Comment by jl2718

That’s a good point. The selection from a lineup then gives less information. So the real issue is poorly understood priors.

I’m pretty sure that this can be used fairly with a rigorous Bayesian treatment.