Driving Simulator Experiments: Power for Repeated Measures vs. Completely Randomized Design

Driving simulator studies are usually designed to collect repeated measures on each participant. This design has different implications for the power of within and between-subject effects that needs to be recognized by researchers. The power for between-subject variables decreases when additional measures are collected on the within-subject variables. However, the power for the main and interaction effects of the within-subjects variables increases as more observations are collected on one participant. If the main interest of the experiment is a between-subject effect, such as age, then a completely randomized design can provide the same power with fewer participants. Through a case study, this paper demonstrates how the power changes between a repeated measures design and a completely randomized design.