A simple, intuitive method for computing confidence intervals in within-subject designs: Generalizing Loftus & Masson (1994) and avoiding biases of alternative accounts

摘要

Repeated-measures designs are common in the literature on motor behavior and, more general, in experimental psychology. Because of the correlational structure in these designs, calculation and interpretation of confidence intervals is nontrivial. One solution was provided by Loftus and Masson (1994). This solution, although widely adopted, has the limitation of implying the same-size confidence intervals for all factor levels and therefore does not allow assessment of variance homogeneity assumptions (i.e., the circularity assumption, which is crucial for the repeated measures ANOVA). This limitation and the method?′s perceived complexity has sometimes led practitioners to use a simplified variant, based on a per-subject normalization of the data (Morrison & Weaver, 1995; Bakeman & McArthur, 1996; Cousineau, 2005; Morey, 2008). We show that this normalization method leads to biased results, and we provide a simple, intuitive generalization of the Loftus and Masson method that allows assessment of the circularity assumption. Using typical data from our own grasping experiments, we show to which extent these effects can affect the interpretation of experimental data.