Hi all, I have conducted a study where respondents were asked two similar tasks (choosing colours/ tastes) about two different product categories (in random order). These two variables can take values from 1 to 6. Before these two tasks (within-subjects), respondents were randomly assigned to 4 different treatments (between subjects). I have conducted ANOVAS on each product category task, with IV the treatments (4 dummies) and the tasks order. Unfortunately I find a significant effect of tasks order so, even if I randomised the tasks, I can only consider the first one valid...
Therefore, in order not to loose that much data I wanted to run a linear mixed regression (for "repeated measures", though it is not exactly the same measure repeated twice a it refers to different product categories). I have transformed my data from wide to long and now have 2 rows per each respondent and a single dependent variable that includes the two within-subjects tasks. I also have a dummy that indicates which category the tasks refers to. However, when I try to run the mixed model, I cannot set this new DV as it is also the "repeated measure" one. Any though how I can fix this problem and/or how I could analyse the data instead, without loosing the information included in the second task?
Thank you so much in advance for any suggestion!
------------------------------
TV
------------------------------
#SPSSStatistics