Exact chi-square tests versus asymptotic assumption

Discussion in 'SPSS' started by LogReg, Feb 3, 2006.

  1. LogReg

    LogReg Guest

    Are there conditions when you would favour an aymptotoic chi-square
    (X^2 or G^2) p-value over the exact? Any references or weblinks on this
    topic would also be appreciated.

    Thanks in advance. This forum is great!
     
    LogReg, Feb 3, 2006
    #1
    1. Advertisements

  2. LogReg

    Thom Guest

    I'll have a stab at it.

    If the tests make the same assumptions and test the same hypothesis I'd
    suggest: never. OTOH think I'd prefer an asymptotic test which makes
    more plausible assumptions and tests a hypothesis I'm interested in
    over an exact test that doesn't.

    In practice issues like robustness would also come into it.

    Thom
     
    Thom, Feb 8, 2006
    #2
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.