Estimation of Non-Sampling Errors for the Integrated Survey of Households (1990)  

 

Abstract

The Integrated Survey of Households (ISH) is major statistical activity conducted by the National Statistics Office (NSO) on a continuing basis that provides current information on labor force, employment and other socio-economic characteristics of the population.  This research addressed the role of NSE in quality of data from the ISH by dealing on the following objectives: (1) to develop a conceptual framework on the problems of NSE as applied to the ISH operations; (2) to identify the sources and types of NSE in the ISH and develop/prepare operational instructions for their measurements; (3) to propose appropriate methodologies for the control and evaluation of the different components of NSE; (4) to investigate and incorporate into the ISH/NSO Manual(s) on Concepts, Definitions and Methodology, pertinent procedures to continuously detect these sources and types of NSE, and institute procedures necessary to measure and control them; (5) to provide certain standard measures of reliability and accuracy of the ISH data for evaluation and presentation to users (6) to propose methods for and raise certain theoretical and empirical issues in addressing the problem of adjustments and imputations of missing units and data items in the ISH; and (7) to promote awareness of NSE in the NSO organization through seminars/workshops on NSE during the project’s life.  Sources and types of NSE in the ISH and recommendations were given on: coverage errors, concepts and  definitions of household members, nonresponse substitution, response errors, minimizing response burden in rotation of survey samples, language of questionnaire/ interview and respondent interaction, strict supervision in conduct of survey, and methods of weighing adjustments and imputations.