Chapter Summary

Choosing the correct measure of association depends on the level of measurement of the two variables. The association between two interval-level variables can be measured with Pearson’s r. If you have two ordinal variables, you can measure their association using gamma (or tau). For nominal variables, you can use lambda (or Cramer’s V). Most measures of association are set up so they give a value of one for a perfect relationship and a value of zero for no relationship. But when you are describing the relationship between two variables that have order (either ordinal or interval), you have the ability to distinguish between positive and negative relationships. In that case, you would want your measure to find an association of –1 for a perfect negative relationship. Both Pearson’s r and gamma do that. If the two variables have different levels of measurement, you would generally use the measure of association for the lower-level variable. But if one variable is dichotomous, you generally use the measure of association appropriate for the other variable.

Learning Objectives

After reading this chapter, you should:

• Know how to choose the correct measure of association for a contingency table
• Know how to translate a level of association into a qualitative evaluation of the strength of the relationship
• Be able to calculate Pearson’s r
• Be able to calculate Gamma
• Be able to recognize a positive relationship in a contingency table as opposed to a negative relationship
• Be able to calculate Lambda
• Know how to interpret Lambda as a PRE measure
• Understand the limitations of Lambda
• Be able to produce a correlation matrix in SPSS