Testing for monotonicity in the Hubble diagram
Résumé
General relativistic kinematics and the cosmological principle alone imply a monotonicity constraint in the Hubble diagram, which we confront to present-day supernova data. We use the running gradient method of statistical inference by Hall & Heckman (2000). We find no significant departure from monotonicity. The method seems well adapted and we recommend its use with future data.