Hello, I have attached a permutational table. Intent of this exercise is to determine which statistical test should be used when you've a combination of different Independent (Explanatory) and Dependent (outcome) variables.
I would appreciate if you can help me fill this table in.
The variables' categorization as independent or dependent does not mean anything for testing. It only matters for modeling.
You are thinking about only tests with categorical DV?
As you mention, there is chi-square test of association. There are other tests used in similar circumstances, like G-test and Fisher's exact test.
For three-way tables, there is Cochran-Mantel-Haenszel.
For other tables and kinds of dependence, you might look in to log-linear models.
Traditional logistic regression is used when the DV is binary, but you can have all kinds of IV's.
Similarly, there is multinomial logistic regression for when the DV has more than 2 levels.
There are also tests for paired cases or repeated measures: McNemar, and Cochran's Q.
You might also specify that you mean _nominal_ categorical DV. There's a whole other set of tests and models for _ordinal_ categorical data.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA