census.gov Notification
Due to the lapse of federal funding, portions of this website are not being updated. Any inquiries submitted via www.census.gov will not be answered until appropriations are enacted.

Maximum Norm Ratio Test

Written by:
RRS2025-01

Abstract

Tsao and Wright (1983) proposes the use of a metric called the “maximum ratio” to (deterministically) test whether one of K competing estimates of an unknown population parameter θ is outside of an “acceptable” region, whose radius is a multiple α × θ for some α ∈ (0, 1). Hall (2024) generalizes this (deterministic) test to a (deterministic) “multidimensional maximum ratio” test that assesses whether some vector composed of elements from K G-dimensional estimates of an unknown G-dimensional population parameter θ is outside of an “acceptable” region, whose radius is a multiple α × ∥θ∥ for some α ∈ (0, 1). This test flags collections of elements taken from these estimates as unacceptable, rather than flagging individual G-dimensional estimates. As a result, it can give false positives when there are unacceptable collections of elements but all estimates are acceptable. This paper proposes a (deterministic) maximum ratio test for G-dimensional parameter estimates that will not give false positives, proves its validity, and compares its performance to the multidimensional maximum ratio test discussed by Hall (2024).

Page Last Revised - July 22, 2025