R: Robust self-calibration

selfcalibrate_rob

R Documentation

Robust self-calibration

Description

Reload the datasets written to a file with the function

Usage

selfcalibrate_rob(design, formula, population, k, bounds = c(0, 1e4), Q = TRUE, R = TRUE, verbose = FALSE, maxit = 50, eps = 1e-7)

Arguments

`design`	a `survey.design` object
`formula`	one-sided formula (e.g., `~myvariable`)
`population`	true population mean
`k`	robustness tuning constant
`bounds`	vector of size two, defining the boundaries of the interval to which the sampling weights are constrained to
`Q`	logical, switch on/off Q-part of QR-estimator (default: `TRUE`)
`R`	logical, switch on/off R-part of QR-estimator (default: `TRUE`)
`verbose`	logical, if `TRUE` the function prints out details (default: `FALSE`)
`maxit`	maximal number of iterations (Newton-Raphson algorithm, default: `50`)
`eps`	numerical tolerance in matching sample mean and population mean(default: `1e-7`)

Details

The function selfcalibrate_rob requires the following R packages:

More explanations

Value

A (calibrated) survey.design object

warning

You must not call this function unless you really know how it works! Also, the function does not check whether the provided arguments are meaningful. It is the user's responsibility to make sure that the arguments are properly specified.

Author(s)

Tobias Schoch

References

Schoch, T. and A. Müller (2020. Treatment of sample under-representation and skewed heavy-tailed distributions in survey-based microsimulation: An analysis of redistribution effects in compulsory health care insurance in Switzerland, AStA Wirtschafts- und Sozialstatistisches Archiv 30, pp. 2–23.