 |
 |
Due to getting better acquainted with the task and the surroundings, participants answer faster as the experiment progresses. Basically, I thought, that modelling reaction time this way leads to best results, but someone told me, due to different lengths of the conditions the modelling of events as lasting from onset until reaction leads to a statistical error. That is why I wanted to incorporate 4 regressors with the mean reaction time per condition and person. Since this approach didn't work, I don't really know, how I can incorporate reaction time differences between condition. Maybe important to know: the visual stimulation was shown for a fixed amount of time, but I didn't model that time period but instead the time period in which the interesting search processes should take place. I guess, I am only confused because I thought, that my model of the EVs takes into account the reaction time differences thereby accounting for more variance without inducing an error.