O'Reilly logo

Statistical Inference: A Short Course by Michael J. Panik

Stay ahead with the world's most comprehensive technology and business learning platform.

With Safari, you learn the way you learn best. Get unlimited access to videos, live online training, learning paths, books, tutorials, and more.

Start Free Trial

No credit card required

13.7 A Measure of Strength of Association in a Contingency Table

We noted in Section 13.5 that the Pearson chi-square statistic is used to test for association between two attributes A and B. To conduct the test, we hypothesize independence among the A and B factors and then, for a given level of significance α, we see if we can reject the null hypothesis of independence in favor of the alternative hypothesis that the A and B characteristics are not independent but, instead, exhibit statistical association. If we detect statistical association between A and B, then the next logical step in our analysis of these attributes is to assess how strong the association between the A and B categories happens to be. That is, we need a measure of the strength of association between factors A and B.

One device for depicting statistical association in an r × c contingency table is Cramer's phi-squared statistic

(13.19) equation

obtained by dividing Pearson's chi-square statistic U by its theoretical maximum value n (q − 1), where n is the sample size and q = min{r,c}. Here Φ2 measures the overall strength of association, that is, for Φ2 = 0, the sample exhibits complete independence between the A and B attributes; and when Φ2 = 1, there exists complete dependence between these factors.

Example 13.7

Given the chi-square test results determined in Example 13.5 (we had a 4 × 3 contingency table involving ...

With Safari, you learn the way you learn best. Get unlimited access to videos, live online training, learning paths, books, interactive tutorials, and more.

Start Free Trial

No credit card required