The week in stats (Nov. 18th edition)

2 comments

  1. Can goodness of fit test statistics be used for association/independence?

    There are some goodness of fit test statistics like Pearson chi-square, G square , Freeman-Tukey etc..
    Can these test statistics be used for association/independence? ( I mean to see if there is an association between 2 variables)

    The chi-square test statistic can be used for independence.
    But what about the other test statistics, like G square(log-likelihood ratio), Kolmogorov-Smirnov, Freeman-Tukey, Power divergence test statistic? These are goodness of fit test statistics. But I need to know if they can also be used to determine independence?