Which model?

#1
Hi all,

First time poster at the end of their tether with their stats at the minute.

So, I'm working with a relatively large dataset (n=5,048) of survey respondents. I have split these respondents into three categories, A (n=2072), B (n=1056), C (n=1920), and want to compare the three.

I have created three numeric composite scores from my dataset, two of which are explanatory and one of which is the response variable, all based on current literature.

I want to complete two separate analysis, investigating if there is any influence of one explanatory variable on the response variable, taking into consideration the three categories.

I also have a number of variables which I would like to include as random effects, as current literature suggests they might have an effect, although they are not the focus of this study.

I am using SPSS to complete this analysis.

My issue is I literally have no idea which model to run to show me this. Once I know, I'm happy to work out how to do it and interpret the output, I just feel like everywhere is telling me to try a different test?

If anyone needs any more information to help talk me through this I am happy to oblige.

Thank you for reading this at the very least!
 

Karabiner

TS Contributor
#2
Not sure what your exact question is. The most simple thing would be an analysis of covariance, with
your group variable as factor, and your explanatory variables as "covariates" in SPSS. You can add the
remaining variables if you wish. You can also tell SPSS which interactions between variables you want to
include or exclude, respectively.
With kind regards

Karabiner
 
#3
Thanks so much for the response. Maybe my issue is that I haven't defined my question well enough then ... ? Although I've had a little look into ANCOVAs and it seems very appropriate. Going to have a play around with it today. :) Thanks again!