I have a little question and need some clarification on chi square's strength of association (using the Phi & Cramer's V). Say I have 4 parameters, all of them show significant relationship with the e.g., Quality i.e., p
I tend to use logistic regression over chi^2 tests. Chi^2 tells you if there is an association. Logistic regression tells you how that association works. It also tells you which associations are significantly different. They also let you add in extra information fairly easily.
After performing the Chi square test of independence that results into a positive association, we generally assume that the high chi square values or very low P values signifies a strong association. This is not always the case.
That is where the Phi and Cramers V come in. Phi is generally suitable for 2 by 2 tables and cramers V for 2 by 2 tables and larger tables. A cramers V value of O = No relationship, 0.2 or less = Weak relationship, from 0.21 to 0.3 = moderate, and above 0.3 = Strong relationship. Take note that is applicable only for Nominal data.
In SPSS, there is an option to estimate risk (odd ratios). This is als0 automatically generated in Epi Info but only limited to 2 by 2 tables.
A logistic regression covers all these if you are looking for a causal relationship.
I hope this helps!!
after performing chi square test, we should determine pvalue and then 2x2 contingency tables proves sensitivity and specificity
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS