We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

math.ST

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Mathematics > Statistics Theory

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

Abstract: Friedman's chi-square test is a non-parametric statistical test for $r\geq2$ treatments across $n\ge1$ 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 $n^{-1}$, and also has an optimal dependence on the parameter $r$, in that the bound tends to zero if and only if $r/n\rightarrow0$. From this bound, we deduce a Kolmogorov distance bound that decays to zero under the weaker condition $r^{1/2}/n\rightarrow0$.
Comments: 39 pages. This is a technical report. This version differs from the original version by correcting some numerical mistakes and presentation issues
Subjects: Statistics Theory (math.ST); Probability (math.PR)
MSC classes: Primary 60F05, 62E17
Cite as: arXiv:2111.00949 [math.ST]
  (or arXiv:2111.00949v3 [math.ST] for this version)

Submission history

From: Robert Gaunt [view email]
[v1] Mon, 1 Nov 2021 13:48:13 GMT (39kb)
[v2] Mon, 9 May 2022 10:01:32 GMT (40kb)
[v3] Thu, 30 Jun 2022 15:03:54 GMT (40kb)

Link back to: arXiv, form interface, contact.