Dear all, I have a question my Prof also has problems with and I canít find any answer in the web. Maybe I find somebody here who can help:

I run a model with four effects: HostSex, HostPopulation, ParasitePopulation, and Temperature. I know that I have to be careful to interpret the significant main effects when those effects are also included into a significant interaction effect. What to do in this case is described here for example quite well: http://www.theanalysisfactor.com/int...s-interaction/

But what to do when you have a significant three-way interaction AND a significant two-way interaction? For my example I could imagine two different cases:

1. A significant two-way interaction which includes two terms that are included both into the three way interaction.

For example:
- HostSex x HostPopulation
- HostSex x HostPopulation x Temperature

2. A significant two-way interaction which includes just one term that is also included into the three-way interaction.

For example:
- HostSex x HostPopulation
- HostSex x ParasitePopulation x Temperature

I think in both cases I can interpret the three-way interaction. But how to deal with the two way interaction? Do I have to test it in both cases (1 and 2) or just for case 1 and how do I test if it stays interpretable?

Thanks a lot for your help!