This paper explores eight hypotheses regarding trends in CALL literature relative to research subjects: small number of subjects, overuse of questionnaires, limited time, lack of experience with the application, task type and CALL in general, and lack of training before and during the study. Based on a review of 78 articles from four CALL-oriented journals from 2000-2003, it is concluded that (1) researchers commonly leave out or fail to gather relevant data on these characteristics; and (2) most of what we know about CALL is probably based on studies of untrained, novice users. Recommendations for improving this situation are offered.