Designing Experiments and Analyzing Data: A Model Comparison Perspective

Designed to serve as either a graduate-level text or reference book for experimental design, this book should be of use in psychology, education and statistics departments. The book promotes a model comparison perspective rather than the more traditional variance partitioning approach. This approach offers several advantages. Firstly, it better prepares students to understand the logic behind choosing the best technique for a particular design. Secondly, focusing on definitional versus computational formulae allows for a heavier reliance on statistical packages to perform the actual computations, freeing the reader to spend more time on the conceptual understanding versus the number crunching. Finally, the model comparison approach allows for the introduction of more complex topics, generally omitted from other books.