I have been searching for an efficient way to analyze simple effects in a 2 x 2 mixed ANOVA--that is, one between subjects factor and one within subjects factor. There is a way in SPSS, but it is cumbersome. It appears PHIA package in R may be the best option. If you have worked with this package and have an example of how to probe a 2 x 2 interaction with simple effects analyses, I would like to talk to you. If you've integrated code for a figure, that would be a plus.
Hello Nicholas. How are you going about it in SPSS? I would add two EMMEANS subcommands to the GLM command, as follows:
/EMMEANS=TABLES(A*B) COMPARE(A)
/EMMEANS=TABLES(A*B) COMPARE(B)
You'll have to replace A and B with the names of your variables (or factors). The first line gives the simple main effects of A at each level of B, and the second the simple main effects of B at each level of A. HTH.
Likewise, in R, it's relatively straightforward to fit a mixed effects model with lmer or nlme, and then use emmeans to tease out what you need from the effects.
I haven't used the phia package, but it's really pretty easy to specify contrasts in emmeans. So my suspicion is this package won't help you much on that front.
Bruce Weaver, thanks for the code. That worked well! I added a few lines of code, modeled after your code. I will attach the data and the SPSS syntax in case others find it useful. SPSS syntax:
DATASET ACTIVATE DataSet1.
GLM pretest posttest BY TherapyType
/WSFACTOR=Time 2 Polynomial
/METHOD=SSTYPE(3)
/EMMEANS=TABLES(OVERALL)
/EMMEANS=TABLES(TherapyType)
/EMMEANS=TABLES(Time)
/EMMEANS=TABLES(TherapyType*Time)
/EMMEANS=TABLES(TherapyType*Time) Compare(TherapyType)
/EMMEANS=TABLES(TherapyType*Time) Compare(Time)
/CRITERIA=ALPHA(.05)
/WSDESIGN=Time
/DESIGN=TherapyType.
You can find lots of example figures and regression models with full R code in my book: https://www.researchgate.net/project/Book-New-statistics-for-the-design-researcher
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA