A second popular method of estimation depends less on in- formation known to individuals and more on researchers ability to reach hidden populations repeatedly (by means, for example, such as successive waves of Respondent Driven Sampling). According to the logic of capture-recapture studies, successive samples that discover a proportion of identical individuals can be used to estimate the total population size by the well-known Lincoln-Peterson formula (discussed below). Multiple resam- pling increases the accuracy of these predictions. Where RDS has proven capable of reaching large samples of hidden popula- tions, it would appear ideally suited to such tasks. Problems arise, however, where initial sampling paths can be seen to affect subsequent referral paths, thus skewing the “recapture” process to those in the original sample (and resulting in an in- accurate recapture number, see Berchenko & Frost, 2011). Given these issues, what seems needed is a process that is less susceptible to discovery bias around stigmatized behaviors (a problem for NSUM) and not dependent on resampling proce- dures that may be biased by initial sampling (as is the issue for RDS-based capture-recapture methods), and finally, one that is capable of retaining respondent anonymity throughout the re- search process. Below we propose such