Skoči na glavni sadržaj

Prethodno priopćenje

The choice of a sampling procedure for a (too) small target population: the case of Croatian public hospitals

Berislav Žmuk
Ivana Dražić Lutilsky
Martina Dragija


Puni tekst: engleski pdf 492 Kb

str. 19-44

preuzimanja: 3.152

citiraj


Sažetak

In some cases, a sampling approach must be used instead of a census even if the target population is small. In such a case, the drawn sample will also be small and every sample unit can significantly influence the results. This paper aims to address the aforementioned issue from the perspective of different sampling procedures. The observed small population consists of 57 Croatian hospitals. The averages of total due obligations for medicines of Croatian hospitals in 2014 are estimated using three different sampling procedures and compared to the real population parameter value. The analysis has shown that the minimum recommended overall sampling rate, which would result in satisfactory precision and accuracy levels of parameter estimates, for simple random sampling and stratified random sampling is 70%. The two-stage cluster sampling procedure, on the other hand, did not achieve a satisfactory precision and accuracy level at any of the observed sample sizes because of a high level of heterogeneity between clusters.

Ključne riječi

complex survey design; sampling; clustering; stratification; Croatian hospitals

Hrčak ID:

170199

URI

https://hrcak.srce.hr/170199

Datum izdavanja:

7.12.2016.

Posjeta: 4.122 *