ANALYSIS OF DESIGN EFFECT FOR INDONESIAN NATIONAL LABOUR FORCE SURVEY
Keywords:clustering effect, rate of homogeneity, weighting effect
The implementation of multistage sampling design is a good strategy to achieve the gain in efficiency of survey cost. However, in terms of sampling efficiency, it leads to the loss of precision indicated by the higher sampling variance compared to SRS design. Design effect measures the ratio of actual variance to the variance of SRS and can be decomposed to the effect of sample weight and the effect of clustering. This study aims to analyse the effect of sample weight and the effect of clustering on the estimation of labour variables resulted from the National labour Force Survey of Indonesia. The analysis is provided at the national level, stratum level, and province level. In general, the study finds that the design effect varies between labour variables. The effect of clustering is higher than the effect of the sample weight. There is also a high variability of the clustering effect between provinces and between strata (urban-rural). In contrast, the design effect due to the sample weight is similar between provinces, but it differs between strata. Allocating sample size proportionally to each stratum could be a good strategy for dealing with the high effect of weighting. On the other hand, for the future specific survey that measures the variable with a high clustering effect and high rate of homogeneity, the alternative strategy is increasing the sample size of the cluster and declining the sample size of households per cluster
Cochran, W. G., & William, G. (1977). Sampling Techniques. New York: John Wiley& Sons. Inc.
Kalton, G. (1979). Ultimate cluster sampling. Journal of the Royal Statistical Society: Series A (General), 142(2), 210–222.
Kalton, G., Brick, J. M., & Lê, T. (2005). Chapter VI Estimating components of design effects for use in sample design.
Kish, L. (1995). Survey sampling (Vol. 60). Wiley-Interscience.
Mecatti, F. (2014). Sampling elusive populations: Applications to studies of child labour by Vijay Verma, 2013 International Labour Organization, Department of Statistics-Geneva: ILO. Sampling Elusive Populations: Applications to Studies of Child Labour by Vijay Verma, 2013 International Labour Organization, Department of Statistics-Geneva: ILO., 113–114.
Park, I., & Lee, H. (2006). Design effects for the weighted mean and total estimators under complex survey sampling. Quality Control and Applied Statistics, 51(4), 381–384.
Pettersson, H., & Silva, P. (2005). Analysis of design effects for surveys in developing countries. Household Sample Surveys in Developing and Transition Countries, 123–143.
Verma, V., & Lê, T. (1996). An analysis of sampling errors for the demographic and health surveys. International Statistical Review/Revue Internationale de Statistique, 265–294.
Verma, V., Scott, C., & O’Muircheartaigh, C. (1980). Sample designs and sampling errors for the World Fertility Survey. Journal of the Royal Statistical Society: Series A (General), 143(4), 431–463.