Repeated Measures Analysis of Variance

Terminology

There are several statistical terms commonly used to describe a repeated measures ANOVA. When investigations involve variables pertaining to studied participants, a sampled member is often referred to as a subject. In education, subjects are often students. When the same dependent variable (outcome) is measured repeatedly for all subjects across a set of conditions, the set of conditions is referred to as a within-subjects factor. For studies involving one group’s standardized test scores on three occasions, the within-subjects factor would be the Time (e.g., Time 0, Time 1, and Time 2). The conditions that contextualize this factor is often referred to as trials. When the outcome of interest (dependent variable) is measured three or more times on different groups (such as control and intervention groups), the set of conditions is called the between-subjects factor. For educational studies involving the assessment of a control and an intervention group’s standardized test scores on three occasions, the between-subjects factor would simply be groups. In this case, the research design would be a two-way repeated measures ANOVA.

When to Use Repeated Measures ANOVA

In the context of educational measurement, the repeated measures ANOVA is generally used in two different types of research conditions: studies that investigate (1) changes in means over three or more time points or (2) differences in means under three or more conditions.

For the first example, we may be investigating the effect of a new mathematics program on students’ performance on a standardized test at three separate time points: Time 0, Time 1, and Time 2 (pre-, midway-, and postprogram intervention). This would enable us to develop an understanding of the possible timing and extent of improved mathematics ability. In this case, the within-subjects factor would be Time with three levels.

For the second example, we may be interested in the ability of students to recall historic events and associated dates and make use of three learning strategies. This might help us determine which strategy might best suit the students in the class. In this case, the within-subjects factor could be deemed the study condition.

...