Paper detail

Model Comparison tests of modified gravity from the Eöt-Wash experiment

Perivolaropoulos et al (arXiv:1904.09462) (P19) have argued that the residual torque data in the Eöt-Wash experiment is consistent with an oscillating signal. This could either be a signature of non-local modified gravity theories or some other systematic error in the data. We independently assess the viability of such an oscillating signal in the same data using Bayesian and information theoretical criterion, to complement the frequentist analysis in P19. We fit this data to three different parametrizations (an offset Newtonian, Yukawa, and an oscillating model), and assess the significance of the oscillating model using AIC, BIC, WAIC, and Bayes factor. All these techniques provide decisive evidence for the oscillating model compared to the Newtonian model, provided the phase is fixed to the same value as in P19. If the phase is allowed to vary, then significance from BIC, WAIC, and Bayes based tests reduces to strong evidence, whereas only AIC still shows decisive evidence. Our analysis codes have been made publicly available.