Book Review: Logistic Regression – A Self-Learning Text

By David G. Kleinbaum and Mitchel Klein (3rd edition). Springer, 2010.
I picked up this book with the intention of refreshing my knowledge of logistic regression as I worked on an article regarding the use of predictive analytics in the human resources field.

The book describes itself as a “self-learning” textbook, which appealed to me since I wanted to read up some material by myself rather than take a class. The formula worked for me. First, the text itself is very methodical and explicit, not requiring you to fill in any gaps by triangulating with other references, like so many textbooks do. Second, the whole book is formatted in two columns, one of which is the text and the other explanatory notes. It’s almost as if there’s a tutor or a graduate assistant boiling down the essence of the text into even more bite size chunks. Third, The notations in the explanatory columns include notes that you might write down yourself – very simple, but structured in a way to help you remember and quickly revise the material before a test.

One of the things I look for in descriptions of logistic regressions is how the authors explain the logit, which is the computational foundation of a logistic regression model. It is a completely synthetic mathematical device of convenience, but you somehow have to get your head around its existence and interpretation. I thought the authors of this book did a good job, though they came at it from a very functional standpoint, which requires some comfort with statistics. They also explained the mechanics underlying interaction variables and the interpretation of their coefficients quite well.

The book is part of a series on Statistics for Biology and Health and uses only medical and epidemiological examples. This made for interesting reading, since the examples are quite intuitive. Despite the simplicity of the exposition, the book is two inches thick and deals with some heavy-duty statistics. You need at least an advanced undergraduate course in statistics and econometrics to make the most of it. Among other things, the book covers modeling strategies, goodness of fit, polytomous logistic regression, ordinal logistic regression and logistic regression for special cases, such as matched data or correlated data.

The book does a reasonable job of explaining maximum likelihood estimation and provides computer programming examples in SAS, Stata and SPSS. Since each program is written to exploit the same data and run the same models, one can get a feel for the other packages if you are familiar with one. I remain partial to Stata.

My biggest take-away was that the book’s pedagogical approach might be appropriate for introducing logistic regression to human resource professionals interested in leveraging the power of logistic regressions for predictive human capital analytics. That’s the book I’ll start on after finishing the article.

Note: This review was cross-posted to My Reading List by Amazon on my Linked In profile. It’s the first statistics book I’ve reviewed.


About thenelsontouch

Amit is the founder and managing director of Nelson Touch Consulting, a management consultancy that specializes in maximizing the return on human capital through strategy, incentives and analytics.
This entry was posted in Analytics and tagged , , , , , , . Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s