Skip to contents

sampcompR (development version)

Changes in version 0.2.6 (2024-14-11)

  • As the wooldridge package was archived on CRAN, and our examples rely on the card data of wooldridge, we added the card data to our package

Changes in version 0.2.5 (2024-14-11)

  • Added the possibility to input benchmarks as a named vector of means in univariate comparison
  • Fixed small errors in the uni_compare_table functions that occurred when only one benchmark and survey were compared.
  • Changed example to show bias comparison (estimating bias if only white respondents or if only north respondents would have been recruited)

Changes in version 0.2.4 (2024-10-11)

  • Added parameters that allow to bootstrap both benchmark and survey
  • Added parameter to allow to choose between percentile or normal bootstrap confidence intervals and p-values based on those intervals.

Changes in version 0.2.3 (2024-19-08)

  • Additional small Fix for a test

Changes in version 0.2.2 (2024-22-07)

  • Small Fix for test on CRAN for MacOS

Changes in version 0.2.1 (2024-14-07)

  • We added biv_per_variable() a function to calculate the average bias per variable for the bivariate comparison, and an average bias per variable across comparisons.

  • We added multi_per_variable() a function to calculate the average bias per coefficient and per model, for the multivariate comparison, and an average biases per coefficient and per model across comparisons.

Changes in version 0.2.0 (2024-08-07)

  • We implemented better bootstrapping, that will use weighting in every bootstrap iteration, for all main functions (, , .

  • The functions are much faster during bootstrapping now.

  • The possibility to weight the dataset to the benchmark using and .