Problems with the Mixed-Design ANOVA

An earlier section indicated that the proper use of a control group in a mixed-design investigation can help remedy some of the weaknesses associated with the single-group repeated-measures design. However, there are other problems that can affect a repeated-measures analysis even when it includes a control group.

For example, Chapter 12, “One-Way ANOVA with One Repeated-Measures Factor,” describes a number of sequence effects that can confound a study with a repeated-measures factor. Specifically, repeated-measures investigations often suffer from order effects (effects that occur when the ordinal position of treatments introduces response biases) as well as carryover effects (effects that occur when the effect from one treatment changes, or carries over, to participant responses in the following treatment conditions).

You should always be sensitive to the possibility of sequence effects when conducting research with any type of repeated-measures design. In some cases, it is possible to successfully deal with these problems through the proper use of counterbalancing, spacing of trials, additional control groups, or other strategies. Some of these approaches are discussed in the “Sequence Effects” section of Chapter 12.