Hi all,
I have a few questions about two separate problems.
Firstly, I have a regression analysis I'd like to solve. It involves one explanatory continuous variable (A) and two categorical (age, sex, coded as 0/1) & one continuous (B) covariates. My model would be the following, correct?
-model (A) : (age) (sex) (B) 0 -bucket 0 ${output}
Now, in my output, the F-stat and R^2 would indicate the degree to which my explanatory data fits the data? Or, would I look at the Coef # est & tstat of my explanatory variable for this? I ask this because the coef est / tstat map of one of my covariates appears to display clusters I had anticipated the explanatory variable to modulate.
Covarying for this particular coefficient data introduces a similar expected cluster in what I believe to be the full regression model, but leaving it out of the model removes the cluster. What is an explanation for this? If the coef est / tstat map of my explanatory variable doesn't show significance at this area but the map of aforementioned covariate does, then is it safe to assume that this is just an effect of my covariate? Not sure what to interpret from the cluster in the regression subbrik.
Secondly, I have a separate dataset consisting of three groups: controls, patient responders, and patient nonresponders. I would like to find areas that discriminate between the groups in that nonresponders<controls<responders, and also in the opposite direction. Is the best choice to attack this question to run ttests on each pair of groups, determine their pairwise significance levels, and run a conjunction in which both nonresponders<controls significantly and controls<responders significantly, and vice versa? Or, is there a more sophisitcated way to address this?
Many thanks!
Thomas
