British Journal of Social Psychology (2003), 42, 221–223 © 2003 The British Psychological Society www.bps.org.uk Response Social dominance theory comes of age, and so must change: A reply to Sidanius & Pratto, and Turner & Reynolds Marc Stewart Wilson* and James H. Liu Victoria University of Wellington, New Zealand We find ourselves in the curious position of having fired a shot across the bows of social dominance theory (SDT) but enlisting in its defence in a full-fledged war between SDT and social identity theory (SIT). The combination of Schmitt, Branscombe, and Kappan’s intricate experiments, Turner and Reynolds’ ferocious rhetoric, and our own evidence for moderation effects appear to be a devastating attack on SDT. However, it could also be read as a milestone in the theory’s development—the long-term reactions of Sidanius and Pratto to this forum will determine how posterity will view these papers and the theory. First, in defence of our own work, Sidanius and Pratto (2003) make the claim that, based on their own reinterpretation of our measures, the findings are entirely consist- ent with SDT. Specifically, their argument is that rather than presenting a comparison of males and females on the same construct, we are comparing male-dis-identified males with female-identified females, and male-identified males with female-dis- identified females. While that might be an interpretation, we feel that it represents a contrived shifting of the goal posts from the position Sidanius and Pratto have previously adopted when evaluating potential covariates (Sidanius & Pratto, 1999). Specifically, many variables could be interpreted in this fashion. For example, Sidanius and Pratto (1999, p. 278) present the same type of analysis with political conservatism as the potential moderator. In this case a successful moderation result could be inter- preted in the same way—that is to say, republican-dis-identified males being compared with liberal-identified females, and so on. Yet their analysis, identical to our own, is presented as a legitimate test of the invariance hypothesis. Secondly, as relates to the implications that these papers present for SDT—Schmitt et al.’s (2003) work is (on the surface) most damning, but on reflection is reminiscent of attacks in other unwinnable wars; for example, that between state and trait theorists (e.g. Mischel, 1968) about the utility of personality as a predictive construct. Schmitt et al. represent a long-standing tradition in experimental social psychology of treating individual differences as error variance but, while it makes for fun reading, such a position is untenable in the long term. In personality psychology, the emergence of the Big Five (e.g. McCrae & Costa, 1987) has temporarily stilled debate about whether *Requests for reprints should be addressed to Dr Marc Stewart Wilson, School of Psychology, Victorian University of Wellington, PO Box 600, Wellington, New Zealand. 221