Characteristics of autism research were examined in this survey of a subset of studies published between 1971 and 1982. Results indicated that replicability of research, and appropriate generalization of findings, are likely to be impaired because of inadequacies of subject selection procedures and omission of important descriptive information. Although there were improvements in these areas over this 12-year span, increased use of objective, quantitative measures of diagnosis of autism and assessment of intellectual abilities of autistic subjects are needed.