Author: Douglas A. Luke
Series: Quantitative Applications in the Social Sciences (Book 143)
Paperback: 88 pages
Publisher: SAGE Publications, Inc; 1 edition (July 8, 2004)
Product Dimensions: 0.2 x 5.5 x 8.5 inches
Taking a practical, hands-on approach to multilevel modeling, this book provides readers with an accessible and concise introduction to HLM and how to use the technique to build models for hierarchical and longitudinal data. Each section of the book answers a basic question about multilevel modeling, such as, "How do you determine how well the model fits the data?" After reading this book, readers will understand research design issues associated with multilevel models, be able to accurately interpret the results of multilevel analyses, and build simple cross-sectional and longitudinal multilevel models.
About the Author
Douglas A. Luke is currently a Professor at Washington University, George Warren Brown School of Social Work. He directs the Center for Tobacco Policy Research, has led the doctoral program at St. Louis University School of Public Health, and has long served on the key community health behavior study decision at the National Institute of Health.
In 1990 he received his Ph.D. in clinical/community psychology with a minor in quantitative psychology from the University of Illinois. While at the University of Illinois, he studied under a number or notable quantitative scientists and authors, including Phipps Arabie (co-author of Three-way Scaling & Clustering), Stanley Wasserman (editor of Advances in Social Network Analysis), Larry Jones, Larry Hubert, and Ledyard Tucker. His 1991 article, Expanding Behavior Setting Theory: Setting Phenotypes in a Mutual Help Organization, was recently selected as one of the ten most influential methodology articles published in the first 25 years of the American Journal of Psychology.
2.1 By A. Tierman on October 15, 2009
Sage university papers on quantitative applications are presented as brief and inexpensive treatments of specialized topics in statistics and data analysis. Some are well worth the price, while some leave you wishing you used the money towards acquiring a full-length treatise or textbook. If you need to learn about multilevel modeling on your own, Douglas Luke's Multilevel Modeling is worth much more than its price, especially if you buy it from Amazon.com, because it is a model of compositional economy in addressing a complex idea, and of what a truly introductory textbook should be. Luke maintains focus, precision, and masterful clarity in a fashion that is rarely encountered among books which claim to be "An Introduction to ... " a topic as specialized, intricate, and novel as is multilevel statistical modeling. Luke defines the terms more lucidly than some of the most popular full-sized books which aim to introduce multilevel analysis (and which still leave the reader mired in ambiguity). The author does not attempt to impart any gratuitous complexity to his exposition and manages to integrate textual clarity with statistical notation and equations, figures, and tables which are equally clear for someone who, while familiar with concepts beyond one-variable statistics and simple linear regression and ANOVA, has never studied or engaged in this type of data analysis or research design before. You may need to proceed to thicker treatises to make a thorough analysis and find out how to use your favorite software, but if you begin with one or more of those and find the topic still unclear in its elements - either the big picture or the basic details - you will find Luke's 78 pages (including reference to data online) enlightening.
2.2 By not a natural on July 26, 2009
Given the brevity imposed by Sage's little, green paperbacks, Luke's book is remarkably informative. This is especially true of models with more than three levels. Though Luke devotes only four pages to these more complex models, his examples are among the best I've seen. They lend credibility to the oft-repeated, but sometimes hard to see judgment that three-level models are almost as easy to specify, estimate, and understand as two-level models.
Though most analysts are primarily interested in fixed effects, which makes perfect sense, interpretation of random effects can be instructive. Thanks to Luke's examples, random effects make a good deal more sense after reading his book. It's to his credit that he accomplishes this in an informal, almost off-handed way. Luke is not the sort of author who takes himself or his subject too seriously -- nothing deadeningly grave here -- and his presentation benefits from his relaxed approach.
Even if a reader is benefiting from one of the much longer texts that are currently available, this inexpensive little book is worth buying and looking over. Luke has the gift of succinctness -- saying a great deal with unusual clarity while using remarkably few words -- making his book a pleasure to read. If you've been away from multilevel analysis for awhile and are just getting back into, Luke's presentation will provide a fine review.
3. Buying Link: