Tests of Statistical Significance and Measures of Association

1. Formulate a hypothesis
2. Run appropriate stats
3. Answer whether or not the results support the hypothesis
1. Check statistical significance
2. Check strength and direction
3. Provide a substantive interpretation

4. Add some control variables to the model. Answer whether or not the original association remains the same.

Measurement Types	Statistics	Range	Strength	Significance level
Two nominal variables	Lambda /Cramer’s V	0 ~ +1	~0.1: Little to very weak 0.1~0.19: Weak 0.20~0.29: Moderate 0.30~: Strong	Prob<.05 means significant (95% confidence) Reject the null
One nominal one ordinal variables	Cramer’s V	0 ~ +1
Two ordinal variables	Kendall’s tau-b (Square crosstab) Kendall’s tau-c (Not square)	-1 ~ +1
Two I/R variablesR	Pearson’s corr. coeff and regression	-1 ~ +1 for corr. coeff	~0.25: Little to very weak 0.25~0.34: Weak 0.35~0.39: Moderate 0.40~: Strong (p244)	* : p<0.05 (95% confidence) **: p<0.01 (99% confidence) Reject the null
Dependent: I/R Independent: One or more I/R variables	Regression: R^2, slope, and beta coefficients	R^2 : 0 ~ +1