Raymond L. Horton. The General Linear Model, Data Analysis in the Social and Behavioral Sciences. New York: McGraw-Hill International Book Company. 1978. pp.xi + 274.

It is a parsimonious book for the massive material it contains. Written for advanced students of social and behavioural sciences, it presents the analytic uses, limitations, and assumptions underlying the application of five techniques: factorial analysis of variance designs, latin square designs, repeated measures designs, analysis of covariance, and general regression analysis. The slimness of the book is made possible partly by the parsimony of style – a parsimony which did not impair the clarity of exposition – but mainly by the unifying factor of the general linear model (GLM)presented in the first two chapters. In this respect, the work reviewed resembles a 1978 book on demographic technique of analysis by Guillaume J. Wunsch and Marc G. Termote (plenum Press). There, too, in introductory chapters the common elements of cohort analysis and period analysis were first presented in a general manner and then applied to the four fields of mortality, nuptiality, natality, and migration. These examples of generalizability are recommended to expositors of intricacies in our profession. Such methodological generalizations are a condition of real interdisciplinary exchanges, of which this journal is a notable example. There is something genuine about that kind of generalizations in comparison with the more desperate attempts to generalize about the society so beloved of recent Ph.Ds.