Assessment of sexual dimorphism: a critical discussion in a (paleo-) anthropological context

Viewed as a biological concept, sexual dimorphism does not appear to motivate debate. However, with the aim of measuring such dimorphism, several indexes have been proposed, and not all of them have been regarded as equally reliable by the biological community. The main divergence between the indexes is that their definition is based on appraising only partial features (e.g., the mean parameter) of the set of measurements corresponding to each sex. Provided that sexual dimorphism can be satisfactorily analyzed when random variables and their distribution functions are involved, it is also likely that the conjecture that the two sexes making up a population are independent enables such indexes to be clearly distinguished. We examined and compared the following measures of sexual dimorphism: the quotient of sample means, the sample range, the sample coefficient of variation, the overlapping area between two independent normal distributions, and the overlapping area between the functions making up a mixture of two normal distributions. We especially consider their inferential structures.