Hello Ernesto,

Some folks would likely take issue with a single, Likert-type response score (1-7) being characterized as continuous (or interval strength). You could easily evaluate this model via ordinal regression.

Good luck with your work.

Hi David,

Thank you very David Morse much for taking the time to read my question and answer. My response (DV) is consisting of 3 different construct each measured with a validated 10 item-scale, I don't know if this would make a difference (you have a lot more data points).

I will have a look at ordinal regression, currently I am not familiar with this method, for this reason I was leaning towards the three way interaction repeated measures ANOVA.

Kind regards,

Ernesto

As Prof Morse indicated that would be a very poor way. Follow his advice and get a much better result. BTW regression models and ANOVA models are equivalent so you can follow his advice with your ANOVA type design. All of this can be found by google search. Programs are available in all major stat packages. Prof Morse knows his business so follow his advice.

Thank you very much David Eugene Booth

I will have a further look at an ordinal regression!

Thank you very much for taking the time to read my question and interesting article Haroon R Abbu .

It seems that the model you are proposing is a mediated moderated mediation however my model is a moderated moderation. I have already decided to approach this with an ordinal regression, but for another project this literature is really useful. Do you perhaps know some literature on interpreting/conducting a three way interaction ordinal regression?

Most of the literature I could find on forums such as stackoverflow and leard statistics focus on a simple moderation. However I am quite new in the research world so it could be I am not looking in the right place.

Best,

Ernesto