Abstract
In biostatistical studies, capture-recapture experiments have been widely used to estimate various parameters of biological populations. In capture-recapture experiments, there are two types of models: discrete-time and continuous-time models. In a discrete-time model, the target population is sampled over a certain number of occasion and traps are checked only once for each trapping occasion. In a continuous-time model, the target population is sampled in any time and we record the exact capture times for each animal in the tagging process. Accordingly, continuous-time models need a huge expenditure for sampling. On the other hand, there is no paper studying the relations between discrete-time and continuous-time models. For these reasons, we will compare discrete-time and continuous-time estimators of population size in this thesis. Mainly, we consider a series of capture-recapture models which are postulated depending on sources of variability in capture probabilities due to behavior, time, heterogeneity, or behavior and time. In these models, we find the differences between discrete-time and continuous-time estimators and compare the efficiencies of discrete-time and continuous-time estimators. Moreover, we use the method of simulation to analyze these estimators. Finally, we give an example of software failure data of Moek (1984) concerns an information system for registered aircraft movements to compare discrete-time and continuous-time estimators of population size in the various models we mention above.