Marginalised posterior distributions of noise hyperparameters obtained with the EPTA DR1, DR2full and DR2new. Top left: PSR J1730–2304 falls in Category 1, where the DR2full/New datasets are able to constrain chromatic noise (DMAmp, DMSlope). Top right: For PSR J1857+0943, belonging to Category 2, it is clear that the DR1 data could not disentangle chromatic from achromatic noise (RedAmp, RedSlope), which is resolved better with the DR2 datasets. Bottom left: PSR J1713+0747 in Category 3, shows clear inconsistencies in the achromatic noise estimates. Bottom right: PSR J1909–3744 shows a much more complicated behaviour as described in the text motivating its inclusion in Category 3.