Robustness in Survey Sampling Using the Conditional Bias Approach with R Implementation
Abstract
The classical tools of robust statistics have to be adapted to the finite population context. Recently, a unified approach for robust estimation in surveys has been introduced. It is based on an influence measure called the conditional bias that allows to take into account the particular finite population framework and the sampling design. In the present paper, we focus on the design-based approach and we recall the main properties of the conditional bias and how it can be used to define a general class of robust estimators of a total. The link between this class and the well-known winsorized estimators is detailed. We also recall how the approach can be adapted for estimating domain totals in a robust and consistent way. The implementation in R of the proposed methodology is presented with some functions that estimate the conditional bias, calculate the proposed robust estimators and compute the weights associated to the winsorized estimator for particular designs. One function for computing consistently domain totals is also proposed.
Keywords
Conditional bias Design-based Huber-function Mean squared errorReferences
- 1.Beaumont, J.-F., Rivest, L.-P.: Dealing with outliers with survey data. In: Rao, C. R., Pfeffermann, D. (eds.) Handbook of Statistics, Sample Surveys: Theory Methods and Inference, vol. 29, pp. 247–279 (2009)Google Scholar
- 2.Beaumont, J.-F., Haziza, D., Ruiz-Gazen, A.: A unified approach to robust estimation in finite population sampling. Biometrika 100, 555–569 (2013)MathSciNetCrossRefzbMATHGoogle Scholar
- 3.Dalén, J.: Practical estimators of a population total which reduce the impact of large observations. R and D Report, Statistics Sweden (1987)Google Scholar
- 4.Demoly, E.: Les valeurs influentes dans lénqute TIC 2011: Traitements et expérimentations en cours. Acte du sminaire de méthodologie statistique de l’INSEE, juillet (2013)Google Scholar
- 5.Deroyon, T., Favre-Martinoz, C.: A comparison of Kokic and Bell and conditional bias methods for outlier treatment. In: Work Session on Statistical Data Editing. The Hague, Netherlands, Unece (2017)Google Scholar
- 6.Deville, J.-C., Särndal, C.E.: Calibration estimators in survey sampling. J. Am. Stat. Assoc. 87(418), 376–382 (1992)MathSciNetCrossRefzbMATHGoogle Scholar
- 7.Favre-Martinoz, C., Haziza, D., Beaumont, J.-F.: A method of determining the winsorization threshold, with an application to domain estimation. Surv. Methodol. 41(1), 57–77 (2015)Google Scholar
- 8.Kokic, P.N., Bell, P.A.: Optimal winzorizing cutoffs for a stratified finite population estimator. J. Official Stat. 10, 419–435 (1994)Google Scholar
- 9.Moreno-Rebollo, J.L., Muoz-Reyez, A.M., Muoz-Pichardo, J.M.: Influence diagnostics in survey sampling: conditional bias. Biometrika 86, 923–968 (1999)MathSciNetCrossRefGoogle Scholar
- 10.Muñoz-Pichardo, J., Muñoz-Garcia, J., Moreno-Rebollo, J.L., Piño-Mejias, R.: A new approach to influence analysis in linear models. Sankhya Series A 57, 393–409 (1995)MathSciNetzbMATHGoogle Scholar
- 11.Tambay, J.-L.: An integrated approach for the treatment of outliers in sub-annual surveys. In: Proceedings of the Survey Research Methods Section, American Statistical Association, Alexandria, Virginia, pp. 229–234 (1988)Google Scholar
- 12.Tillé, Y.: Sampling Algorithms. Springer (2011)Google Scholar
- 13.Tillé, Y., Matei, A.: The sampling package. Software manual, CRAN (2015). https://cran.r-project.org/src/contrib/Descriptions/sampling.html