Abstract
Tests of fundamental theory are sensitive to the values of the fundamental (physical) constants. These values are determined by a global fit to experimental data and depend on the data, theory, and theory uncertainties that enter the fit. Tests of fundamental theory are thus sensitive, to a greater or lesser extent, to procedural decisions. We label the set of choices that comprise a given global fit procedure as a scheme. No single, best scheme exists. For example, a scheme optimized to give values for the fundamental constants with the smallest possible uncertainties might omit anomalous data, while a scheme optimized to test fundamental theory might include anomalous data. Historically, values for the fundamental constants have been determined by the National Institute of Science and Technology (NIST) by an elaborate fit to world data--- i.e. by reference to a single scheme. The result is that the scheme dependence of the fundamental constants has effectively been removed from phenomenological consideration. Unfortunately the precision of modern experiments is such that the scheme dependence of tests of fundamental theory (occurring through the scheme dependence of the fundamental constants) can no longer be safely ignored. We argue that any test of fundamental theory should be evaluated using values for the fundamental constants determined via a scheme that includes all data and theory relevant to that test. The alternative--- namely, evaluating a given test of fundamental theory using fundamental constants determined via a scheme omitting relevant data or relevant theory or both--- has potentially serious consequences for the phenomenology that are difficult to predict from pure thought alone. We find that an omission of relevant data from the NIST determination of the fundamental constants is responsible for the proton size puzzle, with the puzzle disappearing under a scheme that includes all relevant data. We also find that schemes omitting a relevant theory alternative generally lead to falsely-restrictive exclusion limits--- a result that has implications for many proposed solutions to the muon g-2 anomaly. We have created an online interface called CONSTANT FINDER to enable the community-wide investigation of the scheme dependence of the fundamental constants. With CONSTANT FINDER, anyone interested can adjust experimental and theoretical values and uncertainties, and pose theory alternatives within several global data-fitting frameworks.