Applied Discriminant Analysis
Most books on discriminant analysis focus on statistical theory. But properly applied, discriminant analysis methods can be enormously useful in the interpretation of data. This book is the first ever to offer a complete introduction to discriminant analysis that focuses on applications. It provides numerous examples, explained in great detail, using current statistical discriminant analysis algorithms. It also develops several themes that will be useful to researchers and students regardless of the analytical methods they employ. They are the careful examination of data prior to final analysis; the application of critical judgment and common sense to all analyses and interpretations; and conducting multiple analyses as a matter of routine. To encourage and enable readers to conduct multiple analyses of their data, the accompanying diskette contains the four complete data sets and five special computer programs that are referred to repeatedly in the text and are the subjects of numerous exercise problems. This enables the reader to carry out package analyses on the data sets using a variety of procedural options both within and across computer packages. The term "discriminant analysis" means different things to different people. For statisticians and researchers in the physical sciences, it usually denotes the process through which group membership is predicted on the basis of multiple predictor variables. Behavioral scientists, on the other hand, often use discriminant analysis to describe group differences across multiple response variables. Though closely related, predictive discriminant analysis (PDA) and descriptive discriminant analysis (DDA) are used for different purposes and should be approached in different ways. To accentuate these differences and distinguish clearly between the two, Applied Discriminant Analysis presents these topics separately. For graduate students, this book will expand your background in multivariate data analysis methods and facilitate both the reading and the conducting of applied empirical research. It will also be of great use to experienced researchers who wish to enhance or update their quantitative background, and to methodologists who want to learn more about the details of applied discriminant data analysis, and some still unresolved problems, as well.
What people are saying - Write a review
We haven't found any reviews in the usual places.
PART ONE INTRODUCTION
Discriminant Analysis in Research
PART TWO PREDICTION
20 other sections not shown
ACTN ANALYSIS FOR TALENTB assessment assigned best subset BMDP 7M Canon1cal canonical correlation categorical variable centroid Chapter classification results classification rule CNINT considered context contrast correlation covariance matrix CPSPO criterion CROUP CRP1 CRP2 CRP4 data set deleted denoted descriptive discriminant analysis distance ECPA effects eigenvalues EPROF error ETSA ETSC example F statistic FCPA FRTBK given group g group membership grouping variable hit rate hit-rate estimates Huberty Intarval linear composite LINFO LLINT MANOVA measures method NATRP NCPA obtained OISCRININANT Olscrlalnant outcome variables outliers output A2 parcant pertains population g posterior probability predictive discriminant analysis prior probabilities Probab1l1ty problem PSINT quadratic quadratic rule regression analysis researcher response variables sample SAS DISCRIM scores SCPA SINFO SLNF SNSF SOCBL SPSS Squarad stepwise Table TALENTB OATA tion total-group TRINT two-group unit univariate values variable ordering variable selection variance vector VN/ESA CNS5 VTOIN Wilks yields