Examining Relationships Between Two Categorical Variables

How can we examine if a relationship exists between two categorical variables? We can consider whether the distribution of cases among the values of one of the variable differs depending on the value of the other variable. For example, suppose the two categorical variables are handedness and sex. We could ask: Are the proportions of people who are left-handed different for males and females?

Stream the video without the embedded quiz questions by clicking on the video link below. Closed captions are available.

Notes on the video: Exploring Relationships Between Two Categorical Variables

Accessibility note: Plots at 2:31, 4:47, 10:49 and 11:23 in the video rely on colour to differentiate among areas. See the notes for versions of the plots that use shading to differentiate among the curves.

A point to consider for this video:

At the end of this variable, the effect of a lurking variable is revealed. When examining the relationship between two variables, a lurking variable is a third variable that hasn’t been considered but has an important effect on the relationship. The specific manner in which it has an effect can vary with the circumstance. When examining the relationship between two categorical variables, the combination of a lurking variable and unequal group sizes can lead to contradictory conclusions. This was seen in the smoking example in the video. To avoid this, the effect of lurking variables can be eliminated by holding them constant (for example, if gender is a lurking variable, only study females) or making them part of the study.