Tainted Evidence: Cosmological Model Selection vs. Fitting

Abstract

Interpretation of cosmological data to determine the number and values of parameters describing the universe must not rely solely on statistics but involve physical insight. When statistical techniques such as "model selection" or "integrated survey optimization" blindly apply Occam's Razor, this can lead to painful results. We emphasize that the sensitivity to prior probabilities and to the number of models compared can lead to "prior selection" rather than robust model selection. A concrete example demonstrates that Information Criteria can in fact misinform over a large region of parameter space.

5 pages, 1 figure; v2 minor rephrasings, clarifications; v3 minor changes to match published article under title "Cosmological Model Selection: Statistics and Physics"

References in corpus (5)

• Information criteria for astrophysical model selection
• Reconstructing Dark Energy
• Present and future evidence for evolving dark energy
• Why anthropic reasoning cannot predict Lambda
• Biased Cosmology: Pivots, Parameters, and Figures of Merit

Cited by in corpus (10)

• The Mass Distribution of Stellar-Mass Black Holes
• Precision cosmology with Padé rational approximations: theoretical predictions versus observational limits
• Crossing Statistic: Reconstructing the Expansion History of the Universe
• Nonparametric reconstruction of the cosmic expansion with local regression smoothing and simulation extrapolation
• (Lack of) Cosmological evidence for dark radiation after Planck
• Crossing Statistic: Bayesian interpretation, model selection and resolving dark energy parametrization problem
• The Crossing Statistic: Dealing with Unknown Errors in the Dispersion of Type Ia Supernovae
• A Bayesian analysis of inflationary primordial spectrum models using Planck data
• The contribution of the kinematic Sunyaev-Zel'dovich Effect from the Warm Hot Intergalactic Medium to the Five-Year WMAP Data
• Falsifying Cosmological Constant