External Email - Use Caution Hi Courtney,
I have two general remarks: 1.) For comparing n groups, the general rule is to have (n-1) binary indicator variables (and possibly n-1 interaction variables). One group (a reference group) will be implictly modeled by the intercept (1) and time (2) regressors. As a consequence, you should drop one group regressor and one interaction regressor from your design matrix (assuming that you have exactly 3 groups). As an analogy to your analysis, consider the example given on the Freesurfer LME tutorial page where there are 4 groups (HC, sMCI, cMCI, AD), but only 3 group indicators (sMCI, cMCI, AD) and 3 interaction terms. HC is chosen as the reference group in this example and not explicitly modeled. 2.) Any single (possibly multi-row) contrast matrix is expected to give only one single result, i.e. one F- and p-value. For example, a classical F-contrast that tests whether or not there is a difference between *any* pair of (more than 2) groups would be specified as a multi-row contrast matrix. To test for a difference between a specific pairs of groups, use a single-row contrast matrix (more precisely, vector). Best regards, Kersten On Mi, 2018-07-25 at 16:19 +0200, Courtney Haswell wrote: > External Email - Use Caution > > We are working in Matlab to analyze longitudinal data using the mass- > univariate approach to Linear Mixed Effects (LME) Modeling. > > > Step 1: Below is the design matrix (first few rows) we created in an > effort to compare subjects with low, moderate, and high drinking > behavior with respect to time (3 groups). > > > 1 0 1 0 0 0 0 0 > 1 3.0 1 3.0 0 0 0 0 > 1 0 1 0 0 0 0 0 > 1 1.2 0 0 1 1.2 0 0 > 1 2.3 0 0 0 0 1 2.3 > 1 3.5 0 0 0 0 1 3.5 > 1 0 1 0 0 0 0 0 > > Column 1 - intercept ones; 2 - time (years since baseline scan); 3 - > low drinking group; 4 - low drinking x time; 5 - mod drinking group; > 6 - mod drinking x time; 7 - high drinking group; 8 - high drinking x > time > > > Step 2: To compare low/moderate drinking, mod/high drinking, and > low/high drinking, we created the contrast matrix below: > > > CM=[0 0 0 1 0 -1 0 0; 0 0 0 0 0 1 0 -1; 0 0 0 1 0 0 0 -1] > > > Row 1 - low/mod drinking; Row 2 - mod/high drinking; Row 3 - low/high > drinking > > > Please let us know if the contrast matrix is appropriate for our > design matrix and the comparisons we would like to make. This seems > incorrect since there is only one output from > "lme_mass_F(lhstats,CM);" > > > Thank you for your help. _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer The information in this e-mail is intended only for the person to whom it is addressed. If you believe this e-mail was sent to you in error and the e-mail contains patient information, please contact the Partners Compliance HelpLine at http://www.partners.org/complianceline . If the e-mail was sent to you in error but does not contain patient information, please contact the sender and properly dispose of the e-mail.
