I need suggestions on analyzing (in SPSS) the IV, DV, co-variates and confounders from my study. I have attached a pdf with more details to this aim of my study.
The dependent variable of interest is only the first one (survey toal summed scores) or all three of them?
The dependent variable is the survey which measures perception of pain management care. The independent variable if a patient's length of stay (in months). The co-variate variables are pain intensity (an intensity rating 0 - 10) and depression survey with a score of 2 or greater (0 - 5). I am treating age (in years) and gender (male or female) as confounding variables, but not as co-variate variables.
I saw in the .pdf file you have three variables measured in both time intervals.
All the covariates-independent variables have to be put in the independent variables box except for the binary ones (0-1) which will be put in the between-effects box. The way you interpret the independent variables is your thing then.
This is in the case where all the independent variables are emasured once. In your case I think you have time varying independent variables and thus repeated measures anova will not work. You need to go through the mixed model approach which is a bit more difficult.
One of the problems I have is that my DV and 2 of my Co-variate variables were measured twice (14-days apart), A third co-variate was measured as bring present or not (yes or no). My IV is measured as the number of months each patient was in the hospital. For SPSS and GLM, there is the "with-in" subject area, which is where I would input my DV and 2 of my co-variate variables for T0 and T1. I believe I would put the number of months (the IV) in the "between-subjects" area? The other co-variate, which is not a repeat measure does not seem to work well with GLM, and I am wondering how to manage its effect on the DV? Also, my two confounding variables, gender and age, I thought would go into the co-variate area of the SPSS GLM, but I do not see where that is permitted? It may be that GLM will not work in my case, because of the level of measurement among my seven variables?
I agree as well. I am going to meet with our statistician in order to revise the statistical analysis for the 3rd aim of my study. Thank you for your great input!
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS