Posterior predictive checks on the NANOGrav 15 yr dataset

Five pulsar timing array collaborations have recently released evidence for a gravitational-wave background. Further data taking and a data-combination effort are underway, but in the meantime we can perform checks on the models and assumptions used in the analysis. In this talk, we present results of posterior predictive checks on the NANOGrav 15 yr dataset. We re-evaluate the significance of the frequentist optimal statistic, we test the power-law assumption for the GW background and intrinsic red noise for each pulsar, and we perform leave-one-out cross-validation on each of the individual pulsars.