ABSTRACT
The preceding set of regression methods involved a (linear) relationship between a response variable Y and a regressor X, where Y was assumed to be continuous. However, when dealing with logistic regression, the response variable is taken to be dichotomous or binary (it takes on only two possible values), i.e., Yi = 0 or 1 for all i = 1,…, n. For instance, we can have a situation in which the outcome of some process of observation is either a success (we record a 1) or failure (we record a 0), or we observe the presence (1) or absence (0) of some characteristic or phenomenon. In addition, dichotomous variables are useful for making predictions, e.g. we may ask the following:
1. Will an individual make a major purchase of a particular item in the near future? Here
Yi =
⎧⎨⎩ 1 0 , , yes; no.
