Stess testing pulsar timing array models

The recent common process that has been observed by many pulsar timing arrays could presage a gravitational wave background detection. Such a detection would require carefully testing and checking that the models used for PTAs are appropriate, and we should highlight where mismodelling might occur and whether it can cause a "false" detection. Most pulsar timing array experiments use a power law to model the power spectral density of the gravitational wave background and intrinsic pulsar red noise. These assumptions can be relaxed using so-called "free spectral" Bayesian runs that model the power in each frequency bin individually, but they are time consuming. In this talk, I'll present a fast and efficient method to test the power law assumption in the GWB and intrinsic red noise. I'll show how we can perform fast, repeated simulations using this method to estimate the "significance" of deviations from a power law. I'll close by showing how we can use the formalism to test whether correlations between pulsars follow the Hellings and Downs curve.