Research examining the effects of diversity in workgroups has been inconsistent, and even conflicting, in its findings over the last 50 years, leading some researchers to refer to diversity as a double-edged sword (Milliken & Mrutins, 1996, p. 402). As a consequence, models of diversity have evolved to incorporate moderators and mediators in the relationship between diversity and outcome variables. In this program of research, a series of three field studies utilising both qualitative and quantitative methods were conducted to test and extend Pelled's (1996) model of workgroup diversity, in the Australian Defence Force. Taking a social identity approach, Study 1 aimed to identify the specific conditions (i.e., cognitive group representations) under which employment status diversity (reserve, or part time, military personnel v. full time military personnel) would lead to workgroup conflict, and to test the proposed mediating role of conflict in the relationship between diversity and organisational outcome variables. Qualitative interviews were conducted with 16 personnel (8 reserves, 8 full time) from a maintenance unit of the Royal Australian Navy, and 160 surveys (n = 144 full time, n = 16 reserves) were collected from five unit locations around Australia. In Study 1, sampling full time naval personnel only, both a 'one group' (i.e., dual identity) and individuals (i.e., decategorised) group representation were found to moderate the diversity, conflict relationship, but only in workgroups with a high proportion of outgroup (reserve) members. In diverse workgroups, a significant negative relationship between the 'one group' representation and workgroup conflict was observed, while for the 'individuals' item, this pattern was reversed. Conflict was found to mediate the relationship between diversity and workgroup performance only. In Study 2, 25 (9 full time, 16 reserve) qualitative interviews were conducted with personnel from a reserve-integrated combat battalion in Brisbane, Australia. Surveys were also collected from 90 personnel (17 full time, 73 reserve) that had recently returned from a four-month training deployment to Malaysia. Examining the reserve data, only the individuals group representation moderated the relationship between diversity and conflict. For participants in workgroups with a high proportion of outgroup members, higher scores on the individuals representation measure were associated with higher levels of conflict, while for those in workgroups with a low proportion of outgroup members, a significant negative relationship was observed. Conflict was a marginally significant mediator of the Diversity x 'Individuals' - job satisfaction relationship only, with conflict again a negative mediator as predicted. Extending Pelled's (1996) model, intergroup anxiety was also included as a mediator. Anxiety was found to fully mediate the relationship between a one group representation and both ingroup bias (status relevant) and outgroup evaluation. Specifically, bias was reduced and outgroup evaluations improved through a decrease in intergroup anxiety. In Study 3 a total of 200 military personnel (n = 35 full time, n = 165 reserve) from units of the Australian Army (n = 34 full time, n = 81 reserve) and Royal Australian Air Force (n = 1 full time, n = 84 reserve) located in South East Queensland, Australia, returned surveys. Analyses of reserve personnel revealed only the one group representation moderated the relationship between diversity and conflict, when proportion of outgroup was high. Higher scores on the one group measure were associated with lower levels of conflict, for participants in more diverse workgroups. In this study quality of contact was also found to moderate this relationship in a similar manner to the one group measure. While Pelled's (1996) model was found to be limited in its explanatory power in this program of research, a social identity approach was found to be useful in extending its utility and identifying both the conditions under which diversity and conflict are related, and the processes that intervene between diversity and organisational and intergroup outcome variables.