Dear Jens, thank you for your video. I wonder, since you left the cities out of your analysis, if you could leave them in instead? What would have to be considered as it is a nominal variable. Would you have to create 5 dummy variables (if there are five cities) and only take four of these variables for your analysis, or how can I go about it?
Exactly if I wanted to include the cities I would have to include dummy variables. To avoid multicollinearity problems you always select one of the possible outcomes as the base category and include the dummies for the remaining n-1 possible outcomes. The results are interpreted in relation to the base category.
Sir please share the data ..so, we can a also practice it on spss
Dear Jens, thank you for your video. I wonder, since you left the cities out of your analysis, if you could leave them in instead? What would have to be considered as it is a nominal variable. Would you have to create 5 dummy variables (if there are five cities) and only take four of these variables for your analysis, or how can I go about it?
Exactly if I wanted to include the cities I would have to include dummy variables. To avoid multicollinearity problems you always select one of the possible outcomes as the base category and include the dummies for the remaining n-1 possible outcomes. The results are interpreted in relation to the base category.
@@jensk.perret6794 thank you for your answer! This helps me a lot!
Why are we use discriminate analysis not logistic here?
Because this video is on how to use discriminant analysis. However a multinomial logistic approach would be an equally valid approach.