Estimation and analysis of design effects and intra-class correlation of selectedindicators from Philippine household surveys

• Main Authors: De Veyra, Janna M., Ong, Jovin David C.
• Format: text
• Language: English
• Published: Animo Repository 2014
• Subjects: Physical Sciences and Mathematics
• Online Access: https://animorepository.dlsu.edu.ph/etd_bachelors/17993
• Institution: De La Salle University

Household surveys constitute an essential and strategic component in the organization of needed information in the formulation of the policies in the Philippines. Hence the reliability of survey estimates is needed to be able to yield maximum precision with the least possible cost. The computation of design effect and intra-class correlation of selected indicators, both in the regional and in the national level, of the Family Income and Expenditure Survey (FIES) and Labor Force Survey (LFS) across time was used to test the efficiency of the survey design. Furthermore, to provide the required sample sizes and optimal number of primary sampling units (PSU) needed for a survey at a specified level of precision and cost. The computed design effect and intra-class correlation values revealed inconsistency and variability across space and time. The 2003 Master Sample (MS) is nearly adequate in design and implementation in terms of compensating the need to increase the sample sizes and the number of PSUs. Moreover, it may not be advisable to depend on national design effect and intra-class correlation values of selected indicators from FIES and LFS because there were variability on the behavior and pattern of values across regions and across time.