Team climate mediates the effect of diversity on environmental science team satisfaction and data sharing

Scientific research—especially high-impact research—is increasingly being performed in teams that are interdisciplinary and demographically diverse. Nevertheless, very little research has investigated how the climate on these diverse science teams affects data sharing or the experiences of their members. To address these gaps, we conducted a quantitative study of 266 scientists from 105 NSF-funded interdisciplinary environmental science teams. We examined how team climate mediates the associations between team diversity and three outcomes: satisfaction with the team, satisfaction with authorship practices, and perceptions of the frequency of data sharing. Using path analyses, we found that individuals from underrepresented groups perceived team climate more negatively, which was associated with lower satisfaction with the team and more negative perceptions of authorship practices and data sharing on the team. However, individuals on teams with more demographic diversity reported a more positive climate than those on teams with less demographic diversity. These results highlight the importance of team climate, the value of diverse teams for team climate, and barriers to the full inclusion and support of individuals from underrepresented groups in interdisciplinary science teams.


Introduction
Several trends are transforming contemporary scientific practices. In most disciplines, scientific research is increasingly being conducted in teams [1][2][3], and these teams are becoming increasingly interdisciplinary in order to tackle grand challenges that require multiple disciplinary perspectives [4,5]. The scientific community is also striving to become more demographically diverse and to promote the advancement of groups that have been underrepresented in the sciences [6,7].
However, creating successful teams that are demographically and scientifically diverse is not a simple matter of recruiting more individuals from underrepresented groups and combining team members from a variety of disciplinary backgrounds. Diverse teams can struggle with allocation of credit, differences in perspectives, and unequal power dynamics. For example, women and those from "soft sciences" (e.g., sociology) can be less credited or valued than men or those from "hard sciences" (e.g., physics; [8][9][10]). Philosophical, methodological, and conceptual differences that result from disciplinary diversity can complicate team collaboration [4,5,11]. Finally, power dynamics can be difficult in demographically diverse teams, with individuals from some groups feeling unable to influence team practices and decisions [12]. For the sciences to effectively transition to a more diverse team-based enterprise, contemporary science teams must address these challenges. We propose that team climate is a critical factor for addressing these challenges and promoting the success of diverse science teams. Team climate is the perceived set of norms, attitudes, and expectations on a team [13]. Research indicates that climate is related to individual job attitudes such as organizational commitment and turnover intentions, as well as to job performance [14][15][16]. Climate is also related to positive team performance [17,18]. However, surprisingly few studies have investigated the relationship between climate and diversity specifically within science teams [5]. In one of the few studies of diversity and team climate in the science team context, Li et al. [19] found that cultural diversity was related to greater creativity on engineering teams through the mediating role of information sharing, but that result was only for teams with a climate of inclusion (i.e., equitable employment practices, integration of differences, and collective decision making). To better support science teams with increased demographic and disciplinary diversity, more studies are needed to determine how individual and team diversity are related to climate, perceptions of team functioning, and team satisfaction.
To fill this gap, our study tests a conceptual framework that describes how diversity, climate, and team outcomes are related to each other (Fig 1) in a sample of 266 participants from 105 NSF-funded interdisciplinary environmental science teams. We studied the associations between two forms of individual and team diversity (demographic and scientific; see Fig 2) and team members' satisfaction with their teams, their satisfaction with authorship practices, and perceptions of the frequency of team data sharing. These outcomes are important because diverse, interdisciplinary science teams are less likely to function successfully and to retain members of underrepresented groups if team members are not satisfied and if they do not perceive effective and fair implementation of important team practices such as authorship and data sharing. By measuring diversity in terms of demographic and scientific composites that simultaneously account for multiple underrepresented identities, our study builds on previous work that has focused on single dimensions of diversity, such as gender or race [5], which does not accurately reflect people's lived experience.
We included team climate as a mediator of the associations between diversity and outcomes, and we measured it by assessing individuals' perceptions of procedural justice, collaboration, and inclusion on their teams (see Fig 3). These dimensions of climate are likely to help teams address challenges associated with allocation of credit, differences in perspectives, and unequal power dynamics [20][21][22].
We hypothesized (Fig 1) that: (H 1 ) individuals with more underrepresented demographic and scientific characteristics compared to their counterparts will be less satisfied with their teams, less satisfied with authorship practices specifically, and perceive data sharing to occur less often, and these associations will be mediated by more negative perceptions of team climate; and (H 2 ) regardless of one's own demographic and scientific characteristics, individuals who are on teams with more demographic and scientific diversity will be more satisfied with their teams, more satisfied with authorship practices specifically, and perceive data sharing to occur more often compared to individuals on teams that have less demographic and scientific diversity, and these associations will be mediated by more positive perceptions of team climate. To test these hypotheses, we conducted path analyses to understand how team satisfaction and team practices were affected by diversity composites, and how team climate mediated these associations.

