The two aims of this study were to investigate the use of statistics in the surgical literature and to assess the degree of statistical comprehension possessed by graduating surgical residents. Two hundred journal articles were randomly selected from the 1984 issues of four surgical journals and were reviewed for statistical content. A classification of statistical techniques was created. A reader who has knowledge of descriptive statistics only has access to 44.5% of the articles. The addition of knowledge of t tests, contingency table analysis, other nonparametric techniques, and life table analysis to a reader's repertoire increases the access rate to 80.5%. The data indicate the specific statistical techniques that would best serve the surgeon who is attempting to increase access rate to the surgical literature. Ninety-one surgical residents in their fifth postgraduate year (PGY-5) responded to a questionnaire regarding their knowledge of statistics. While 90% of the respondents thought they would benefit from a course on statistics, 92% reported that they had received less than 5 hours of instruction in statistics during their residency. Both subjective self-ratings and objective testing revealed that the residents surveyed have a suboptimal knowledge of statistics. The results suggest the need for formal instruction in statistics during surgical residency.