I am looking at the impact of circumstance(could cheat/ could not cheat) on the dishonesty by using Repeated measures ANOVA.
Participants played may trails in my study. In each trail , they could freely choose to do Task A or Task B. Then they confront with a particular circumstance(could cheat/ could not cheat), and then they decide whether to be dishonesty or not.
The circumstance was a with-subject factor( could cheat / could not cheat) , the dishonesty was the dependent factor.
Now I wanted to control variables of task-type(Task A / Task B ), which was also a within-subject factor but not control in my study(participants could freely to choose which task to do).
In SPSS, I used Repeated Measures ANOVA. I set circumstance as within-subject in RMA, dishonesty as dependent variable, and task-type as covariates. But I was confused whether i should put task-type*circumstance in the model ?
Actually I tried this and found the circumstance*task-type was significant(
You could run a 'logistic analysis'; i.e., a binomial logistic regression for those situation where all the explanatory variables are categorical. You can very well include the interaction if this makes sence. In your case: does it make sence that e.g. honesty in Task A when they can cheat, is not the same as honesty in Task B when they can cheat.
In your case, run the logistic regression:
honesty = TaskType + Circumstance + TaskType*Circumstance
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS