Customer Reviews

Linear Probability, Logit, and Probit Models (Quantitative Applications in the Social Sciences)

5 star:		(2)
4 star:		(0)
3 star:		(1)
2 star:		(0)
1 star:		(0)

 

 

I think that it is a really good monograph about the logit and the probit models. It is very accessible at the appropriate level. I made a great deal use of it. Although it has a introductory review section about the linear regression model, having a good understanding of it, and also of statistics, is necessary in order to understand the rest of the book well.

A short mathematical tech detailing mathematical techniques that can be used when ordinary regression analysis is not appropriate, valid, or will just not work. The examples given are generally looking at a social science point of view, but the explanation and description is clear and is easily followed for application in other areas.

This is a short book on modeling probabilities using linear and generalized linear models. It walks the conceptual path from least-squares linear regression, through the linear probability model, to logistic and probit regression. This book is not for the statistical novice: A working knowledge of linear models will be necessary to take advantage of this text (knowing something about classification models like logistic regression or discriminant analysis would help, too). The great contribution of this book is that it ties these modeling methods together and answers a number of "why?" questions that anyone with an imagination would ask, after using linear classification models.

Covered include: linear probability model (Goldberger's procedure), basic generalized linear models (notably logistic and probit regressions, though alternative transfer functions are touched upon), both dichotomous and polytomous models and important practical issues.

Note that this book has a copyright date of 1984, hence some of the guidance is dated, being based on the assumption of much weaker computing resources than are cheaply available today.

The bottom line: If you have a passing familiarity with linear classification models and would like some background in a well-written and brief format, this is an excellent choice.