Participants and survey procedures
Potential participants and teams were identified using the National Science Foundation (NSF) database of awards for three interdisciplinary environmental science funding programs. (We do not provide program names to reduce risk of participants' identification.) The NSF database reports contact information for project Principal Investigators (PIs) and Co-Principal Investigators (Co-PIs). To recruit participants who held other roles on these projects (e.g., graduate student research assistants, post-docs, technicians), we emailed the PIs and requested that they provide contact information for all team members. During the summer of 2017, we invited a total of 1,727 individuals from 229 interdisciplinary research teams via email to participate in an online survey using the Qualtrics survey platform. To increase survey responses, participants had the opportunity to win one of five $100 Amazon gift cards, and we sent two follow-up reminder emails to non-respondents. The survey contained our NSF (NSF-14546) and IRB (HUM00128956) identification numbers, contact information for our project Principal Investigator and Research Coordinator, and instructions that participants could skip any questions that they preferred not to answer.
Our final sample contained 266 participants from 105 NSF-funded research teams (response rate = 15.4% of participants, 45.9% of teams). Participants had an average age of 46.56 (SD = 13.15) years. See Table 1 for participants' sex, race/ethnicity, sexual orientation and gender identity, nationality, academic disciplines, career status, and duration of involvement with their NSF research projects; see Table 2 for the demographic and scientific composition of participants' teams. Correlations among all study variables are presented in Table 3. All data included in this study, except for demographic information that could be used to identify participants, are available in a public archive [23].

Characterizing diversity
We used the survey responses to calculate four composite predictor variables characterizing diversity in our models. We computed measures of diversity at both the individual and the team level, as well as to characterize both demographic and scientific diversity (See S1 Table for additional details on the diversity composite variables).
Individual demographic diversity was a sum of the number of dimensions (ranging from 0 to 5) along which participants contributed to their team's demographic diversity in terms of sex, race, sexual orientation, gender identity, and nationality (see Fig 2 for the specific groups coded as contributing diversity in each category). Groups that are underrepresented in the academy relative to their prevalence in the United States were coded as contributing diversity to their teams. Note that for race, we did not count being Asian as contributing to diversity because they are not underrepresented in the academy [24,25], and racial stereotypes portray Asians as intelligent, educated, and hard-working-characteristics that are not as readily attributed to other racial minority groups [26]. As a result, Asians in the academy may have qualitatively different experiences from those of racial minorities who are underrepresented in the academy [27].
Team demographic diversity represents participants' teams' demographic diversity in terms of sex, race, sexual orientation and gender identification, and nationality. Using participant reports of the gender and race of team members, and total numbers of team members, we calculated the proportion of team members providing gender and racial diversity (see Fig 2 for the specific groups coded as contributing diversity in each category). As with individual demographic diversity, Asians were not included when calculating team racial diversity. For sexual orientation and nationality, participants indicated, to the best of their knowledge, the makeup of their team on a 5-point Likert-type scale; sexual orientation ranged from 1 (all straight/heterosexual) to 5 (all lesbian, gay, bisexual, and queer-identified), and nationality ranged from 1 (all from the U.S.) to 5 (all not from the U.S.). "Don't know" and "Prefer not to answer" response options were available for these questions, and these responses were excluded from the measure. Because these variables had different scales, participants' scores on each variable were standardized and then averaged, so that higher scores indicated more team demographic diversity on these four dimensions.
Individual scientific diversity was computed as the sum of the number of dimensions (ranging from 0 to 3) along which participants contributed to their team's scientific diversity in terms of academic discipline, career status, and how long they had been involved with their  team's project. Groups with less status on an environmental science team were coded as contributing diversity to their teams (see Fig 2 for the specific groups coded as contributing diversity in each category). Team scientific diversity represented individuals' reports of the scientific diversity on their teams in terms of career status, discipline, and previous collaboration. Participants reported the number of team members in various career positions and in different disciplines; using their reports of the total number of team members, we calculated the proportion of team members providing diversity in each category. For previous collaboration, participants indicated the proportion of team members who had previously collaborated on a scale from 1 = 0-20%, 2 = 21-40%, 3 = 41-60%, 4 = 61-80%, 5 = 81-100% (see Fig 2 for the specific groups coded as contributing diversity in each category). These three variables were standardized and averaged so that higher scores indicated more team scientific diversity.

