A Multi-Way Analysis of Variance for Single-Subject Designs

way ANOVA for single-subject designs. The Shine-Bower ANOVA views the single subject as a response generator for which the responses to a particular stimulus are statistically independent and normally distributed about a central response value. A fixed trial factor is introduced into the design which is then regarded as a fixed factor, two-way ANOVA with one observation per cell. Only an even number of trials is permitted. The design is presented in Table 1. Any correlations introduced into the data are carried by the pattern of nonerror effects. A modified error term, constructed on the assumption of a slow change across trials for the main effect of trials (expressed by the statement a ¡ = 0~+1 for odd i), is used to construct an approxi-mate F ratio for testing the main effect of conditions. The Mean