Parameters of the basic model consisting of the primary power law and the thermal plasma component, where present. More complex model parameters are presented in Tables 4–8. Column (1): galaxy name. Column (2): χ²/dof of the fit. ^(F) indicates that the full band spectral fit method was used, ^(D) indicates that although the χ²_r of this fit is bad (χ²_r > 2), the data/model ratio never exceeded 5 per cent, which is the level of uncertainties in the XMM–Newton calibration, ^(H) indicates that only the 2.5–10 keV fit parameters have been presented here due to spectral complexities below 2 keV. For the 2.5–10 keV fits, only NGC 1068, NGC 1365 and M82 have bad fits. These sources have notoriously complex spectra, which are high signal-to-noise ratio in this case, with several emission and absorption features. We do not attempt to fit these additional complexities, however, and present only the basic hard band parameters for these fits. No errors are quoted because the standard formalism does not apply unless the model fits the data (Lampton et al. 1976). Column (3): galactic column density (10²⁰ cm⁻²). Column (4): the neutral column density measured in the primary power law (10²² cm⁻²). * indicates that an ionized absorber has been detected in this spectral fit, details of which are in Table 7. Column (5): the power-law index of the primary power law. † indicates that this parameter was fixed in the fit. Column (6): the normalization of the primary power law (×10⁻⁵ photons cm⁻² s⁻¹ keV⁻¹ at 1 keV). Column (7): the temperature, in keV, of the thermal plasma component. Column (8): the ratio of the normalization of the thermal plasma component to the normalization of the primary power law. Column (9): observed flux in the 2–10 keV band in units of 10⁻¹⁴ erg s⁻¹ cm⁻² and the logarithm of the intrinsic luminosity in this band in units of erg s⁻¹.