Characterizing team climate
We assessed participants' perceived climate on their team via measures of procedural justice, team collaboration, and how much the team values inclusion. For all three of these measures, we adapted questions from published scales (see S2 Table for all scale items). For each scale, we computed a mean score such that higher values reflect a more positive climate.
To measure procedural justice, we adapted four items from the Procedural Justice subscale of Colquitt's [21] Organizational Justice Scale. Participants responded to items (e.g., "Have you had the ability to influence your NSF team's policies and/or practices related to conducting and publishing research?"; M = 4.13, SD = .82; Cronbach's alpha = .84) on a scale from 1 (not at all) to 5 (almost always).
To measure team collaboration, we adapted six items from Carson et al.'s [20] assessment of Internal Team Environment for Shared Leadership. Participants responded to items (e.g., "The members of the team spend time discussing our team's purpose, goals, and expectations for the project"; M = 4.02, SD = .67; Cronbach's alpha = .87) on a scale from 1 (strongly disagree) to 5 (strongly agree).
To measure team value of inclusion, we adapted six items from Pugh et al.'s [22] Diversity Climate measure. Participants responded to items (e.g., "Our team makes it easy for people from diverse backgrounds to fit in and be accepted"; M = 4.17, SD = .72; Cronbach's alpha = .88) on a scale from 1 (strongly disagree) to 5 (strongly agree).
In order to measure team climate, we combined these three mean scale scores such that higher scores indicated more positive climate on the team. Cronbach's alpha for the composite was 0.93 for our sample. The correlations between all subscales were significant at p < .001 (S1 Table).

Characterizing team outcomes
We used the survey responses to calculate three response variables for our models that characterize individuals' experiences on their interdisciplinary science teams. We measured individuals' satisfaction with their team, their satisfaction with its authorship practices, and their perceptions of its data sharing practices. We based the survey questions on prior qualitative interviews of the population [28]. To measure the extent to which participants were satisfied with their experiences on their interdisciplinary science teams, participants responded to the question "Overall, how satisfied are you with your experiences on your interdisciplinary NSFfunded science team?" on a scale from 1 (not at all satisfied) to 5 (very satisfied).
To measure participants' satisfaction with their interdisciplinary science teams' authorship practices, we asked them to answer three questions about authorship credit: "To what extent do you think that you personally received appropriate credit (in terms of being included as an author or not) on the papers published by your team?" 1 (inappropriate) to 5 (appropriate); "In your personal opinion, to what extent do you think your interdisciplinary NSF-funded science team is typically fair in deciding who to include as authors on papers?" 1 (not at all fair) to 5 (extremely fair); and "How often do you think your interdisciplinary NSF-funded science team has excluded people from being authors even though they contributed sufficiently to the paper?" 1 (never) to 5 (always). The exclusion item was reverse-scored and then the three questions were averaged such that higher scores indicate more satisfaction with team authorship practices. The correlations between all items were significant with p < .001 and Cronbach's alpha was 0.75 for our sample.
To measure the data sharing practices, we asked participants to indicate how often their team shared data within sub-teams (i.e., a smaller group of team members working on a specific task within the larger research team) and with the entire team using a 5-point Likert-type scale of 1 (never) to 5 (always). The correlation between data sharing within sub-teams and with the entire team was significant r = .51, p < .001. Scores for the two items were averaged and higher scores indicate more data sharing within the team.

