I have to analyze results from an experiment and I’m not sure about how could I do it… The experiment is about the effect of some particularly distorted image typology on response time. The design is as follows: 30 participants watched two movies (A/B), each of them in a different image quality (distorted/no-distorted). Relationship between movie and IQ, order of presentation of IQ, and order of presentation of movie are counter-balanced between participants. Therefore, I understand it is a mixed design with 2 within-subjects IV (movie and IQ) and 3 between-subjects VI (relationship movie-IQ; order of movies; order of IQ). Dependent variable is reaction time (RT) to beeps inserted on movies. There are 10 beeps in each movie.
I think it is not a completely mixed design, as all the participants watched both movies AND both IQ, but both movies IN both IQ. This is because we are mainly interested on the effect of IQ on RT, and not in movie effect, and we cannot use the same movie twice within a participant to avoid a “repetition effect”. To analyze results I have calculated mean RT of each participant in each IQ condition, and conducted a repeated measures mixed ANOVA: 1 within-subjects factor (IQ) and 3 between-subjects factors (relationship movie-IQ, order of IQ, order of movie). Then I have conducted another similar ANOVA, but with Movie as within-subjects factor. Results are marginally significant, so I would like to carry out a more powerful analysis. I think that I could do that if I do not collapse means of 20 beeps in each IQ condition, and compare Beep 1 in no-distorted condition to beep 1 in distorted condition, and so for each beep, and finally compare the differences between the two groups of beeps (the 2 IQ levels). But I am not sure about how could I do that in SPSS: I guess I have to create a column for each beep (1, 2,…10) in each condition, but then I don’t know how to create the both groups of beeps (one for each IQ condition), that is my actual interest. I hope someone could help me! Thank you!
