Estimates of Sampling Error
Accuracy and reliability of estimates in a survey such as HIES are measured by computing the sampling errors, specifically standard errors and coefficient of variations. Although survey estimates are subject to both sampling and non sampling errors, only sampling errors can be calculated. Non sampling errors are minimized by means of quality control and operational control measures in all phases of the survey. In practice, estimates with coefficient of variations (CV) of less than 20% are considered as statistically ‘accurate’. On the other hand, those estimates with (CV) more than 20% are ‘indicative’ and should not be used at face value.
Note: Standard errors and coefficient of variation for key variables are presented in Appendix 2 of the 2008 HIES final report.