I have pairwise comparisons data that I converted into Bradley terry parameters. If I have BT parameters from two different occasions, what is the right way to test the null hypothesis that both models come from the same underlying data, that they are the same? Rejecting this would mean I can say the models are different. Does the Swait-Louviere test do this? I am not familiar with it and can't find any information on it.



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.