Hello to community,   Than you for your time, the problem is following  I am doing classical Pavlovian human conditioning which in very general terms is changing of attitude  of a person to some stimulus.   One of the way to check it is simply to ask the person to rate the attitude to the symbol (in my case : the level of excitement on the scale from one to nine) .   It seams  rather easy task but due to the complexity of paradigm it becomes a little bit complicated.   First  we show a person two symbols: one we associate with money the other one is neutral and ask person to rate those symbols. then we switch association and previously neutral symbol becomes associated with money and previously associated with money becomes neutral. Then we do the same, but use pain instead of money.   In the end we have ratings for 8 different  associations each made by 20 participants.  In my model I have 4 categorical predictors. 1. Stimulus : can be associated with pain\money, or not (CS+, CS-). 2. Stage : stimulus before switch (original) and after switch (inverted). 3. Phase: stimulus associated with money (pleasant), associated with pain (unpleasant). 4 Subjects: 20 of them, each makes 8 ratings. My very general question is: what are your advises?  More actual question is: Definitely it is a repetitive measurement, but what parameters should be treated as repetitive.   Than you for your time. ... View more