Description
Reverberation mapping (RM) of Active galactic nuclei is a multi-messenger time domain that provides one of the most direct probes of the geometry and kinematics of the AGN broad-line region. It combines broad-band photometry and emission line spectroscopy to construct light curves to measure time delays between the continuum and line variability, with this lag time acting as a ruler of the broad line region's characteristic scale. However, modern multi-year "industrial scale" RM surveys like OzDES and SDSS are impacted by the problem of "aliasing", wherein the multimodal lag posterior distribution confounds tools like JAVELIN into returning incorrect or entirely spurious false positive lags. In this work we use a new fully Bayesian lag fitting tool, LITMUS, to construct a new RM lag measurement framework that leverages LITMUS's Bayesian evidence measurements to identify false positive lag measurements in a principled way. Using our new pipeline to re-analyse the full OzDES reverberation mapping sample, we find that previous reverberation mapping studies are likely to have markedly overestimated the confidence of recovered lags, and that, while the H-beta and CIV emission lines reverberate in the full sample, the MgII line does not exhibit detectable reverberation in a large fraction of AGN.