help
help > RE: Between & Within subjects design for Repeated Measures
Apr 3, 2019 12:04 AM | Andrew Zalesky
RE: Between & Within subjects design for Repeated Measures
Hi Danie,
I'm not quite sure of the between-subject effect that you have in mind, however, if you are interested in a grouping that remains consistent across the 4 time points (e.g. patients vs controls), you don't need a repeated measures design to test for this kind of between-group effect. For each subject, you could simply average the 4 connectivity matrices so that each subject is associated with a single matrix.
If you really do want to model each time point, this is also perfectly fine. You would include an explanatory variable for each subject to model the within-subject effect. Each of the explanatory variables would comprise fours 1's and 0's elsewhere. You would then include one additional explanatory variable to model the between-subject effect (e.g. patients vs controls).
Andrew
Originally posted by Daniel Brennan:
I'm not quite sure of the between-subject effect that you have in mind, however, if you are interested in a grouping that remains consistent across the 4 time points (e.g. patients vs controls), you don't need a repeated measures design to test for this kind of between-group effect. For each subject, you could simply average the 4 connectivity matrices so that each subject is associated with a single matrix.
If you really do want to model each time point, this is also perfectly fine. You would include an explanatory variable for each subject to model the within-subject effect. Each of the explanatory variables would comprise fours 1's and 0's elsewhere. You would then include one additional explanatory variable to model the between-subject effect (e.g. patients vs controls).
Andrew
Originally posted by Daniel Brennan:
Hello,
I have a repeated measures design, 19 subjects x 4 timepoints (1 covariate and 1 connectivity matrix for each timepoint). I believe I understand how to model the within-subjects effects at this stage; however, I'm unclear, given the 4 timepoints, how to model the between-subjects effects other than simply averaging the connectivity matrices + covariates to now have 19 datapoints each instead of 76. Is there a way to set up the design matrix so that between-subjects effects can be modeled?
Thank you,
Danie
I have a repeated measures design, 19 subjects x 4 timepoints (1 covariate and 1 connectivity matrix for each timepoint). I believe I understand how to model the within-subjects effects at this stage; however, I'm unclear, given the 4 timepoints, how to model the between-subjects effects other than simply averaging the connectivity matrices + covariates to now have 19 datapoints each instead of 76. Is there a way to set up the design matrix so that between-subjects effects can be modeled?
Thank you,
Danie
Threaded View
| Title | Author | Date |
|---|---|---|
| Daniel Brennan | Apr 2, 2019 | |
| Andrew Zalesky | Apr 3, 2019 | |
| Daniel Brennan | Apr 3, 2019 | |
| Andrew Zalesky | Apr 3, 2019 | |
| Daniel Brennan | Apr 4, 2019 | |
| Daniel Brennan | Apr 3, 2019 | |
| Andrew Zalesky | Apr 3, 2019 | |