One of the long-standing issues in HLA research is whether there is se
gregation distortion in the HLA complex in human populations. In this
paper we study some simple statistical models aimed at detecting segre
gation distortion. We present a statistic to test the Mendelian null h
ypothesis of equal transmission probabilities. To assess the possible
contribution of multiple alleles to segregation distortion, we employ
a specific log-linear model for transmission probabilities equivalent
to the Bradley-Terry model in the literature of paired comparisons. We
also provide a simple method for detecting a single allele effect, if
present.