Free Trial

Safari Books Online is a digital library providing on-demand subscription access to thousands of learning resources.

Share this Page URL

Nine: Examples > 9.5 Horvitz-Thompson Open-population Size Estimates - Pg. 241

EXAMPLES 241 estimates provided by the model, this difference is not trivial. As pointed out previously, Manly et al. (1999) reported that Test 2 and Test 3 have low power. White (2002), on the other hand, performed simulations sug- gesting that bootstrapped estimates of GOF may be unreliable and that Test 2 and Test 3 are preferable. Based on the results of White (2002) and until further investigation of these techniques can be performed, we rec- ^ ommended that c be estimated based on the results of Test 2 and Test 3. 9.5 Horvitz-Thompson Open-population Size Estimates As pointed out above, one difference between the JS model and the CJS model approach is that the CJS model deals only with the marked com- ponent of the population. Because it does not make inference to the un- marked portion the CJS model provides estimates of capture and survival probabilities only. Capture probabilities are, however, related to popula- tion sizes. If the investigator is willing to assume that all European dippers in the study area had positive, but not necessarily equal, capture proba- bilities, then the Horvitz-Thompson estimates of population size given by equation 5.5 can be computed.