number of cuts- ologit

[ali tgh]

hello to all
I want to categorized a data sample based on the income into categories such as " low class" " high calss" "middle class" and then run ordered logit to see propability that people with different gender and race fall into different category. Also, I can increase my categories to 5 and add " lower middle class" and " higher middle class". How can I find out which one is better, a model with 3 categories or 5 categories?which one performs better ?

[Phil Bromiley]

You'll increase your chances of a useful answer by following the FAQ on asking questions - provide Stata code in code delimiters, readable Stata output, and sample data using dataex.

In general, if you have finer data that is better than coarser data so 5 is probably better than 3. However, when you ask "which one performs bettter", you would need to specify the criterion for "better". By the way, if I were working on ordinal logits, I'd look at Professor William's website and material associated with gologit2.