Crossover-Repeated Measures Designs: Clarifying Common Misconceptions for a Valuable Human Factors Statistical Technique

摘要

Human Factors experiments often involve complex experimental designs that require complex statistical analysis. In practice, however, these complex models are often evaluated using oversimplified analyses that do not adequately account for statistical factors that can impact the interpretation of results. This article discusses a useful design for Human Factors experiments: the crossover-repeated measures design. It illustrates the importance of oft-ignored analytical steps in this design and how they can lead to different interpretations of the data and misleading conclusions. The article is intended for use as a refresher and includes explanations of statistical terminology and the components specific to crossover-repeated measures designs. Finally, it provides a case study of how proper and improper statistical methods can lead to drawing different conclusions from the data. SAS Code of the analyses performed for the case study can be found at .