I have a sample size of 30, representing three different groups (one intervention and two controls). I would like to see if there is an association between my dependent variable and 4 self-report measurements when controlling for group membership. In other words, could the impact of the intervention be explained by differences in the constructs measured by self-report measurements. Is it possible to control for group categories in this way?
Dear Dan Kolubinski,
I suggest you to see links and attached files on topic.
Lecture 24: Partial correlation, multiple regression ... - Ernesto Amaral
www.ernestoamaral.com/docs/soci420-17fall/Lecture24.pdf
Linear or rank partial correlation coefficients - MATLAB partialcor https://www.mathworks.com › ... › Descriptive Statistics
Example 2.9 Computing Partial Correlations :: Base SAS(R) 9.4 ...
support.sas.com/documentation/.../procstat_corr_examples12.htm
SDTM mapping becomes more EASY with SAS Clinical ... - Lex Jansen
https://zdoc.site/sdtm-mapping-becomes-more-easy-with-sas-clinic...
Best regards
Hello Dan,
An easy way to get at the answers you seek would be to do the following:
1. Dummy code the group membership into two variates. For example, if case is in intervention group, code the two variates as 1, 0. If case is in control group A, code the two variates as 0, 1. For control group B cases, code the values as 0, 0.
2. Run a hierarchical multiple regression. In stage 1, enter the two dummy coded variates as IVs, and use your dependent variable as the DV. Note the obtained R-squared.
For stage 2, add the four self-report measurements as IVs. Note the new R-squared.
3. The explanatory power of the self-report measures, controlling for group membership, is quantified as Stage 2 R-squared minus Stage 1 R-squared.
Good luck with your work!
Hi Dan,
Have a look at multilevel (partial) correlation in:
https://github.com/easystats/correlation#hierarchical-partial-correlations
Best wishes,
Mariano
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size