Psychometrika

, Volume 34, Issue 1, pp 85–110 | Cite as

Restricted ranking

  • James E. Norman
  • H. A. David
Article

Abstract

This paper is a study of certain aspects of restricted ranking, a method intended for use by a panel ofm judges evaluating the relative merits ofN subjects, candidates for scholarships, awards, etc. Each judge divides theN subjects intoR classes so thatni individuals receive a gradei (i = 1, 2, ...,R; Σn i =N) where theR numbersn i are close toN/R (n i =N/R whenN is divisible byR) and are preassigned and the same for all judges. This method is superior in several respects to other likely alternatives. Under the null hypothesis that allnR =N subjects are of equal merit, four tests of significance are developed. The effectiveness of the method is investigated both theoretically by means of the asymptotic relative efficiency and more generally by simulation studies. When the numbersn i are not restricted to values close to or equal toN/R but instead are given values conforming to a normally distributed pattern, the resulting method is known as theQ-sort, so designated by certain investigators in psychotherapy. The simulation studies reveal that restricted ranking is only slightly inferior to complete ranking and generally superior in the cases considered to theQ-sort, although there are likely to be other situations when the latter is superior.

Keywords

Null Hypothesis Simulation Study Public Policy Statistical Theory Relative Efficiency 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. David, H. A., & Norman, J. E. Exact distribution of the sum of independent identically distributed discrete random variables.Journal of American Statistical Association, 1965,60, 837.Google Scholar
  2. Doornbos, R., & Prins, H. J. On slippage tests.Indagtiones Mathematicae, 1958,20, 438.Google Scholar
  3. Friedman, M. The use of ranks to avoid the assumption of normality implicit in the analysis of variance.Journal of American Statistical Association, 1937,32, 675.Google Scholar
  4. Hannan, E. J. The asymptotic powers of certain tests based on multiple correlations.Journal of the Royal Statistical Society B, 1956, 18, 227.Google Scholar
  5. Harter, H. L. Expected values of normal order statistics.Biometrika, 1961,48, 151.Google Scholar
  6. Mowrer, O. H.Psychotherapy: Theory and Research. New York: Ronald Press Co., 1953.Google Scholar
  7. Noether, G., & Van Elteren, Ph. The asymptotic efficiency of the chi-square test for a balanced incomplete block design.Biometrika, 1959,46, 475.Google Scholar

Copyright information

© Psychometric Society 1969

Authors and Affiliations

  • James E. Norman
    • 1
  • H. A. David
    • 2
  1. 1.University of GeorgiaUSA
  2. 2.University of North CarolinaUSA

Personalised recommendations