Practical Multivariate Analysis
 This version cannot be shipped to your selected country.
Preview
Book Description
This new version of the bestselling ComputerAided Multivariate Analysis has been appropriately renamed to better characterize the nature of the book. Taking into account novel multivariate analyses as well as new options for many standard methods, Practical Multivariate Analysis, Fifth Edition shows readers how to perform multivariate statistical analyses and understand the results. For each of the techniques presented in this edition, the authors use the most recent software versions available and discuss the most modern ways of performing the analysis.
New to the Fifth Edition
 Chapter on regression of correlated outcomes resulting from clustered or longitudinal samples
 Reorganization of the chapter on data analysis preparation to reflect current software packages
 Use of R statistical software
 Updated and reorganized references and summary tables
 Additional endofchapter problems and data sets
The first part of the book provides examples of studies requiring multivariate analysis techniques; discusses characterizing data for analysis, computer programs, data entry, data management, data cleanup, missing values, and transformations; and presents a rough guide to assist in choosing the appropriate multivariate analysis. The second part examines outliers and diagnostics in simple linear regression and looks at how multiple linear regression is employed in practice and as a foundation for understanding a variety of concepts. The final part deals with the core of multivariate analysis, covering canonical correlation, discriminant, logistic regression, survival, principal components, factor, cluster, and loglinear analyses.
While the text focuses on the use of R, SPLUS, SAS, SPSS, Stata, and STATISTICA, other software packages can also be used since the output of most standard statistical programs is explained. Data sets and code are available for download from the book’s web page and CRC Press Online.
Table of Contents
PREPARATION FOR ANALYSIS
What Is Multivariate Analysis?
Defining multivariate analysis
Examples of multivariate analyses
Multivariate analyses discussed in this book
Organization and content of the book
Characterizing Data for Analysis
Variables: their definition, classification, and use
Defining statistical variables
Stevens’s classification of variables
How variables are used in data analysis
Examples of classifying variables
Other characteristics of data
Preparing for Data Analysis
Processing data so they can be analyzed
Choice of a statistical package
Techniques for data entry
Organizing the data
Example: depression study
Data Screening and Transformations
Transformations, assessing normality and independence
Common transformations
Selecting appropriate transformations
Assessing independence
Selecting Appropriate Analyses
Which analyses to perform?
Why selection is often difficult
Appropriate statistical measures
Selecting appropriate multivariate analyses
APPLIED REGRESSSION ANALYSIS
Simple Regression and Correlation
Chapter outline
When are regression and correlation used?
Data example
Regression methods: fixedX case
Regression and correlation: variableX case
Interpretation: fixedX case
Interpretation: variableX case
Other available computer output
Robustness and transformations for regression
Other types of regression
Special applications of regression
Discussion of computer programs
What to watch out for
Multiple Regression and Correlation
Chapter outline
When are regression and correlation used?
Data example
Regression methods: fixedX case
Regression and correlation: variableX case
Interpretation: fixedX case
Interpretation: variableX case
Regression diagnostics and transformations
Other options in computer programs
Discussion of computer programs
What to watch out for
Variable Selection in Regression
Chapter outline
When are variable selection methods used?
Data example
Criteria for variable selection
A general F test
Stepwise regression
Subset regression
Discussion of computer programs
Discussion of strategies
What to watch out for
Special Regression Topics
Chapter outline
Missing values in regression analysis
Dummy variables
Constraints on parameters
Regression analysis with multicollinearity
Ridge regression
MULTIVARIATE ANALYSIS
Canonical Correlation Analysis
Chapter outline
When is canonical correlation analysis used?
Data example
Basic concepts of canonical correlation
Other topics in canonical correlation
Discussion of computer program
What to watch out for
Discriminant Analysis
Chapter outline
When is discriminant analysis used?
Data example
Basic concepts of classification
Theoretical background
Interpretation
Adjusting the dividing point
How good is the discrimination?
Testing variable contributions
Variable selection
Discussion of computer programs
What to watch out for
Logistic Regression
Chapter outline
When is logistic regression used?
Data example
Basic concepts of logistic regression
Interpretation: Categorical variables
Interpretation: Continuous variables
Interpretation: Interactions
Refining and evaluating logistic regression
Nominal and ordinal logistic regression
Applications of logistic regression
Poisson regression
Discussion of computer programs
What to watch out for
Regression Analysis with Survival Data
Chapter outline
When is survival analysis used?
Data examples
Survival functions
Common survival distributions
Comparing survival among groups
The loglinear regression model
The Cox regression model
Comparing regression models
Discussion of computer programs
What to watch out for
Principal Components Analysis
Chapter outline
When is principal components analysis used?
Data example
Basic concepts
Interpretation
Other uses
Discussion of computer programs
What to watch out for
Factor Analysis
Chapter outline
When is factor analysis used?
