{"id":"arxiv/2601.02529","url":"https://arxiv.org/abs/2601.02529","tags":["stat.ME","math.ST"],"score":0,"title":"A novel finite-sample testing procedure for composite null hypotheses via pointwise rejection","domain":"stat.ME","source":"arxiv","authors":["Joonha Park","Ming Wang"],"reviews":0,"subject":"stat.ME","summary":"We propose a novel finite-sample procedure for testing composite null hypotheses. Traditional likelihood ratio tests based on asymptotic $χ^2$ approximations can exhibit substantial size distortions in small samples. Our procedure rejects the composite null hypothesis $H_0: θ\\in Θ_0$ if the simple null hypothesis $H_{0,t}: θ= θ_t$ is rejected for every $θ_t \\in Θ_0$, using a suitably inflated significance level. We derive a general formula for calibrating this inflated level based on the geometry of the null region $Θ_0$. By employing finite-sample methods for testing the simple hypotheses…","category":"research","comments":0,"disputed":false,"division":"stat","mentions":0,"discussed":0,"published":"2026-01-05","reactions":{}}