Results
To test our hypotheses, we used the conceptual model described in Fig 1 and conducted two path analyses, one with each set of diversity composites (i.e., individual and team demographic diversity or individual and team scientific diversity) as the predictor variables. In all analyses, the mediator was team climate and the outcomes were satisfaction with the team, satisfaction with authorship practices, and data sharing practices. Path analysis, an extension of multiple regression, tests the strength of relationships among variables. Because it can test a hypothesized model with multiple independent and dependent variables, and mediating (i.e., indirect) effects, it is appropriate for our study [29]. All analysis used MPLUS Version 8 [30].

Demographic diversity and team climate
The first model (Fig 4A) examined the effects of demographic diversity on outcomes at the individual and team levels. Testing H 1 and H 2 for demographic diversity, we found that diversity composites were not directly related to our outcomes. However, individual demographic diversity was associated with more negative perceptions of team climate, and team demographic diversity was associated with more positive perceptions of team climate. Further, tests of indirect effects indicated that team climate perceptions mediated the relationship between individual demographic diversity and all three outcomes. Team climate also mediated the relationship between team demographic diversity and all three outcomes, but with opposite effects (See Table 4). Specifically, participants with more underrepresented demographic characteristics (e.g., women who are Black, gay men not born in the US) perceived their team climate to be more negative, which was associated with lower satisfaction with the team and more negative perceptions of authorship and data sharing on their teams. In contrast, participants on more demographically diverse teams perceived team climate to be more positive, which was associated with their greater satisfaction with the team and more positive perceptions of authorship and data sharing on their teams.

Scientific diversity and team climate
The second model (Fig 4B) examined the role of scientific diversity at the individual and team levels. Testing H 1 and H 2 for scientific diversity, we found that individual scientific diversity was directly related to satisfaction with authorship practices. As with individual demographic diversity, individuals who contributed more scientific diversity to their teams perceived their team climate more negatively. However, unlike team demographic diversity, team scientific diversity was unrelated to team climate. Tests of indirect effects indicated that team climate mediated the relationship between individual scientific diversity and all three outcomes. However, team climate did not mediate the relationship between team scientific diversity and any outcome. Thus, individuals with more underrepresented or low status scientific characteristics perceived their team climate more negatively, which was associated with their lower satisfaction with the team and more negative authorship and data sharing perceptions. In Team climate mediates the effect of diversity on environmental science team satisfaction and data sharing contrast, team scientific diversity was not related to team climate or any of the outcomes examined.

