Includes index.

Bibliography: p. 433-457.

1. Vectors of random variables -- 2. Multivariate normal distribution -- 3. Linear regression: estimation and distribution theory -- 4. Linear regression: hypothesis testing -- 5. Confidence intervals and regions -- 6. Departures from underlying assumptions -- 7. Straight-line regression -- 8. Polynomial regression -- 9. Analysis of variance -- 10. Analysis of covariance and missing observations -- 11. Computational techniques for fitting a specified regression -- 12. Choosing the "best" regression.

This book explains the theory and application of research techniques used in linar regression analysis. Dr. Seber gives a full discussion on the assumptions underlying regression models, and presents a variety of graphic and computational techniques for investigating these assumptions. His geometric approach enables the reader to deal with full rank and less than full rank models at the same time, and he varies the material by using the theory of generalized inverses to explain other approaches. Seber gives special attention to cases in straight-line and polynomial regression, analysis of variance and co-variance models associated with experimental designs in a theoretical framework. He also includes a number of topics which are usually omitted from most books, but are important in this area: optimal design, ridge estimators, two-phase regression, spline functions, and missing observations-with up-to-date sources of the literature available in these areas. Over 200 carefully selected problems, outline solutions, a reference bibliography, and appendices make this one of the most useful and informative books available on the subject of linear regression.