On February 15, 2019, Eva Zijlmans will defend her thesis, titled:
Item-Score Reliability – Estimation and Evaluation
Measurement instruments, such as tests and questionnaires, for psychological attributes, such as intelligence and personality traits, must be reliable and valid. In classical test theory, reliability refers to the degree to which a measurement value is repeatable under precisely the same test administration conditions (Lord & Novick, 1968). Validity refers to the attribute the test measures and also to practical use of tests, such as prediction of educational and job success and suitability for therapy (Lissitz, 2009). In this project, we concentrate on psychometric issues and problems of reliability, and we provide solutions.
- The first project provides an overview of the different approaches to reliability and the varying methods for estimating the reliability coefficients that are associated with these approaches.
- The second project assesses estimating the reliability of single items.. The reliability of a single item may provide a more direct and interpretable contribution to the test-‐score reliability.
- In the third project, standard errors are obtained for a variety of reliability methods, such as the lambda coefficients, the MS-‐statistic, and stratified alpha (Cronbach, Schönemann, & McKie, 1965; Rae, 2007). Additionally, the bias of these estimation methods is investigated.
- The fourth project deals with reliability of nominal data, by focusing on the degree to which nominal scores are replicable over repeated observations.
- The final project extends the framework developed in the previous project such that it can be applied to item scores having an ordinal level of measurement.
Prof. Dr. K. Sijtsma, Dr. J. Tijmstra & Dr. L.A. van der Ark