I am currently analyzing data and I am not sure of what would be the best function and script I should use in SAS. I am looking if there is a difference in body composition (BMI Z-score) between children with congenital heart diseases depending of the physical activity restrictions they have. These restrictions can be imposed by the doctor (MDany), the parents (PAany) or both (Bothany).
Thank you !
You could calculate biserial correlations between the (continuous variable) BMI Z-scores and the 3 dichotomous variables MDany, PAany, and Bothany. In attachment a SAS example on how to calculate (a.o.) the biserial correlation.
To test for the differences between these 3 biserial correlations, I guess you would first have to convert these to Z-scores. Maybe this can help:
https://www.psychometrica.de/correlation.html
Thank you very much Jos Feys . Biserial correlation was the term I was looking for.
If you only measure effect of restriction on BMI once for each restriction type,why not just do a one way ANOVA of the BMI differences (Before-after) with the comparison groups Md, Pa, Both. .
- Badges
- Science topic
- Similar topics
- Social Psychology
- Attitude