Estimating the size of a hidden finite set: large-sample behavior of estimators (Q2286344)

In this paper, the authors review models and methods for estimating the size of a hidden set, the elements of which are neither directly enumerable nor the size of the set can be estimated by a deterministic direct survey. Such a problem arises in several fields such as public health, epidemiology, demography, ecology, sociology, and intelligence analytics. After defining a finite population setup, the authors introduce two asymptotic regimes one of which is called the infill asymptotic regime which consists of drawing samples repeatedly (with intersection among different samples) from a set of fixed finite size population. Since it is difficult to sample infinitely many times from the population due to practical constraints, outfill asymptotic regime is introduced where the size of the sample as well as the size of the hidden set population asymptotically go to infinity. Statistical properties of the estimators under the two regimes are discussed for the case of ordered and unordered sets with illustrations from German tank problem, failure time models, network scale-up estimator, Horvitz Thompson estimator for unequal probability sampling, multiplier and capture-recapture methods. After characterizing consistency property and rates of estimation error, a discussion at the end provides theoretical considerations for the preference among the estimators considered for the size of the hidden set.