Nominal Level Data Analysis

What are the appropriate measures of association for this situation?

1. Cramer's V 2. Phi 3. Yule's Q 4. Lambda

What are the assumptions made in Chi-Square?

1. Sample size needs to be greater than 50 2. Each of my expected cell frequencies should be greater than 5 3. Independent random samples 4. Used mostly nominal and ordinal level variables

Lambda

B is the number of errors made before taking into account the independent variable. A is the number of errors made after taking into account the independent variable. Gives you the lowest result. Good for any size table. This is a precise interpreration. Makes a difference which are the independent and dependent variables. You can have a false zero.

What is an appropiate test of significance or relating two nominal level variables?

Chi-Square Test

Cramer's V

Good for any size table. A measure of association between two nominal variables. Based on Pearson's chi-squared statistic. No precise interpretation of results. The results you get reflects the value you get for the Pearson's R

Yule's Q

Only good for a 2x2 box. No precise interpretation. Derived from gamma. Gives you a higer result.

Phi

Phi coefficient is used as measure o association if at least one of its variables in a particular contingency table is nominal and table is 2x2. only good for 2x2 tables No precise interpretation of results Results resembles what you would get from Pearson R Range of scores is 0-1; The closer phi is to 1, the greater the relationship is between the variables Phi is symmetrical statisitic, meaning that its value does not depend on which variable is the independent variable and which is the dependent variable.