Paper detail

Quantifying the Biases of Spectroscopically Selected Gravitational Lenses

Spectroscopic selection has been the most productive technique for the selection of galaxy-scale strong gravitational lens systems with known redshifts. Statistically significant samples of strong lenses provide a powerful method for measuring the mass-density parameters of the lensing population, but results can only be generalized to the parent population if the lensing selection biases are sufficiently understood. We perform controlled Monte Carlo simulations of spectroscopic lens surveys in order to quantify the bias of lenses relative to parent galaxies in velocity dispersion, mass axis ratio, and mass density profile. For parameters typical of the SLACS and BELLS surveys, we find: (1) no significant mass axis ratio detection bias of lenses relative to parent galaxies; (2) a very small detection bias toward shallow mass density profiles, which is likely negligible compared to other sources of uncertainty in this parameter; (3) a detection bias towards smaller Einstein radius for systems drawn from parent populations with group- and cluster-scale lensing masses; and (4) a lens-modeling bias towards larger velocity dispersions for systems drawn from parent samples with sub-arcsecond mean Einstein radii. This last finding indicates that the incorporation of velocity-dispersion upper limits of \textit{non-lenses} is an important ingredient for unbiased analyses of spectroscopically selected lens samples. In general we find that the completeness of spectroscopic lens surveys in the plane of Einstein radius and mass-density profile power-law index is quite uniform, up to a sharp drop in the region of large Einstein radius and steep mass density profile, and hence that such surveys are ideally suited to the study of massive field galaxies.