{"id":"arxiv/2607.08415","url":"https://arxiv.org/abs/2607.08415","tags":["math.ST","math.PR"],"score":0,"title":"An Exact Distribution-Free Test for Means of Nonnegative Random Variables","domain":"math.ST","source":"arxiv","authors":["Nikos Vlassis","Philip S. Thomas"],"reviews":0,"subject":"math.ST","summary":"Let $X=(X_1,\\ldots,X_n)$ be independent nonnegative random variables, not necessarily identically distributed. Let $D=(D_0,D_1,\\ldots,D_n)\\sim\\operatorname{Dir}(1,\\ldots,1)$ be independent of $X$, and define $K(x)=\\mathbb{P}\\{\\sum_{i=1}^n x_iD_i\\le1\\}$. We prove that, for every $n\\ge1$, whenever $\\mathbb{E} X_i\\le1$ for every $i$, $\\mathbb{P}\\{K(X)\\leα\\}\\leα$ for all $0\\leα\\le1$. Thus $K(X)$ is a finite-sample, distribution-free $p$-value for testing the null hypothesis $\\mathbb{E}X_i \\le 1$ for all $i$. This proves a conjecture of Gaffke (2005).","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-07-09","reactions":{}}