Firstly I am very very new to stats...
I am trying to find a relationship between a dependent numeric factor and have 6 nominal, 1 ordinal and 4 numeric independents.
I have been using the category regression or optimal scaling on spss have been using group discretization for nominal varaibles with the number of categories and ranking discretization for ordinal using intial configuration as "random" given my limited exposure my questions are
I am running this on 200 points, is that gonna give adequate results?
My adjusted r square lies between 25% to 38%. How do I judge whether this is sufficient? Some psych journals claim it to be good others find it pathetic.
Also is there a better method available to find a predictive relationship between the variables described above?
Are there any other issues or mistakes with optimal scaling and the way I am using it?