What's your experience and explanation.
In conducting logistic regression analysis, a variable may be significant in
a selected context/model. If it is supected that there may be some
intercation with another variable in the model, an intercation term can be
entered.
If on entering the interaction term (maintaining the main effects) that
variable loses its significance and the interaction term is also not
significant - how is this explained (partial?)
Also, with SPSS, the dialogue box allows for a correlation output matrix with
logistic regression. What type of correlation is this and what exactly does
it say (I thought it allowed to check for multicollinearity). I obserbed that
for two variables (interval scale) used in this model the coefficient was
-8.8 and when I checked the pearson correlation coefficient for those two
variables by themselves the coefficient was .47 (sig).
How can this be explained.
Thank you
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|