Data example
Basic concepts
Initial extraction: principal components
Initial extraction: iterated components
Factor rotations
Assigning factor scores
Application of factor analysis
Discussion of computer programs
What to watch out for
Cluster Analysis
Chapter outline
When is cluster analysis used?
Data example
Basic concepts: initial analysis
Analytical clustering techniques
Cluster analysis for financial data set
Discussion of computer programs
What to watch out for
LogLinear Analysis
Chapter outline
When is loglinear analysis used?
Data example
Notation and sample considerations
Tests and models for twoway tables
Example of a twoway table
Models for multiway tables
Exploratory model building
Assessing specific models
Sample size issues
The logit model
Discussion of computer programs
What to watch out for
Correlated Outcomes Regression
Chapter outline
When is correlated outcomes regression used?
Data example
Basic concepts
Regression of clustered data
Regression of longitudinal data
Other analyses of correlated outcomes
Discussion of computer programs
What to watch out for
Appendix
References
Index
A Summary and Problems appear at the end of each chapter.
Reviews
"I have come to know this book, and its several precursors, very well indeed. I have used it – in its various editions – for almost thirty years, as the principal text in my doctorallevel statistics course, S052: Applied Data Analysis, at the Harvard University Graduate School of Education. My course served around 6575 doctoral students annually, drawn from the socialscience disciplines around Harvard and MIT. Over those years, students told me repeatedly what an excellent book they found it to be, as a resource both during and after their successful completion of the course. In my view, the book’s main strength is embodied in its title – "Practical Multivariate Analysis" – with a strong emphasis on the "practical."
In the book, a thoughtful set of powerful multivariate methods are presented and described. Explanations in the text are built upon realworld dataexamples and their accompanying research questions. There is strong emphasis on perceptive datadisplay and good dataanalysis, on the testing of assumptions and on the credible interpretation of results. These very things are modelled deeply and repeatedly in the book itself, so that it is an exemplar for the rigorous conduct and clear reporting of sophisticated quantitative analyses. The writing is clear and understandable, while remaining technically stringent. It is in these important roles – as informant, explainer and model – that the book has always shone, in my view. Consequently, it has been a critical companion to many cohorts of new scholars and been instrumental in insuring the quality and rigor of the work that they then went on to do. Who knows how far its influence stretches? Overall, I do not think that there many other books at this level – maybe none – that have these same qualities. I recommend it very highly indeed."
—John B. Willett, Charles William Eliot Research Professor, Harvard University Graduate School of Education"First of all, it is very easy to read. … The authors manage to introduce and (at least partially) explain even quite complex concepts, e.g. eigenvalues, in an easy and pedagogical way that I suppose is attractive to readers without deeper statistical knowledge. The text is also sprinkled with references for those who want to probe deeper into a certain topic. Secondly, I personally find the book’s emphasis on practical data handling very appealing. … Thirdly, the book gives very nice coverage of regression analysis. … this is a nicely written book that gives a good overview of a large number of multivariate techniques and also tries to give the reader practical advice on how to perform statistical analyses."
—Australian & New Zealand Journal of Statistics, 56(4), 2014"I found the text enjoyable and easy to read. The authors provide a sufficient description of all the methodology for practical use. Each chapter includes at least one real world dataset analysis and the software commands summary tables included at the end of every chapter should be particularly helpful to a practitioner of statistics. … I would recommend the text for practitioners of statistics looking for a handy reference, particularly those performing basic analysis in the health sciences."
—Thomas J. Fisher, Journal of Biopharmaceutical Statistics, Issue 6, 2012Praise for Previous Editions:
For the past 20 years, whenever I had an occasion to review a multivariate method…this was the book that I grabbed first. These books kept the mathematical content to the minimally necessary material and used a wealth of nice examples. One of its attractions is that it is a practical text that works well with nonstatisticians who have had a decent statistics course. It also continues to be an excellent book for the statistician's bookshelf.
—Technometrics, November 2004
This book is an excellent presentation of computeraided multivariate analysis. I believe that it will be a very useful addition to any scholarly library … it provides a comprehensive introduction to available techniques for analyzing data of this form, written in a style that should appeal to nonspecialists as well as to statisticians.
—Zentralblatt MATH 105
This is a text for a broad spectrum of researchers … who may find it very useful as it stresses the importance of understanding the concepts and methods through useful real life illustrations.
—Journal of the RSS, Vol. 168, 2005
A key feature of this book is that it can be used in conjunction with any or all of the following very wellknown software tools: SPlus, SAS, SPSS, STATA, and STATISTICA.
—Pat Altham, University of Cambridge, UK, Statistics in Medicine, 2005
Support Material
Ancillaries

Data.zip
Additional materials

Instructor Resources
To gain access to the instructor resources for this title, please visit the Instructor Resources Download Hub.
You will be prompted to fill out a regist