Comparison of estimation methods for the size of a closed population when there is variation between repeated samples and there is time and behavior dependence and heterogeneity in a capture-recapture situation
Abstract:
The objective of this study was to compare the accuracy of the population size
estimators from capture recapture sampling. The estimation methods were sample
coverage method, conditional sample coverage technique and maximum quasilikelihood
estimator. The Trap-shy and Trap-happy values were fixed at 0.5 and 1.2,
respectively. The investigation was done using 10,000 Monte Carlo simulation runs with
various population sizes, number of trapping samples, the first-capture probabilities and
multiplicative e time effect of the jth trapping sample. The evaluation criteria was the root
mean square error. The results show that.
1. The sample coverage method was appropriate when Trap-shy = 0.5 and the
first capture probabilities were equal or when Trap-happy = 1.2 and the first capture
probabilities were not equal.
2. The conditional sample coverage technique was appropriate when Trap-shy =
0.5 and the first capture probabilities were not equal.
3. The maximum quasi-likelihood estimator was appropriate when Trap happy =
1.2 and the first capture probabilities were equal.