robsurvey-package {robsurvey}

Package Overview

Description

A key design pattern of the package is that the majority of the estimating methods is available in two "flavors":

• bare-bone methods

• survey methods

Bare-bone methods are stripped-down versions of the survey methods in terms of functionality and informativeness. These functions may serve users and package developers as building blocks. In particular, bare-bone functions cannot compute variances.

The survey methods are much more capable and depend, for variance estimation, on the survey package.

Basic Robust Estimators

Trimming

• Bare-bone methods: weighted_mean_trimmed and weighted_total_trimmed

• Survey methods: svymean_trimmed and svytotal_trimmed

Winsorization

• Bare-bone methods:

  • weighted_mean_winsorized and weighted_total_winsorized

  • weighted_mean_k_winsorized and weighted_total_k_winsorized

• Survey methods:

  • svymean_winsorized and svytotal_winsorized

  • svymean_k_winsorized and svytotal_k_winsorized

Dalen's estimators (weight reduction methods)

• Bare-bone methods: weighted_mean_dalen and weighted_total_dalen

• Survey methods: svymean_dalen and svytotal_dalen

M-estimators

• Bare-bone methods:

  • weighted_mean_huber and weighted_total_huber

  • weighted_mean_tukey and weighted_total_tukey

  • huber2 (weighted Huber Proposal 2 estimator)

• Survey methods:

  • svymean_huber and svytotal_huber

  • svymean_tukey and svytotal_tukey

  • mer (minimum estimated risk estimator)

Survey Regression (weighted least squares)

svyreg

Robust Regression and Ratio Estimation (weighted)

• Regression M-estimators: svyreg_huberM and svyreg_tukeyM

• Regression GM-estimators (Mallows and Schweppe): svyreg_huberGM and svyreg_tukeyGM

• Ratio M-estimators: svyratio_huber and svyratio_tukey

Note: The functions svyreg_huber and svyreg_tukey are deprecated, use instead svyreg_huberM and svyreg_tukeyM, respectively; see also robsurvey-deprecated.

Robust Generalized Regression (GREG) and Ratio Prediction of the Population Mean and Total

• Regression predictors: svymean_reg and svytotal_reg

• Ratio predictors: svymean_ratio and svytotal_ratio

Utility functions

• weighted_quantile and weighted_median

• weighted_mad and weighted_IQR

• weighted_mean and weighted_total

• weighted_line, weighted_median_line, and weighted_median_ratio