On a Nonparametric Test for Parallelism against Ordered Alternatives

  • Song, Moon Sup (Department of Computer Science and Statistics, Seoul National University) ;
  • Kim, Jaehee (Department of Computer Science and Statistics, Seoul National University) ;
  • Jean, Jong Woo (Department of Computer Science and Statistics, Seoul National University) ;
  • Park, Changsoon (Department of Applied Statistics, Choongang University)
  • Published : 1989.11.25

Abstract

A nonparametric test for testing the parallelism of regression lines against ordered alternatives is proposed. The proposed test statistic is based on a linear combination of robust slope estimators. It is a modified version of the Adichie's test statistics based on scores. A snail-sample Monte Carlo study shows that the proposed test is compatible with the Adichie's test.

Keywords

Acknowledgement

Supported by : Ministry of Education