Interval Estimation of the Size of a Small Population from a Mark-Recapture Experiment

SUMMARY We consider the problem of estimation of the size of a small population based on the results of a certain type of capture-recapture experiment. Seven methods for constructing confidence intervals for the population size, N, are considered. One of the methods, the method, is new, while the other six are maximum likelihood methods. For large confidence levels (i.e., 90% or more) extensive computer simulation of the experiment indicates that the ad hoc method tends to yield more reliable results for small N. In particular, for N - 500 the ad hoc method tends to give actual confidence levels that are close to the desired level and tends to give shorter intervals than the other methods when the probability of capturing individuals is small.