# Phi MOA, Chi-Square and Null Hypothesis, What can be learned?

In short, is there anything to be learned from performing a Phi measurement of association (MOA) on a Chi-Square test where it failed to reject the null hypothesis?

$$\phi = \sqrt{\chi^2/N}$$

where $\chi^2$ is the value returned by Chi-Square and $N$ is the number of observations.

Or is the phi coefficient only useful on significant tests?

I haven't touched stats in a few years but I couldn't find a explicit remark on the usefulness of a measurement on non-significant tests.

Thanks.

-
Could you tell us what a "Phi measurement" is and what "MOA" might stand for? –  whuber May 1 '12 at 19:35
Sorry, didn't spell that out very well. I have always heared it as the Phi measurement of association. or p = (X^2/N)^(1/2). Where X is the Chi-Square return and N is observations. The square-root of Chi-squared divided by observations. –  RomaH May 1 '12 at 19:40