Multiple comparisons, selection bias, missing data, name tests, Bayesian priors, and false precision in sparse historical records.