Run regressions to determine if the potential confounding variables have any effect on the DVs of interest. If they do they may be confounders. Then look at the attached screenshot search for more detailed information. Best wishes David Booth

You can convert age to a linear variable by assigning each category its midpoint (e.g., 30-35 gets a score of 33). For education, you can score it as years of education, so that a Master degree gets a score of 18, etc.

Are the three MMQ scores that you are analyzing completely separate? If so, there is no problem with the finding that they have different degrees of association with your independent variable.

Apologies to Prof Morgan but what is the definition of a linear variable? Further what does completely separate mean? Thanks in advance, David Booth

OK, instead of linear I meant interval-level. For "completely separate," I simply wanted to know if they were part of a larger 3-part scale, or whether each one was meant to be used by itself.

And as long as we are asking questions, how would you do the analysis in question if both education and age had 10 or more ordered categories? And would you allow the three scale variables to be treated as interval, or would they have to be multi-category ordinal as well? I'm especially curious about what the partial coefficients would look like in such an analysis.

David Eugene Booth Thank you very much for your answer,. I'm using Jamovi for my analysis and wondered if the regression you're suggesting would be the partial correlation option? I'm also interested as David Morgan has asked how these variables would be categorised.

David L Morgan thank you, I had wondered how best to convert age so this is very helpful. With regards to the MMQ, it is 3 separate subscales, each supposed to be treated independently.

Thanks Ashleigh for the question. Your model is described by the entries in the screen shot attached. See also Montgomery Design and Analysis of Experiments available in the z-library. Best wishes David Booth

Dear Ashleigh Hopwood, in a study in which the effect of several factors had to be evaluated, the statistical consultant used ANOVA and K-Square analysis and the effect of each variable was determined by controlling the effect of other factors; If the results of the analysis were significant, linear regression was used to find the group that made the results meaningful.