1

As part of a larger analysis, I'm interested in examining the relationship between an interval level variable and a four category nominal variable. In this case, the interval variable is the independent variable and the four category nominal variable is the dependent variable. I'm interested in whether there is a statistically significant relationship between the two variables, as well as the strength of the relationship if one exists.

Does anyone know whether a measure of association exists for this situation? I'm aware of eta, but my understanding is that eta is appropriate when the IV is nominal and the DV is interval. I'm also aware of the point bi-serial correlation, but my understanding is that the nominal variable must be binary.

Any help is greatly appreciated.

Matt
  • 11
  • 1
  • 2
    Sounds like you might be looking for [multinomial logistic regression](http://en.wikipedia.org/wiki/Multinomial_logistic_regression). – Nick Stauner Jun 26 '14 at 01:10
  • Thanks for the comment. I'm familiar with multinomial logistic regression, but I'm wondering whether there are other methods. For the first part of my analysis, I'm simply looking for a measure of association, sort of in the way you might look at simple correlations before moving onto regression. – Matt Jun 27 '14 at 14:39

0 Answers0