Guide to Sullivan’s Chapter 4

Summary: Now we turn to two-variable (bivariate) numeric data. Correlation analysis answers the question how close do the data points stick to a straight line? and regression analysis answers the question which line fits the data points best? Historically the two were separate, but your calculator does them together.

Here’s the agenda for this chapter:

• scatter plot (a/k/a scattergram)
• correlation coefficient, r
• regression line and residuals
• coefficient of determination, R²
• predicted average, ŷ
• computed residual, y−ŷ
• inference using decision points

Handout: The handout Correlation and Regression replaces Chapter 4 of the textbook.

If there’s time in class. we’ll work through pages 205–206 problem 23 as a second example. Of course you’re free to delve further into the textbook if you want more explanation and illustrations. Please note, however, that the second paragraph of page 186 is wrong.

Announce: The Shoe-Size Lab is due in two weeks and will use material from Chapter 4 only. Start now with the data collection at least.