Diversity: âGoodâ or âBadâ for Teams and Organizations?
In the 1990s and early 2000s diversity management in developed economies continued to emerge, as organizations considered HR and management practices like diverse teams, diversity training, minority employee resource or affinity groups, mentoring, leadership development, and so on (e.g. Jayne & Dipboye, 2004).
Relatively early on, reviews of management research on diversity and teams acknowledged that diversity had simultaneous mixed effects and could act as a âdouble-edged swordâ on thoughts, emotions and behaviors affecting performance (Milliken & Martins, 1996). One persistent limitation seemed to be that such research was fairly intent on trying to establish whether diversity was positive or negative, or âgoodâ or âbadâ, in its âmain effectsâ on various outcomes like performance and well-being, despite such questions inevitably being subject to many caveats (van Knippenberg & Schippers, 2007). Papers worked at the problem by trying to establish âtypesâ or âdimensionsâ of diversity, difference or attribute that might operate differently. One classic dichotomous example of this was to distinguish surface diversity (overt, demographic, and biological) from deep diversity (flexible, attitudinal, and informational) (Harrison, Price, & Bell, 1998). The implication was dressed up in theoretical language but really often boiled down to a sense that organizational actors needed to get past visible categories, demographics, and stereotypes if they were to appreciate the more functional, deeper, meaningful, and useful human attributes underneath.
In tandem with this typing of different differences, quantitative management research, almost always on diverse teams, would try to model moderator and mediator variables to shed further light on diversityâs âgoodâ and/or âbadâ influences. Moderator variables could be included to show when, where, and for whom diversity would operate differently, and mediators could be included to show how/why (e.g. Mohammed & Angell, 2004; Pelled, Eisenhardt & Xin, 1999). Arguably one problem here is that as studies and findings accumulate, a mixed picture and long list of team member variables emerges â such as conflict, time, interdependence, team personality traits, debate and decision norms, information sharing, leadership, beliefs, and more. Associations may feel unreliable, inconsistent or difficult to reconcile to the reader of such research. The implications tend to be of the unwieldy form that diversity types are variously good, bad or neutral in their effects at different times, for different teams, and under different conditions.
Overall, and in review, the âbadâ and âgoodâ perspectives of this body of team or workgroup diversity research became referred to in theoretical terms as the âsocial categorizationâ perspective of disruptive ingroup/outgroup classifications, and the âinformation/decision-makingâ perspective of creative, valuable expert discussions, respectively (van Knippenberg & Schippers, 2007). In short, diversity was either driving a wedge through teams or leading to useful discussions -but with a tangle of processes in between and a potentially large range of different types of differences and perceptions of them among team members to start with.
In terms of theory, the main theories were mostly social psychological and located behind the negative predictions of the social categorization theory â similarity-attraction theory, social identity theory, self-categorization theory. Furthermore, researchers were potentially misusing the theories by deploying them in a simplistic way to hypothesize that diversity leads to âus-themâ feelings within a team and thus has negative effects as people feel more liking and attraction towards others they deem similar to themselves (van Knippenberg, de Dreu & Homan, 2004). In their categorization-elaboration model (CEM), van Knippenberg et al (2004) reemphasize this point and argue through their model that any attribute of team diversity can lead to positive or negative effects. Ultimately, they argued that the effects of diversity depend on the working context and if team members notice and process such differences in ways that positively help with completion of a task or negatively feed into intergroup biases. To some extent, this appeared to be simply rephrasing the overall problem and urging caution with this type of team diversity research.
Returning to the typing of diversityâs different differences (i.e. gender, age, expertise, tenure, personality etc.) â the typologies were inconsistent, proliferating, and not shown to be of consistent or clear explanatory value when findings were reviewed in several meta-analytic papers and reviews, which aggregated findings doing studies of compiled sets of studies. These category or classificatory labels for diversity were seemingly arbitrary, except they tried to make some rough distinctions between background characteristics, in terms of demography, skills, and attitudes/values. For example:
It is quite alarming in some ways that such a thriving and ongoing program of research has described its own sense of progress over decades as equivocal, inconclusive, and inconsistent on several, recurrent occasions in top-tier international reviews of world-class management research (Joshi, Liao, & Roh, 2011). Furthermore, Bell et al. (2011), based on their work, seemed to be suggesting that efforts to aggregate diversity into categories of different difference should be scrapped, and researchers should return to where they started â trying to gather data on individual demographic variables and their own particular black boxes.
Variants on this kind of research emerge, as authors try to circle the problem afresh, but often with the same or only slightly modified theories, designs, and statistical tools. For example, Shemla, Meyer, Greer, and Jehn (2016) argue that going forward what matters is perceptions of group diversity, and gaining more clarity and consistency on how people see the (dis)similarities and splits in groups and their effects. In a way, this again reflects a return to the initial âblack boxâ problem, which presumably demographers had hoped they would be able to skirt around by collecting and analyzing supposedly âobjectiveâ demographic workforce records and relating them to other perceptions and outcomes. There is also the challenging issue of considering diversity in terms of multiple levels of analysis, where the vast majority of this kind of research has focused on the team level (e.g. Nielsen, 2010). One exception to this is the work of Orlando Richard, who has linked diversity to strategies and industry environments at the firm or organization level of analysis (e.g. Richard, Murthi, & Ismail, 2007). However, arguably this work reproduces many of the issues of team-level research but at a higher level of analysis.
In sum, despite an impressive body of work and hypothesis testing in management research over the last few decades, findings on team and organizational diversity havenât stood up well to reviews and the hopes of accumulating reliable and meaningful knowledge about its effects. At best, research seems to circle around addressing ideas of context and conceptualization, repeating that diversity can mean different things to different people in different settings, which future research should keep trying to capture. Moderators and mediators in quantitative analyses give some further insights, but only in a piecemeal way, as they offer up lots of âifs and butsâ to diversityâs effects. We end up only being certain that diversity â of race and gender in teams, for example â can be sometimes good, sometimes bad, depending on a myriad of other factors around measurement, perceptions, context, task, team norms, strategy, time, and industry.
This research leads to a certain kind of progress and understanding but is ultimately only one kind. Other chapters of this book will argue that critical approaches to studying diversity offer various alternatives that can overcome some of the issues around appreciating the context and meaning of diversity in organizations. However, given that a pluralism of diversity research in management is likely to continue, much of it building on the studies mentioned previously but also taking other paths...