Computer File
Applied multivariate statistics for the social sciences (Fifth Edition)
This edition has several major changes, and I would like to mention those first.
There are two new chapters (15 and 16) on two very important topics. Chapter 15 on
the Hierarchical Linear Model was written by Dr. Natasha Beretvas of the University of
Texas at Austin. This model deals with correlated observations, which occur very fre
quently in social science research. The general linear model assumes the observations are
INDEPENDENT, and even a small violation causes the actual alpha level to be several
times the nominal level. The other major topic, Structural Equation Modeling (Chapter 16),
was written by Dr. Leandre Fabrigar of Queen's University and Dr. Duane Wegener of
Purdue (both were former students of Dr. MacCallum). Among the strengths of this tech
nique, as they note, are the ability to account for measurement error and the ability to
simultaneously assess relations among many variables. It has been called by some the
most important advance in statistical methodology in 30 years. Although I have a concern
with equivalent models, SEM is an important technique one should be aware of.
Tidak tersedia versi lain