Hello,
I've been having issues interpreting the results from a set of data I have. Basically, it's a within-subjects design with two experimental conditions (Condition 1 and 2), and one continuous between subjects predictor. The DV is a measure of bias. The interaction between the experimental conditions and the predictor variable on the bias measure is nonsignificant, however my parameter estimates suggest that in Condition 1, people who were higher on the predictor variable were lower in bias (significantly), but not in Condition 2 (it was non-significant). How do I interpret this in my results without sounding like I am talking up a non-significant interaction?
Hi Olivia,
If your research question is based on the overall experimental conditions, it might be of value to present the non-significant result, which will still be useful research!
You might want to go further and stratify your results based on the specific experimental conditions, highlighting the significant finding.
Regards,
I don't quite understand. Have you done the interaction plot? If that is nonparallel you should expect that the interaction is significant unless the sample size is too small. This could give you a clue if there is a problem somewhere. Interaction plots are discussed and software is available from the CRAN website in
https://www.amazon.com/Analysis-Experiments-Chapman-Statistical-Science/dp/1439868131/ref=sr_1_fkmrnull_1?keywords=John+Lawson+Experimental+Design+with+R&qid=1557107902&s=books&sr=1-1-fkmrnull
An interaction plot would make your comments quite a bit more clear.
Best wishes, David Booth
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Regression