Experimental design : procedures for the behavioral sciences /
Roger E. Kirk.
- fourth edition.
- Thousand Oaks : Sage, (c)2013.
- xiii, 1056 pages : illustrations ; 27 cm
Includes bibliographies and index.
Research strategies and the control of nuisance variables -- Experimental designs : an overview -- Fundamental assumptions in analysis of variance -- Completely randomized design -- Multiple comparison tests -- Trend analysis -- General linear model approach to ANOVA -- Randomized block designs -- Completely randomized factorial design with two treatments -- Completely randomized factorial design with three or more treatments and randomized block factorial design -- Hierarchical designs -- Split-plot factorial design : design with group-treatment confounding -- Analysis of covariance -- Latin square and related designs -- Confounded factorial designs : designs with group-interaction confounding -- Fractional factorial designs : designs with treatment-interaction confounding -- Appendix A : rules of summation -- Appendix B : rules of expectation, variance, and covariance -- Appendix C : Orthagonal coefficients for unequal intervals and unequal ns -- Appendix D : matrix algebra -- Appendix E : tables -- Appendix F : answers to starred exercises.
"Roger E. Kirk shows how three simple experimental designs can be combined to form a variety of complex designs. He provides diagrams illustrating how subjects are assigned to treatments and treatment combinations. New terms are emphasized in boldface type, there are summaries of the advantages and disadvantages of each design, and real-life examples show how the designs are used"--