Page 323 - Contributed Paper Session (CPS) - Volume 6
P. 323
CPS1942 Daniel D. M. P.
2.2 Method of Evaluation
For each combination of model restriction and sampling rate, the point
estimate and standard error is generated for the SRSWOR and PPAS sampling
methods, while PPSS estimates are only generated for sampling rate 1%.
Standard errors serve as basis of efficiency of estimates, while the point
estimates are further used to calculate the absolute percent difference
defined by
where T is the true population total, and ̂ is the estimator of sampling
procedure. The estimates are expected to be unbiased in the long run, but a
single sample may incur discrepancy between the true value and the estimate.
The better estimator corresponds to a lower value.
3. Results
3.1 Estimation of Population Total
The estimates of population total using SRSWOR, PPAS, and PPSS
incorporating the three sampling rates (1%, 5%, 10%). There is notable bias in
the estimates, but a clearer comparison is presented using the in tables
3.3-3.5.
3.2 Standard Error of Estimates
An optimal estimate is characterized by small variability, hence, the
lower the standard error the more efficient the estimate. As expected, the
estimates improve as sampling rate increases, but under 1% sampling rate,
312 | I S I W S C 2 0 1 9