•••• NEVV BOOKS IN REVIEVV

EDITOR: Robert A. Peterson

ASSOCIATE EDITOR: Roger A. Kerin

ADVANCES IN FACTOR ANALYSIS AND vanced and less relevant (e.g., Duncan, An Introduc-
STRUCTURAL EQUATION MODELS, Karl G. tion to Structural Equation Models, New York:
Joreskog and Dag Sorbom (Jay Magidson, editor). Academic Press, 1975; Heise, Causal Analysis, New
Cambridge, MA: Abt Books, 1979,242 pp. York: John Wiley and Sons, Inc., 1975), or somewhat
too specialized (e.g., Goldberger and Duncan, Struc-
New statistical and research procedures continually tural Equation Models in the Social Sciences, New
surface, although few have the scope and power of York: Seminar Press, 1973) to be of much use to
the latest methodological development: structural marketers.
equation models. Actually, structural equation models The primary audience for the book is marketing
have existed in well developed forms for at least two researchers. The techniques are particularly relevant
decades. Until recently they have evolved indepen- for problems in (1) scaling and measurement, where
dently in the subfields of statistics, econometrics, they can be used to determine internal consistency
psychometrics, and measurement and as a conse- and test-retest reliability while explicitly allowing for
quence many different terms describe essentially the measurement errors; (2) construct validation, where
same phenomenon. Among these are path analysis, they provide more powerful means to assess conver-
simultaneous equation systems, causal modeling, and gent, discriminant, and predictive validity than are
confirmatory factor analysis. currently available; (3) causal and predictive modeling
It was perhaps inevitable that a synthesis of the of phenomena as systems of relationships and hypoth-
diverse models would arise, and Professors Joreskog eses; (4) quasi-experimental designs such as the none-
and Sorbom provide a concise compilation of nine quivalent control group problem; and (5) comparisons
key articles spanning the past decade. But the book of phenomena across population groups. Many com-
represents much more than a synthesis, and the mon specialized procedures can be seen to be select
procedures they present extend earlier work into cases or less elegant forms of structural equation
structural equation models in a number of very impor- models, and thus the approach has the appeal of
tant directions. For example, structural equation generality and wide applicability to problems con-
models now lend themselves to the investigation of fronting marketers (e.g., the procedures accomplish
experimental as well as survey research. They are what two-stage least squares, full information maxi-
equally applicable to cross-sectional, time-series, or mum likelihood, factor analysis, and analysis of co-
longitudinal designs. They accomodate recursive rela- variance methods do).
tionships as well as reciprocal causation. Perhaps most To fully comprehend the book, the reader should
important, they permit the separation of errors in have knowledge of statistics and regression at an
variables from errors in equations, and thus provide intermediate level. The book would make a good
a means for representing unobservable or latent varia- addition to the advanced marketing models and multi-
bles. All of these features are accompanied by explicit variate statistics class at the Master's and advanced
test statistics, and a computerized maximum likelihood graduate levels. It is recommended that chapters 1,
estimation procedure-LISREL IV-is readily 4, 5, and 2 be read in sequence before the more
available in an interactive computer mode. specialized parts of the remainder of the book.
Advances in Factor Analysis and Structural Equa- Marketers have lagged behind other social scientists
tion Models is now the only text to bring together in the use of structural equation models. One reason
a high level and thorough exposition of the procedures is that the procedures appear at first to be both overly
in one source. Other texts in the area are either less complicated and specialized. That they are not can
focused and dated (e.g., Blalock, Causal Models in be seen in the many examples presented by Joreskog
the Social Sciences, Chicago: Aldine, 1971), less ad- and Sorbom. Indeed, from simple two-variable pre-
133

