Statistical Models for Causal Analysis

This volume provides an overview of statistical models commonly used in the causal analyses of non-experimental data in the social and biomedical sciences. Models covered include simple bivariate regressions, multiple regression, multiple classification analysis, path analysis, logistic regression, multinomial logistic regression and survival models. The goal is to impart a working knowledge of these models with a minimum of mathematics. General principles and procedures of statistical inference and model testing are covered, but derivations of sampling distributions and standard errors are omitted, thereby circumventing the need for a great deal of complex mathematics. On the other hand, a fair amount of mathematical detail is retained in the discussions of model specification and interpretation. The book also contains an appendix of computer programs that are written for the major statistical packages, including SAS, BMDP, and LIMDEP.