Generalizability Theory of Evaluation and Its Applications

Author: Yang Zhiming, Zhang Lei
Publisher:
Publish Date: 2003-05-01
Features: Whether it is diagnostic tests in the field of education, talent selection in enterprises and institutions, employee performance evaluation, performance-based assessment, or target-referenced tests, non-standard testing activities are becoming increasingly common. Classical test theory and item response theory can no longer meet these demands. As a result, a generalizability theory, which is based on classical test theory and incorporates randomization techniques, balancing or matching techniques, and statistical adjustment techniques to control measurement error, has undergone rapid development in the recent decade. Generalizability theory is a measurement theory that treats error variables as model parameters. Its significant feature is that for the same set of measurement data, it can provide multiple different measurement error estimation indicators based on the different inference ranges or purposes of the measurement results, making decision-making more appropriate and reasonable. Additionally, compared to other measurement theories, generalizability theory requires weaker prerequisites. This book is the first specialized work in China dedicated to the study of generalizability theory and its applications. It covers fundamental issues in measurement, the basic principles and methods of generalizability theory, and application examples. It holds guiding value for theoretical and applied work in psychological assessment, educational testing, personnel selection, employee performance evaluation, and social survey statistics.