Lurking Variable Detection via Dimensional Analysis

Lurking variables represent hidden information, and preclude a full understanding of phenomena of interest. Detection is usually based on serendipity -- visual detection of unexplained, systematic variation. However, these approaches are doomed to fail if the lurking variables do not vary. In this article, we address these challenges by introducing formal hypothesis tests for the presence of lurking variables, based on Dimensional Analysis. These procedures utilize a modified form of the Buckingham Pi theorem to provide structure for a suitable null hypothesis. We present analytic tools for reasoning about lurking variables in physical phenomena, construct procedures to handle cases of increasing complexity, and present examples of their application to engineering problems. The results of this work enable algorithm-driven lurking variable detection, complementing a traditionally inspection-based approach.