I conducted 2*2*2 experiment and only main interaction is significant (2*2*2). All the other interactions and direct effects are not significant. Please let me know the steps of interpreting the results.
I used the two charts to show the mean difference in three variable interaction setting. However, when i use 't' statistics to show the difference again they are insignificant. What should i do.
It is not easy to test a three-way interaction and interpret them.
First, test the three way interactions (X1*X2*X3) using a regression analysis.
constant
b1 (X1)
b2 (X2)
b3 (X3)
b4 (X1*X2)
b5 (X1*X3)
b6 (X2*X3)
b7 (X1*X2*X3)
Then, use the attached excel file...then you can make the figures for the three-way interactions.
Hi,
Thank you very much but, my factors are dichotomous in nature.
Rasika:
First of all, testing dichotomous predictors in regression is perfectly fine, so Tae-Yeol's suggestion still stands.
Secondly, it looks to me like the plots you attached illustrate the nature of the three-way interaction pretty clearly. There is a different sort of similarity*crowdedness interaction in the 'honest' condition than in the 'deceitful' condition.
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis