Model Selection and the GREG Estimator Bias in a Small Business Survey
DOI:
https://doi.org/10.15678/ZNUEK.2017.0971.1101Keywords:
GREG, business statistics, model-assisted estimation, outliersAbstract
Estimation for a very skewed population containing extreme values is problematic, especially at a low level of aggregation. Traditional direct estimation methods do not provide satisfactory results. The growing demand for detailed information and the wider possibility of using data from administration registers has increased the importance of recognising more sophisticated estimation methods. Generalised Regression (GREG) estimation is an example of one such type. The paper examines the importance of the model chosen in GREG estimation in dealing with highly variable and outlier-prone populations. The model-assisted GREG estimator is applied to a real business survey. Lagged variables from administrative registers were used as the auxiliary variables. The variable of interest – mean revenue of small companies – was estimated for provinces cross-classified by categories of economic activity.
Downloads
References
Bracha C. [2004], Estymacja danych z badania aktywności ekonomicznej ludności na poziomie powiatów dla lat 1995–2002, GUS, Warszawa.
Chambers R., Chandra H., Salvati N., Tzavidis N. [2014], Outlier Robust Small Area Estimation, „Journal of the Royal Statistical Society: Series B (Statistical Methodology)”, vol. 76, no 1, https://doi.org/10.1111/rssb.12019. DOI: https://doi.org/10.1111/rssb.12019
Chambers R. L., Falvey H., Hedlin D., Kokic P. [2001], Does the Model Matter for GREG Estimation? A Business Survey Example, „Journal of Official Statistics”, vol. 17, nr 4.
Clark R.G., Kokic P., Smith P.A. [2017], A Comparison of Two Robust Estimation Methods for Business Surveys, „International Statistical Review”, vol. 85, nr 2, https://doi.org/10.1111/insr.12177. DOI: https://doi.org/10.1111/insr.12177
Dehnel G. [2014], Winsorization Methods in Polish Business Survey, „Statistics in Transition – New Series”, vol. 15, nr 1. DOI: https://doi.org/10.59170/stattrans-2014-007
Dehnel G. [2016], M-estimators in Business Statistics, „Statistics in Transition – New Series”, vol. 17, nr 4. DOI: https://doi.org/10.59170/stattrans-2016-041
Dehnel G. [2017], GREG Estimation with Reciprocal Transformation for a Polish Business Survey [w:] Proceedings of the 11th Professor Aleksander Zelias International Conference on Modelling and Forecasting of Socio-Economic Phenomena, eds M. Papież, S. Śmiech, Foundation of the Cracow University of Economics, Cracow.
Lehtonen R., Särndal C.E., Veijanen A. [2016], Generalized Regression and Model-calibration Estimation for Domains: Accuracy Comparison, https://www.researchgate.net/publication/228665672_Generalized_regression_and_model-calibration_estimation_for_domains_Accuracy_comparison (data dostępu: 17.12.2017).
Ludność. Stan i struktura demograficzno-społeczna. Narodowy Spis Powszechny Ludności i Mieszkań 2011 [2013], GUS, Warszawa. DOI: https://doi.org/10.59139/ws.2011.11.2
Rao J.N.K., Molina I. [2015], Small Area Estimation, 2nd ed., Wiley Series in Survey Methodology, Wiley, Hoboken, New Jersey.
Särdnal C. E., Swensson B., Wretman J. [1992], Model Assisted Survey Sampling, Springer Verlag, New York.
Wykorzystanie danych administracyjnych w badaniu: Ocena bieżącej działalności gospodarczej przedsiębiorstw [2016], GUS, Warszawa.