Discussion
Our findings indicate that positive perceptions of team climate are associated with satisfaction with teams, as well as perceptions that authorship practices are fair and that data are shared openly within teams. However, individuals with more dimensions of demographic or scientific diversity (e.g., women, LGBTQ team members, early-career scientists) perceived team climate to be more negative than their more represented counterparts, and as a result they reported less satisfaction with their teams, team authorship practices, and the frequency of team data sharing.
These findings support our hypotheses and suggest that efforts to maximize the benefits and minimize the challenges of diverse science teams should take into account the mediating effects of team climate. In accordance with our first hypothesis, we found that those who contributed more demographic or scientific diversity tended to perceive climate less positively than those who did not contribute as much diversity. As predicted by our second hypothesis, one of the factors related to positive climate perceptions was team demographic diversity, although team scientific diversity did not have this effect. Thus, although our results support ongoing efforts within the scientific community to incorporate individuals who can contribute diversity to scientific teams, we add the important caveat that it is critical to provide these individuals with adequate support and recognition. Moreover, in order to promote positive team outcomes, greater attention needs to be directed at understanding the range of factors that influence the climate of science teams.
As predicted by H 1 and H 2 , perceptions of team climate on diverse science teams may drive outcomes such as satisfaction with teams, satisfaction with authorship practices, and frequency of data sharing. It makes sense that these outcomes can be improved by addressing the three dimensions of climate examined in this study: procedural justice, collaboration, and inclusion. Having clear, openly-discussed, and collaboratively developed team policies and practices is likely to promote data sharing and encourage fair credit allocation related to authorship [12,28,31]. In addition, fair and transparent policies and procedures are likely to alleviate power imbalances that can diminish satisfaction with teams [12]. The importance of promoting positive climate also accords with the finding that diversity can have varied effects on team outcomes, and what matters is whether organizations support diversity by recognizing the contributions of all individuals through fair processes and rewards [32]. We found that perceptions of the climate on teams with greater demographic diversity were more positive than on less demographically diverse teams. These positive effects of demographic diversity are aligned with previous research indicating that diversity can have a number of beneficial effects on team outcomes [5,33]. Demographic diversity might improve team climate because team members from traditionally underrepresented groups may be particularly likely to identify concerns about power dynamics and unfair or exclusive practices on these teams [34,35]. By doing so, they could help to prevent and alleviate policies and practices that damage team climate, but they may feel frustrated or burdened with the need to be the individuals performing these extra duties.
Although demographic diversity is generally beneficial for teams, the outcomes are less positive for the individuals who contribute diversity. The less positive outcomes for these individuals may be the result of "token effects," which occur when group members experience stresses such as performance pressure and social isolation because they have characteristics that are unique within their groups [27,36,37]. The somewhat counter-intuitive difference between group-level and individual-level results for teams with greater demographic diversity might be occurring because participants who contributed diversity to the teams we studied made up a low proportion of their teams. Therefore, their negative perceptions did not overwhelm the overall positive perceptions associated with diverse teams.
Although some scholars have theorized that token effects could be addressed by increasing the proportion of underrepresented individuals on teams [36], other research suggests that the problems experienced by token team members are related not just to low numbers but also to low status [12,38]. This accords with our findings, insofar as those who are from scientifically or demographically underrepresented groups (e.g., having early career status; being on the team for less than half the project duration; identifying sex as female; identifying race as Black, Latinx, or American Indian) are also likely to have comparatively low status on scientific teams. Thus, in addition to recruiting more individuals from underrepresented groups to science, it is important to take additional steps at the team and institutional levels to support and value the contributions of all team members [27,39,40]. Over the long term, changes to institutional cultures (i.e., the basic underlying assumptions and espoused values) in which science teams operate could help improve climate and facilitate the development of more inclusive practices [41].
Our research moves beyond previous studies in three important ways. First, very little previous work investigates the role of climate in science teams, and none of this research investigates the effects of climate on team practices like authorship and data sharing. Second, whereas previous studies have focused on single dimensions of diversity (primarily gender or race), our study answers recent calls to examine multiple dimensions of both demographic and scientific diversity [42]. Third, we examined diversity in the composition of teams at both the individual and team level (albeit aggregated from individual reports). Our findings indicate that investigations into divergences between individual and team level diversity are very important in order to promote the interests of underrepresented groups in science.
Our findings also suggest that we should reframe the current dialogue surrounding science teams and diversity. This conversation should focus less on whether diverse teams are good for team outcomes (which appears to depend on the outcomes and the dimensions of diversity) and more on the factors that contribute to positive outcomes both for diverse teams and for individual team members. We found that team climate perceptions are one of the important factors related to positive or negative outcomes on science teams. Therefore, science teams will benefit from additional research on steps to improve team climate for all members on science teams, especially those who are underrepresented or marginalized.
Supporting information S1