How to measure intersexual dominance?

Abstract

Intersexual dominance (dominance between the sexes) is often assumed to be binary with species categorized as either male- or female-dominant. Yet in many species, the degree of intersexual dominance falls somewhere in the middle of these two extremes. There are several measures of intersexual dominance, but in empirical studies, it is not possible to evaluate which is best because the real degree of intersexual dominance is unknown. This evaluation is possible, however, in the agent-based model, DomWorld, because individuals have internal dominance values that drive their agonistic behavior. In the present study, we defined the accuracy of measures of intersexual dominance in DomWorld by the strength of the correlation between the degree of intersexual dominance based on A) their internal dominance values and B) observations of their competitive interactions (similar to observations in empirical studies). We examined the four measures that have been most commonly used in the literature: the proportion of intersexual conflicts won and initiated, the Female Dominance Index, and the proportion of female-dominant dyads. The Female Dominance Index was highly accurate, possibly because it was based on the outcomes of intra- and intersexual conflicts, both of which influence an individual's dominance. The proportion of intersexual conflicts initiated was similar in its accuracy to the Female Dominance Index and it was the only measure to be unaffected by missing data. Measures were more accurate when groups were smaller, or the intensity of aggression was higher, but their accuracy did not depend on the degree of sexual dimorphism. To best represent dominance relations between the sexes, we recommend reporting both the Female Dominance Index and the proportion of intersexual conflicts initiated.