Journal of Marketing Research

Vol. XVII (February 1980), 133-9
13.4 JOURNAL OF MARKETING RESEARCH, FEBRUARY 1980

test / post-test models to more complicated causal models and the application of models to the study
interactions with correlated errors and simultaneity, of consumer behavior. After diagrammatic and brief
the illustrations show that the approach has wide narrative descriptions of the Engel-Kollat-Blackwell
applicability and power for developing and testing and Howard-Sheth models the authors suggest these
theory. are too complex for the beginning student. They then
Advances in Factor Analysis and Structural Equa- present a simplified framework for organizing an
tion Models provides marketers with a vehicle for introductory study of consumer behavior. This frame-
entering a new period of multivariate research in the work, as might be expected, becomes the basis for
discipline. the organization of the text.
RICHARD P. BAGOZZI Before proceeding with the development of their
Massachusetts Institute of Technology framework, the authors insert some fundamental as-
pects of consumer behavior research methodology and
CONSUMER BEHAVIOR: CONCEPTS AND AP- a discussion of market segmentation. The bases for
PLICATIONS, David L. Loudon and Albert J. Della segmentation included in the text are demographic,
Bitta. New York: McGraw-Hill Book Company, geographic, socioeconomic, product usage, benefit,
1979, 545 pp.
life-style, and psychographic.
The authors have succeeded in producing a very The authors' organizational scheme purports to
readable introductory consumer behavior textbook. include the "variables or processes that have been
A wide variety of topics are discussed in terms readily identified as the most important general influencers
understood by students with little or no previous of consumer behavior." The text is organized into
experience in marketing or the social sciences from the following three major components.
which many of the concepts of the study of consumer External environmental influences of consumer behav-
behavior are drawn. The text is written in a pleasant, ior include cultural, subcultural (with special atten-
easy-to-read style. It does not seem to talk down to tion given to ethnic and age influences), social class,
the reader, nor does it bore one already familiar with reference group, family, and personal (opinion lead-
the subject matter. "Technical" terms (jargon of the ership, diffusion of innovation, etc.).
marketer and social scientist) are defined in layman's Individual determinants of consumer behavior include
language, but these terms are subsequently used with motivation, perception, learning, personality, atti-
minimal additional explanation in what appears to be tudes and attitude change (the latter as a function
a successful effort to help the reader make them part of persuasive communication).
of his or her vocabulary. Consumer decision processes include considerations of
problem recognition, information search and alterna-
A major strength of the book is the attention given tive evaluation, actual purchasing decisions, and
to specific applications of the many concepts intro- postpurchase behavior.
duced. Frequently actual or potential applications are
given not to just one, but to several or all elements This book will be used primarily as a text (for
of the marketing mix. The inclusion ofthe large number marketing students and for nonmarketing students
of applications both in the text and in the short cases taking a survey course in consumer behavior). It would
at the end of each section enhances the salience of also be a good selection for the marketing executive
the concepts presented and should effectively and who wants a very readable introduction to consumer
meaningfully implant the concepts in the minds of behavior.
readers. THOMAS E. NESS
Like most survey textbooks, it must cover a great University of South Florida
variety of concepts and consequently most topics are
presented in capsule form. The material in this text
is generally sufficient for meaningful comprehension MARKETING RESEARCH, TEXT AND CASES,
and discussion by beginning students of consumer K. L. McGown. Cambridge, MA: Winthrop Pub-
behavior, and it also allows instructors latitude to lishers, Inc., 1979, 432 pp.
supplement the text as they may choose. Footnoted A textbook should not be judged by its unique
references for further study are abundant and conve- contribution to a field or discipline without considering
niently placed at the ends of the chapters. the objectives of the author. McGown's overall objec-
The authors begin with a short but effective ex- tive is to present marketing research concepts in simple
planation of what the study of consumer behavior and understandable terms for students who have had
entails. They note that it is their intention to focus only one semester of marketing and no statistical
on the ultimate consumer who purchases goods and background. Thus this text is not in direct competition
services for personal or household use. Considerations with Green and Tull's outstanding but quantitative
of industrial or institutional decisions or purchasing Research for Marketing Decisions or with Churchill's
behavior are not included. excellent but less quantitative Marketing Research:
The reader is next introduced to the concept of Methodological Foundations. Given this objective, the