In SPSS, I conducted a general linear model with repeated measures to determine changes in various blood markers, BMI, waist circumference, and blood pressure in youth from pre to post intervention comparing an intervention and control group using SPSS.
The within-subject factor is time with 2 levels (pre and post), the within-subject variables (dependent variables) include various blood markers, BMI, waist circumference, and blood pressure. The between-subjects factor is treatment group (intervention or control group), and the covariates were age, sex, and area deprivation.
Under "Tests of Within-Subjects Effects," the multivariate analysis notes there was no main effect for time nor was there a significant interaction between time and treatment group. Under the univariate tests, there was no main effect of time for any of the variables, but there was a significant interaction between time and treatment group for some but not all of the mentioned variables.
- Do I interpret the univariate rather than the multivariate analysis and conclude that there were significant improvements in some of the variables over time and between groups?
- Alternatively, do I interpret the multivariate analysis and conclude since there was not significant main effect or time by treatment interaction, the intervention had no effect on any of the variables over time between groups?