Estimating psychometric reliability with one observation per subject (original) (raw)

Abstract

Test a sample of individuals with the same test on two occasions. Correlate the pairs of scores. The resulting r is called the reliability of the test. More than 100 years ago Charles Spearman proposed a random effects model for interpreting r a model which remains extensively used today. However in many settings obtaining pairs of scores from the same individuals is often expensive or cannot be achieved. A model in the Spearman spirit for estimating reliability based on one observation per subject is proposed.

Loading...

Loading Preview

Sorry, preview is currently unavailable. You can download the paper by clicking the button above.

References (5)

  1. Givens, G. H., & Hoeting, J. A. (2005). Computational statistics. New York: Wiley. Kelly, T. L., (1927). Interpretation of educational measurements. New York: World Book Company.
  2. Spearman, C. (1904). General intelligence objectively defined and measured. American Journal of Psychology, 5, 201-293.
  3. Spearman, C. (1910). Correlation calculated from faulty data. British Jour- nal of Psychology, 3, 271-295.
  4. Thomas, H., Lohaus, A., & Domsch, H. (2011a). Extensions of reliability theory. In Hunter, D., Rosenberger, J., & Richards, D. (Eds.) Advances in statis- tics: Nonparametric Statistics and Mixture Models: A Festschrift in Honor of Thomas P. Hettmansperger. New Jersey: World Scientific Publishing Co. (pp 309-316).
  5. Thomas, H., Lohaus, A., & Domsch, H. (2011b). Stable unstable reliability theory. British Journal of Mathematical and Statistical Psychology. Article first published online : 2 February 2011, DOI: 10.1111/j.2044-8317.2010.02011.x MMLV2012.T E X