Hi everyone! I was wondering if anyone might have any ideas of how to conduct some potential analyses for some research I am working on. I am interested in the moderating role of trait emotion regulation on the relationship between state anxiety prior to solving an interpersonal problem (predictor) and problem solving in a real-life interpersonal problem (outcome). I want to hypothesize that individuals with high levels of state anxiety perform well if they also have good emotion regulation skills, but individuals with high levels of state anxiety do not perform as well if they do not have as good of emotion regulation skills. I have used diary cards for participants to track interpersonal problems that they have faced (the responses are coded) as well as their state anxiety level prior to trying to solve the problem. Each participant rated their state anxiety prior to 4 interpersonal problems that they faced (predictor) and wrote about how they solved that problem (outcome). They filled out one measure about trait emotion regulation skills. I have a sample of 50 participants right now. I do not think I can treat each scenario as if it were a different participant (then I would have a sample of 200 because 4 scenarios per participant and there are 50 participants) and run a test of moderation once. Is there a way to do some kind of repeated measures moderation without averaging participants' scores (this would defeat the purpose of looking at state-anxiety)? Should I just run four different moderation analyses, one for the first scenario, one for the second scenario, and so on? Or is there a way I could use a repeated measures ANOVA to look for an interaction? I don't think it makes conceptual sense to treat the first scenario recorded any differently from the second scenario because there were no experimental manipulations to create hypothesized differences for scenario 1, 2, 3, and 4. I hope this makes sense. I look forward to hearing what anyone has to say!
I am not sure I understand your experiment completely but this looks to me like a study that is often carried out in drug effectiveness studies over time. It was developed originally to follow AIDS patients as they were treated across time. I am attaching a link to course notes produced by Prof. Marie Davidian who worked on this type of study. In general these are called longitudinal models. Hopefully these will be of use to you. Best wishes, David Booth PS I forgot to mention that moderation is the same thing as the interaction discussed in these notes. The link to the notes is:
https://www4.stat.ncsu.edu/~davidian/st732/notes.html
Moderation (interaction) can be tested in a regression model. Since you have four measurements per person (repated measures), you need a mixed-effects regression model.
Stephen Politzer-Ahles, thank you for your response. I looked into using a mixed-effects regression model, but am still confused. I have two within subject variables (state anxiety while solving an interpersonal problem and the effectiveness of the solution to the problem) that I collected four times each. When I look at examples of mixed-effect models they only use one within subject variable. Is there a way to use two within subject variables? Also, I am not interested in whether or not state anxiety or effectiveness is significantly different between times 1, 2, 3, or 4 when the data was collected and it seems like mixed-effect models provide you with this information too. I just want to see how ability to regulate emotions moderates the relationship between state anxiety during problem solving and the effectiveness of the solution of that given problem, but I have four data points for state anxiety and effectiveness.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA