An ex ante adoption model of bovine somatotropin (BST) is estimated with survey data of California milk producers. Theoretical justification is developed for incorporation of socioeconomic explanatory variables in a technology-adoption model. The advantages of a multinomial over a binomial ex ante model also are presented. The multinomial logit model is used to predict BST adoption, to test hypotheses on characteristics associated with knowledge and receptiveness towards BST, and to predict potential structural changes in the California dairy industry due to the release of BST technology.