceive the team process. Past longitudinal research by Harrison and colleagues (Harrison et al., 1998, 2002) would suggest that the effects we found could get stronger over time. Finally, ODT could be applied to examine faultlines within teams empirically, rather than just positing its relevance to the theory of faultlines.
Practically speaking, our results have uncovered a potentially powerful new method for forming groups. Put plainly, being optimally distinct in personality from your team members related to perceptions of team process such as being satisfied with the experience of being on the team, getting along with the team, feeling fairly treated by the team, and wanting to continue working with the team. Forming teams using the method in this study could dramatically improve the experience of being on a team and allow the teams to work together more effectively and efficiently and would not be difficult to implement. If a new organization needed to assemble several new work teams from among a large pool of candidate team members, a computer algorithm could be used to optimize the distinctiveness of each member of each team. The only inputs required for such an algorithm would be personality scores for each of the Big 5. For an existing organization, this algorithm could be used in conjunction with teamwork selection measures (e.g., Stevens & Campion, 1999) to form teams that are most likely to have a positive collective attitude about working together. However, we would suggest determining the generalizability of our results prior to implementing some of the ideas we have proposed. Additionally, the results of the current study imply that a person who may be “optimally distinct” in one group may not be in another. Hence, individual personality examined in isolation may not be the best predictor of perceptions of team process; rather the personality of all team members within a given team needs to be considered to best predict perceptions of team process. Most team-based research which has examined individual differences has not considered the personalities of other team members (see Prewett et al., 2009). It is important to note that Brewer originally posited ODT to balance intra-group inclusiveness with inter-group distinctiveness, and applying it in such a way in an organization would allow more employees within an organization to experience the benefits afforded by being optimally distinct. 3
Study Limitations and Future Research
As with any study, the current study has several limitations. First, our study carries with it the same limitations of all lab studies (e.g., concerns with external validity given the artificial nature of the lab setting). To alleviate this, we did select the team task with generalizability in mind, using an established task that includes common teamwork activities. One benefit of our laboratory setting is that teams were randomly assigned, ruling out any personality effects that might occur in naturally forming teams. A second limitation is that the use of a student sample may also suggest a questionable level of external validity. Given the student sample, our results are most likely to generalize among groups of students who are working together on a class project with a specific goal. Also, they are more likely to be applicable in teams where all members are of equal standing, rather than teams with a specified person in charge of the team. Third, while our discussion of belongingness and uniqueness was germane for understanding optimal distinctiveness theory, we did not measure these variables directly and cannot be certain that those participants deemed “optimally distinct” truly had both of those needs satisfied within their team. Fourth, our ANOVA test for effects of team membership was underpowered, so we cannot say for certain that there was no significant effect of team membership, although the test on residuals and ICC analysis help to support that. Lastly, teams develop and adapt over time. Our research, like most team research, was cross-sectional in nature and is therefore unable to make conclusions regarding relationships across time or any changes in the needs for belongingness and uniqueness that might take place over time.
The results of our study as well as its limitations suggest promising areas for new research. First, and foremost, the results of our study need to be replicated in applied settings to determine the robustness of the findings reported herein. There could be developmental differences in the level of distinctiveness that is optimal between college students and older employees. There could also be cultural differences, especially in collectivist cultures where fitting in with peers is more strongly emphasized. In these cultures, we would expect the optimal level of distinctiveness to fall more toward being similar than being different. Second, a more nuanced longitudinal study could be conducted examining our primary results. For instance, a longitudinal study could examine how team members who are not optimally distinct behave across time. Do these people try to find a way to adapt to or distance themselves from other members of the team? Third, future work could examine the influence of distinctiveness on individual performance within the group. Previous research has found that team member personality appears to relate more to behavioral criteria and team processes than with performance-related criteria (Prewett et al., 2009), which suggests that team member personality may influence performance-related criteria in an indirect manner via these behavioral criteria and team processes (LePine et al., 2011). For example, people at the optimal level of distinctiveness may participate more in the team, thereby leading to better performance outcomes. Similarly, future research could examine whether the relationship between distinctiveness and performance is mediated by perceptions of team process. This research has the potential to shed new light on the bivariate relationships that have been previously reported between these perceptions and group performance (e.g., see Neininger et al., 2010; Sinclair, 2003). Fourth, “distinctiveness” could be examined as a predictor of person-team fit in future studies (e.g., see DeRue & Morgeson, 2007; Kristof-Brown et al., 2005), especially in cases where teams have a formal leader. The optimal distinctiveness of the team leader could have stronger effects than that of other team members. Finally, a larger sample size could allow the testing of hypotheses related to differential effects of different profiles that result in equivalent levels of distinctiveness, for example, whether a distinctiveness score resulting from an individual being distinct on agreeableness and openness has the same effects as the same distinctiveness score resulting from an individual being distinct on conscientiousness and extraversion.
Conclusion
In conclusion, our study has contributed to the group and teamwork literature by providing the first application of an extension of ODT to the intra-team context, predicting outcomes that reflect the process by which the team completes its work based on dispositional diversity than has been previously examined. The results of our study underscore the potential value of this approach to team diversity for predicting relevant organizational criteria.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
