Page 170 - Intermediate Statistics for Dummies

13_045206 ch08.qxd 2/1/07 9:57 AM Page 149
Chapter 8
Yes, No, Maybe So: Making
Predictions by Using Logistic
Regression
In This Chapter
Knowing when logistic regression is appropriate
Building logistic regression models for yes or no data
Checking model conditions and making the right conclusions
veryone (even yours truly) tries to make predictions about whether or
Enot a certain event is going to happen. For example, what’s the chance
it’s going to rain this weekend? What is our team’s chances of winning our
next game? What is the chance that I’ll have complications during this
surgery? These predictions are often based on probability, the long-term per-
centage of time an event is expected to happen. In the end, you want to esti-
mate p, the probability of an event occurring. In this chapter, you see how to
build and test models for p based on a set of explanatory (x) variables. This
technique is called logistic regression.

Setting Up the Logistic Regression Model

Yes or no data that comes from a random sample has a binomial distribution
with probability of success (the event occurring) equal to p. In the binomial
problems you saw in intro stats, you had a sample of size n trials, you had
yes or no data, and you had a probability of success on each trial, denoted
by p. In your intro stat course, for any binomial problem the value of p was
somehow given to be a certain value, but in intermediate stats, you operate
under the much more realistic scenario that it’s not. In fact, because p isn’t
known, your job is to estimate what it is and use a model to do that.