zanna clay posted on Thursday, July 10, 2008 - 4:31 am
the categories of my dv are very different and so im unsure which i should select as my reference category for multinomial logistic regression?
when i use the last reference category, i get sero values (.) for some of my predicotrs, as i think that these predictors are not important in that particular reference category, but in other categories. how do i overcome this porblem and know hw to select a reference category that can reflect the predictors contributions properly?
sorry if this message appears twice
zanna clay posted on Thursday, July 10, 2008 - 4:33 am
i meant multinomial logistic regression, not multiple as written in title!
The choice of reference category makes a difference - and it should because you are discussing different effects with different reference categories. For further information on multinomial logistic regression, see a textbook on categorical data analysis like Agresti.