With an ecological-evolutionary perspective increasingly applied toward the conservation and management of endangered or exploited species, the genetic estimation of effective population size (Ne) has proliferated. Based on a comprehensive analysis of empirical literature from the past two decades, we asked: (i) how often do studies link Ne to the adult census population size (N)? (ii) To what extent is Ne correctly linked to N? (iii) How readily is uncertainty accounted for in both Ne and N when quantifying Ne/N ratios? and (iv) how frequently and to what degree might errors in the estimation of Ne or N affect inferences of Ne/N ratios? We found that only 20% of available Ne estimates (508 of 2617; 233 studies) explicitly attempted to link Ne and N; of these, only 31% (160 of 508) correctly linked Ne and N. Moreover, only 7% (41 of 508) of Ne/N ratios (correctly linked or not) reported confidence intervals for both Ne and N; for those cases where confidence intervals were reported for Ne only, 31% of Ne/N ratios overlapped with 1, of which more than half also reached below Ne/N = 0.01. Uncertainty in Ne/N ratios thus sometimes spanned at least two orders of magnitude. We conclude that the estimation of Ne/N ratios in natural populations could be significantly improved, discuss several options for doing so, and briefly outline some future research directions.