FUTA JOURNAL OF RESEARCH IN SCIENCE

VOLUME 13 NUMBER 1 2017

Paper Details

  • Title :COMPARISON OF NON-PARAMETRIC METHODS FOR PAIRED SAMPLE DATA
  • Author(s) : A.H. Bello and O.I. Teniola
  • Abstract:

This study was on the comparison of non-parametric methods for paired sample data. A sample of fifty healthy adults’ body temperatures were used to analyze the statistical methods used. The study was aimed at comparing a rank based test statistic for testing the equality of two population medians and comparing the Oyeka and Umeh method with sign test and Wilcoxon signed rank test. The three statistical methods were compared with each other to know which was better by testing the efficiency, standard error and minimum variance. The standard error of the proposed method Oyeka and Umeh method was 0.947 as compare to that of Wilcoxon signed rank which was 13.207. The efficiency of the proposed test statistic relative to the Wilcoxon sign ranked test is 175.12, which was greater than 1, therefore the proposed method is more efficient than the Wilcoxon signed rank and sign tests. It also has an added advantage over the two tests because it uses all the available information in the data.
Keywords: Non-parametric, Rank, Sign test, Wilcoxon signed rank test, Paired Sample Data