Summary and Info
Logistic regression is a very handy statistical tool. It does not demand much of the data (no need for variables to be normally distributed). Also, it can be used with dichotomous dependent variables. Multiple regression is often used for such purposes, but--technically--that might provide misleading results. This is one of the classic books outlining the use (and abuse) of logistic regression analysis.
There are some features of logistic regression that are unfamiliar to those who have never used the technique before. This book describes the Wald statistic, pseudo r-square (designed to be something of an analogue to explained variation in multiple regression), goodness of fit measures (such as the eponymous Hosmer-Lemeshow test and chi square), logistic regression with more than two categories of the dependent variable, accuracy of predictions, and so on.
This is one of the best works on the subject, and it has helped me make sense of the results when I use statistical software featuring logistic regression. If interested in the technique, this work will be a nice resource.