I want to conduct a within-subjects experiment with one independent variable with four levels. However, I ran a few informal pilot tests, and exposing participants to all four conditions seems to negatively affect results on the later conditions. That is, participants get bored of the experiment by the second or third treatment, and this shows in their responses.
If possible, I would like to expose participants to only two of the treatments (properly cross-balanced) but have no clue how to analyse the data then as, I think, a normal RM ANOVA would expect participants to have values for all four treatments.
NB: This might have been asked before, but the closest question I could find had different underlying assumptions, namely, that different "treatments" (in this case, different questions on a questionnaire) do not have any influence on each other).