English

Bounds for the chi-square approximation of Friedman's statistic by Stein's method

Statistics Theory 2022-07-01 v3 Probability Statistics Theory

Abstract

Friedman's chi-square test is a non-parametric statistical test for r2r\geq2 treatments across n1n\ge1 trials to assess the null hypothesis that there is no treatment effect. We use Stein's method with an exchangeable pair coupling to derive an explicit bound on the distance between the distribution of Friedman's statistic and its limiting chi-square distribution, measured using smooth test functions. Our bound is of the optimal order n1n^{-1}, and also has an optimal dependence on the parameter rr, in that the bound tends to zero if and only if r/n0r/n\rightarrow0. From this bound, we deduce a Kolmogorov distance bound that decays to zero under the weaker condition r1/2/n0r^{1/2}/n\rightarrow0.

Keywords

Cite

@article{arxiv.2111.00949,
  title  = {Bounds for the chi-square approximation of Friedman's statistic by Stein's method},
  author = {Robert E. Gaunt and Gesine Reinert},
  journal= {arXiv preprint arXiv:2111.00949},
  year   = {2022}
}

Comments

39 pages. This is a technical report. This version differs from the original version by correcting some numerical mistakes and presentation issues