A sample coverage approach for estimating the size of the population with application to death undercount

authors:

avatar Iman Taghadosi , avatar Anoshirvan Kazemnejad ORCID , * , avatar Nahid Senjeri , avatar Soghrat Faghiezade , avatar alireza abadi

Koomesh: Vol.5, issue 1; 37-44
published online: February 15, 2004
article type: Research Article

how to cite: Taghadosi I, Kazemnejad A, Senjeri N, Faghiezade S, abadi A. A sample coverage approach for estimating the size of the population with application to death undercount. koomesh. 2004;5(1):e151993. 

Abstract

Introduction: Existence of numerous statistics and sometimes contradictory from unique phenomena is one of the problems and difficulties that prevent policy makers and administrators to take the right steps and make correct decisions. Generation of numerous statistics and sometimes contradictory by relevant organizations on mortality registration has created much uncertainty among governmental administrators and decision makers. This difference between statistics shows that a type of undercount exists in information collecting system of these organizations. Sample coverage approach is one of the procedures that might enable one to remove this deficiency. This study was accomplished to estimate the number of deaths of Bushehr province in year 2000 and to correct the death undercount. Materials and Methods: First, death information was obtained from three sources: Statistics System Development Department, Health Promotion and Health Network Development Department of Ministry of Health and Medical Education, and Census Organization of Iran. Then information of each source that was based on the first name, last name, gender and residence of dead people was entered to computer. Finally, a number of soft wares such as Access were used for obtaining common cases of sources. The data were then analyzed by SPLUS software and the following results were obtained. Results: Without considering the number of newborn deaths, the total number of death in Bushehr province was estimated at 660 cases with an estimated standard error of 5 cases. However, only 641 cases had been reported by relevant organizations. As a result, 19 cases attributed as undercounts (2.88 %) were estimated by these organizations. Conclusion: Results of this research has been shown that a type of undercount exists on mortality registration and could be obtained us an acceptable estimation for this undercount by using sample coverage approach.