Key points are not available for this paper at this time.
The extensive research examining relations between group member dissimilarity and outcome measures has yielded inconsistent results. In the present research, the authors used computer simulations to examine the impact that a methodological feature of such research, participant nonresponse, can have on dissimilarity-outcome relations. Results suggest that using only survey responders to calculate dissimilarity typically results in underestimation of true dissimilarity effects and that these effects can occur even when response rates are high.
Allen et al. (Mon,) studied this question.