Simple Main Effects in a 2 × 3 ANOVA
Hi,
I'm teaching factorial ANOVA in JASP (v0.98.1) and would appreciate some clarification.
Suppose I have a 2 × 3 ANOVA with:
- Diabetes status (2 levels)
- Wound dressing (3 levels: Hydrocolloid, Silver, Standard)
and a significant interaction.
If I set Simple Effect Factor = Diabetes, JASP tests the diabetes effect within each wound dressing. Since Diabetes has only two levels, the analysis is complete.
If I instead set Simple Effect Factor = Wound dressing, JASP tests the effect of wound dressing within each diabetes group. If a Simple Main Effect is significant, I would normally expect pairwise comparisons among the three wound dressings within that diabetes group.
However, I cannot find a way to perform these follow-up comparisons in JASP. The regular Post Hoc tests average across diabetes status, while the interaction Post Hoc compares all six cell means, neither of which seems to answer this question.
Am I missing something, or is this functionality currently unavailable in JASP?
I've also submitted a GitHub feature request suggesting that pairwise comparisons be added directly to the Simple Main Effects procedure, as I think this would make the workflow much clearer for teaching and more consistent with standard factorial ANOVA practice.
Thanks in advance!
Per the JASP lover🥰
Comments
They way I've been able to handle this in JASP is to use Contrasts--in particular, custom contrasts for the interaction. I set each of the two to-be-contrasted conditions to a weight of 1 and -1, respectively. I set all other weights to zero.
R
You might be able to get some tests by using the interaction term under Post Hoc Tests:
This gives pairwise comparisons between each cell of the design (not just within each level of the moderator)
Actually you know what, I got what you're looking for by additionally checking the box "Conditional comparisons for interactions":
This gave me pairwise comparisons for education level separately for each gender.