An Optimal Two-Stage Procedure to Select the Best out of a Normal Population

  • Dieter RaschEmail author
  • Takuya Yanagida
Original Article


In selecting a candidate with the largest expectation \(\mu_{k}\) (the best one) from a huge number a of candidates or let us say populations \(P_{k}\) with expectations \(\mu_{k} t\), we can first select a smaller subset using Gupta’s procedure and then from this subset the best one by Bechhofer’s approach. A simulation experiment for normal distribution indicates when such a two-stage selection is optimal and preferred over using each procedure alone.


Subset selection Indifference zone selection Two-stage procedure Simulation 


  1. 1.
    Bechhofer RE (1954) A single sample multiple decision procedure for ranking means of normal populations with known variances. Ann Math Stat 25:16–39MathSciNetCrossRefGoogle Scholar
  2. 2.
    Gupta SS (1956) On a decision rule for a problem in ranking means, Mim. Ser., No 150, Univ. North CarolinaGoogle Scholar
  3. 3.
    Gupta SS, Panchapakesan S (2002) Multiple decision problems: theory and methodology of selection and ranking populations. SIAM, University CityCrossRefGoogle Scholar
  4. 4.
    Miescke KJ, Rasch D (1996a) Special issue on 40 Years of statistical selection theory, Part I. J Stat Plan Inference 54(2)Google Scholar
  5. 5.
    Miescke KJ, Rasch D (1996b) Special Issue on 40 years of statistical selection theory, Part II. J Stat Plan Inference 54(3)Google Scholar
  6. 6.
    Rasch D, Schott D (2018) Mathematical statistics. Wiley, Oxford. ISBN 9781119385288CrossRefGoogle Scholar

Copyright information

© Grace Scientific Publishing 2018

Authors and Affiliations

  1. 1.Institute of Applied Statistics and ComputingUniversity of Natural Resources and Life Sciences, ViennaViennaAustria
  2. 2.Department of Applied Psychology: Work, Education, and Economy, Faculty of PsychologyUniversity of ViennaViennaAustria

Personalised recommendations