The purpose of this project is to conduct research in statitical methods and computer techniques with particular emphasis on those appropriate for analyzing data from clinical, diagnostic, and prevention trials and epidemiologic studies of cancer. Many of the problems studied under this project arise from the consultative activities of the Section. During the past year major emphasis has been placed on developing an elaborate computer program for projection of cancer incidence and mortality rates into the future (See Project Z01 CN 00142 SORB). Another major area of emphasis has been the initial development of an interactive user-friendly computer program for determining samples sizes for various kinds of experimental designs. For this project, the best existing programs were combined into a single program and dialogue written to elicit the necessary questions from the user to determine which calculations should be performed. Because of the need for an appropriate design for large-scale prevention studies of smoking cessation in heavy smokers, attention was given to the problems of randomization by groups and a satisfactory design with adequate statistical power was developed for a specific project in this area. Work on designs for prevention studies has continued, particularly the investigation of the new class called partial factorial designs in which certain cells are deliberately omitted from full balanced factorial designs either because of anticipated toxicity or ethical considerations. Work has continuued on the problem of missing covariates in non-linear regression models focused on the situations under which estimates of treatment effect may be biased or tests of the null hypothesis may have incorrect statistical size.
