This is a scenario from a paper in 2nd review. The authors used a one-way anova with obvious dependent measures. For example, if they had 100 participants, and if they were measuring say calories expended, then they observed calories expended while participants (for example) engaged in activity A, activity B, and activity C. AND, observed under those 3 activities under two different conditions (X and Y). So their one-way anova had a sample size (say) of 600. All the makings for a two-way repeated measures anova. Which was my comment on review #1. The authors are however sticking to their guns. And they gave no argument. Their paper is excellent otherwise. So I'm wondering if this could get through somehow as is. Despite the breach of independent observations. The X and Y conditions were assessed on different days. But the A, B, and C conditions were observed all at once. In other words, they engaged in one main task during the X and Y conditions, and observations were able to tease apart calorie expenditure in conditions A, B, and C.
