Abstract
Friedman’s chi-square test is a non-parametric statistical test for r treatments across n trials to assess the null hypothesis that there is no treatment effect. We use Stein’s method with an exchangeable pair coupling to derive a 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 , and also has an optimal dependence on the parameter r, in that the bound tends to zero if and only if . From this bound, we deduce a Kolmogorov distance bound that decays to zero under the weaker condition .
Funding Statement
During this research, RG was supported in part by EPSRC grant EP/K032402/1 and a Dame Kathleen Ollerenshaw Research Fellowship. GR has been funded in part by EPSRC grants EP/K032402/1, EP/T018445/1 and EP/R018472/1.
Acknowledgements
Firstly we acknowledge many helpful comments by anonymous referees. Moreover, we would like to thank Persi Diaconis for bringing this problem to our attention.
Citation
Robert E. Gaunt. Gesine Reinert. "Bounds for the chi-square approximation of Friedman’s statistic by Stein’s method." Bernoulli 29 (3) 2008 - 2034, August 2023. https://doi.org/10.3150/22-BEJ1530
Information