1IV with three levels and multiple DV's - Appropriate analysis?

#1
Hi all



I am a bit confused about the statistical analysis I should perform on my data (when I have it). I have been advised to do a multi-level modelling but I can't seem to get everything to fit and make sense. Alternatively, I have thought about doing a multiple regression. The measures are made up just to give an idea...



My IV is the score on an anxiety test and the three levels are control , at-risk and clinical.

My DVs are IQ score, performance on a breathing exercise and questionnaire and support network.

Confounding variables are stress and depression.

I want to look at if anxiety has a relationship with IQ score, breathing and support network.

I then want to see if IQ and breathing have a relationship with support network.

Finally, I would like to see how the three groups differ on the DVs.

Any help would be appreciated.
 

Karabiner

TS Contributor
#2
I want to look at if anxiety has a relationship with IQ score, breathing and support network.
(...)
Finally, I would like to see how the three groups differ on the DVs.
If you have already established that anxiety is (or: is not) related to the DVs, then
why repeat that analysis with the categorized variable?

I want to look at if anxiety has a relationship with IQ score, breathing and support network.
I then want to see if IQ and breathing have a relationship with support network.
The first you could perform using three linear regressions.
The second you could do also by using linear regression.

But maybe you have some mediator model in mind, such as
IV1 -> IV2, IV3 -> DV ?

With kind regards

Karabiner
 
#3
If you have already established that anxiety is (or: is not) related to the DVs, then
why repeat that analysis with the categorized variable?


The first you could perform using three linear regressions.
The second you could do also by using linear regression.

But maybe you have some mediator model in mind, such as
IV1 -> IV2, IV3 -> DV ?

With kind regards

Karabiner
Hi Karabiner

Thanks for the reply.

Would need to look at the differences between the three groups as it may provide insight into the progression of the disorder.

Yes I suppose a mediator model could be used in the study and there will be some confounding variables included like current mood state.