Model to use for a repeated measures experiment with a continuous between subject variable? (repeated measures ANOVA, ANCOVA or Mulit-Level approach)

For my bachelors degree I did an experiment on the effects of attachment style on a perspective taking task where people first filled out a survey which produces 2 continuous values ( for anxiety and avoidance). Second they did the perspective taking task and their response times where measured. The task constisted of 3 repeated measures per trial (each person did 208 trials) - prime (neutral, threat), consistency (consistent, inconstistent), and perspective (self, other).

Now I am struggling a bit on how to best analyze these data. In SPSS I tried GLM -> repeated measures with 3 factors (prime, consistency, perspective) with the 2 levels each and the anxiety, avoidance values as covariates. There where some significant outcomes, but I am not sure if I can trust that method?
I know that a Multi-level approach is also possible but with that I have absolutely no clue how to do that right and also how to interpret the outcomes correctly.

I really would appreciate your help! If necessary I can also upload my SPSS data and the syntax of what I have done so far.
Not really a problem but I am pretty unexperienced and I read that a multi-level model would fit better to my case. But as I do not have much time left to research on multi-level modelling I will stick to the ANOVA, should be enough for the bachelor thesis then :D.

Anyway, thanks for your reply!


OK now I have a question regarding the contrasts with the repeated measures. I did repeated contrasts and I need to know if I am interpreting those correct. I attached the within-subject contrasts and transformation matrix and as I interpret the first significant contrast (prime, with lvl1 as threat and lvl2 as neutral) means that people reacted significantly slower (higher response times) in the threat-prime (lvl1) condition than in the neutral-prime condition.

Is this interpretation correct?


I think my first interpretations is wrong, I guess making graphs helps with the interpretation right? And when I look at the graphs they actually show the exact opposite of my above mentioned interpretation. So going after the graphs should be the right way to see in which direction the significant interactions in the analysis go?

EDIT 3: I really do not know how to go with the 4-way interaction (at the bottom of the screenshot). Could someone give me a hint on how to start interpreting this? The interaction is still significant just with the avoidant variable as covariate. Unbenannt.PNG


Last edited: