 |
 |
Dear Feng To be comparable the second model would have to have two nuisance regressors (i.e. one for each group padded with zeros in the end, or the beginning) since you are comparing the model to a scenario where you project out the regressor with a group specific slope. The only thing that would then make a difference is that the latter approach has the correct degrees of freedom number i.e. Number of subjects - 4. Best Torben > Den 1. jun. 2018 kl. 03.41 skrev chenhf_uestc <[log in to unmask]>: > > Dear SPM experts, > > I have a question about the way of regressing the nuisance covariate (e.g., age or gender) in the between-group comparison (two-sample t test). Here, I have two methods. First, I regress out the nuisance covariate separately in each group, and then perform two sample t test using the residuals of the two gourp. Second, like SPM, I add the nuisance covariate in the geneal linear model, estimate the parameters, input contrast (1 -1 0) and finally obtain the results. > > However, the two results are not equal. Thus, I am wondering which one is right, or which one is better? > > Thanks in advance! > > Best, > Feng > > > > To unsubscribe from the SPM list, click the following link: > https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=SPM&A=1 <https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=SPM&A=1> To unsubscribe from the SPM list, click the following link: https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=SPM&A=1