Art of Data

3.2 Describing Relationships

So far we’ve learned how to describe one variable at a time. Of course, we can compare these descriptions against each other, but we will learn some formalized methods to directly describe these relationships.

Bivariate Analysis

How do we find and interpret the relationship between two variables? What information can we gain, and what information can’t we gain?

⊕ side-by-side charts
⊕ association
⊕ contingency tables, basic probability, conditional proportions
⊕ scatterplot, Anscombe’s Quartet
▲: chi-square, goodness-of-fit, independence testing

Linear Regression

Once we have an understanding of how two variables are related, how do we use this information to make educated predictions?

⊕ correlation
⊕ regression line, line equation
⊕ residuals, least squares