Statistical methods for the analysis of maximum and minimum scores in repeated measures

Oyeka I. C. A., Okeh U. M. and Okoro C. N.

Physical Sciences Research International
Published: October 30 2013
Volume 1, Issue 4
Pages 104-109

Abstract

A researcher may have a set of sample data that may be dependent on time or space. The research interest may not be on the comparison of the mean or median but the modal or minimum scores in the series of experiments or trials. These problems may arise in situations such as in the economy where interest is in studying transactions at the stock exchange or fluctuations in the exchange rate of money where interest is in comparing the peaks, maximum or highest; trough, minimum or lowest scores. In education and public affairs, interest may be in studying the performance of subjects or candidates in a job interview overtime or space to determine whether the subjects or candidates are equally likely to earn the highest or lowest score or grades at each of the several time points or locations. This paper proposes a non-parametric statistical method for the comparison of modes or troughs of populations matched in time or space that do not require any distributional assumptions. Sample data are used to illustrate the proposed method. The method is compared with existing Statistical methods and shown to compares favorably with the method.

Keywords: C-matched, maximum scores, minimum scores, treatment, observations, Chi-square, Cochran Q test.

Full Text PDF