I have conducted a research with basically a following design. There are three within-subject conditions (A, B, and C) and participants filled in the same 10-item questionnaire (answered between 1 to 7) in each condition. The questionnaire has three factors (1, 2, and 3), so I have calculated factor scores by averaging responses for the items within each factor. So, individual participant has 9 sets of scores (A1, A2, A3, B1, B2, ect.). I want to know how the factor scores differed across conditions. I thought I should either run 3(condition) by 3(factors) ANOVA, or run one way ANOVA (3 levels of factors) separately for each condition.

However, my collaborator said that it is not allowed to regard factors as different levels in ANOVA, because factors are qualitatively different from each other and cannot be compared. So, it is meaninless to figure out, say, factor 1 is higher than factor 2 in condition A but the direction is reversed in condition B.

Is it true? And if so, can anyone explain more about the qualitative differences between factors that prevent us from using ANOVA, please?

Similar questions and discussions