A study examined the efficacy of the keyword method of vocabulary instruction by comparing it with five methods designed to increase semantic processing of the definitions of the vocabulary words. Subjects in all five experiments were college students. In the first three experiments, recall of the definitions from the vocabulary words was the critical dependent measure, with the keyword method producing greater learning than any of the semantic-based or control conditions. Also, none of the semantic-based conditions facilitated definition recall, relative to a no-strategy control condition. In the fourth and fifth experiments, the keyword method, two semantic strategies, and the no-strategy control procedure were compared with respect to associative- and response-learning components of vocabulary learning. Results indicated the keyword method enhanced vocabulary/definition (associative) learning, but not definition (response) learning per se. In contrast, the semantic conditions tended to increase nonassociative learning of the definitions. The results support the case that the keyword method is a vocabulary-learning procedure superior to the semantic-based strategies advocated by reading theorists. (HOD)