×

Optimal predictive selection. (English) Zbl 0938.62128

Mukherjee, S. P. (ed.) et al., Frontiers in probability and statistics. 2nd international triennial Calcutta symposium on Probability and Statistics, Calcutta, India, December 30, 1994 - January 2, 1995. London: Narosa Publishing House. 141-152 (1998).
From the introduction: The use of selection tests for a variety of jobs, college admissions, training programmes, etc. has become common place. There are two basic reasons for the widespread use of selection tests. First, the number of applicants is usually considerably larger than the number of positions or seats available, and a large degree of screening is called for. Secondly, it is generally believed that the academic results or work experiences of the applicants are neither easily comparable nor adequately informative because of varying standards of assessment and having a different focus.
A selection test is used apparently to serve two purposes. First, it is used for considerable screening. Secondly, such a screening is being supported apparently by the belief that the score in a selection test is a good predictor for future performance in relation to the job or the academic programme, or the training programme etc. under consideration.
In this paper we shall try to examine the validity of this general belief. With what degree of confidence can we assert that a high-scorer in a selection test would be a high performer?
For the entire collection see [Zbl 0912.00017].

MSC:

62P15 Applications of statistics to psychology
PDFBibTeX XMLCite