Page 90 - Intermediate Statistics for Dummies

09_045206 ch04.qxd 2/1/07 9:49 AM Page 69
Chapter 4
Getting in Line with
Simple Linear Regression
In This Chapter
Using scatterplots and correlation coefficients to examine relationships
Building a simple linear regression model to estimate y from x
Testing how well the model fits
Interpreting the results and making good predictions
ooking for relationships and making predictions is one of the staples of
Ldata analysis. Everyone wants to answer questions like “Can I predict
how many units I’ll sell if I spend x amount of advertising dollars?”; or “Does
drinking more diet cola really relate to more weight gain?”; or “Do children’s
backpacks seem to be getting heavier each year in school, or is it just me?”
Linear regression tries to find relationships between two or more variables
and comes up with a model that tries to describe that relationship, much like
the way the line y = 2x + 3 explains the relationship between x and y. But
unlike math where functions like y = 2x + 3 tell the entire story about the two
variables, in statistics, things don’t come out that perfectly; some variability
and error is involved (that’s what makes it fun!).

This chapter is partly a review of the concepts of simple linear regression
presented in an intro stats book. But the fun doesn’t stop there. I expand on
the ideas you learned about regression in your intro stat course and set you
up for some of the other types of regression models you see in Chapters 5
through 8.

In this chapter, you see how to build a simple linear regression model that
examines the relationship between two variables. You also see how simple
linear regression works from a model-building standpoint.