Development of an efficient SAS® macro to perform permutation tests for two independent samples

  • Authors:

  • G. K. Balasubramani;Stephen R. Wisniewski;Hongwei Zhang;Heather F. Eng

  • Affiliations:

  • Epidemiology Data Center, Graduate School of Public Health, University of Pittsburgh, Pittsburgh, PA 15261, USA;Epidemiology Data Center, Graduate School of Public Health, University of Pittsburgh, Pittsburgh, PA 15261, USA;Epidemiology Data Center, Graduate School of Public Health, University of Pittsburgh, Pittsburgh, PA 15261, USA;Epidemiology Data Center, Graduate School of Public Health, University of Pittsburgh, Pittsburgh, PA 15261, USA

  • Venue:
  • Computer Methods and Programs in Biomedicine
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

When underlying distributions are unknown and sample sizes are small, permutation tests may have superior features to parametric tests, such as smaller type I error and more accurate probability values when testing the null hypothesis. However, permutation tests are not widely used in clinical trials because of their computational complexity. We developed an efficient SAS macro to generate all possible permutations of the data and to run subsequent permutation tests of difference between means for two independent samples. The macro performs permutation tests and provides the exact probability of significance for a wide range of statistics, including geometric means, medians, mid-ranges, mean-ranks, proportions and variances to meet the needs of data analysis in clinical trials.