1

I am a first year graduate student in biostatistics, and I have somewhat of an idea of the difference between REML and ML. However, I want a more in-depth understanding of each estimation method, especially when it comes to Likelihood ratio tests.

I know that you cannot use REML for likelihood ratio tests, and it has something to do with using the residuals of the means. If you take the residuals of the means, the models are no longer nested (nested models are needed for LRT). However, I really want a more thorough explanation as to why we cannot use REML for LRT's.

I will appreciate it if someone can explain this to me.

Sycorax
  • 76,417
  • 20
  • 189
  • 313
Shyamali Mukerjee
  • 11
  • 3
  • 2
    Possible duplicate of [What is "restricted maximum likelihood" and when should it be used?](http://stats.stackexchange.com/questions/48671/what-is-restricted-maximum-likelihood-and-when-should-it-be-used) and see also [Why does one have to use REML (in stead of ML) for choosing among nested var-covar models?](http://stats.stackexchange.com/questions/99895/why-does-one-have-to-use-reml-in-stead-of-ml-for-choosing-among-nested-var-cov/171529#171529) – Sycorax Apr 20 '16 at 20:38
  • I think you will find the information you need in the linked thread. Please read it. If it isn't what you want / you still have a question afterwards, come back here & edit your question to state what you learned & what you still need to know. Then we can provide the information you need without just duplicating material elsewhere that already didn't help you. – gung - Reinstate Monica Apr 20 '16 at 21:19

0 Answers0