I'm analysing a cross tab using Phi and I'm getting approx sig. of .005

No error is shown in SPSS. However I'm of the understanding that the testing of such a measure is done using Chi Squared. Since I have 3 cells with expected count less than 5, it seems that I cannot use Chi squared but not even Phi since the test of significance is based on Chi Squared. If this is the case, I wonder why SPSS does not flag this problem.

Can anything else be used to solve my problem.. I need non parametric measures.

Topic