NEWS.md

In sampcompR: Comparing and Visualizing Differences Between Surveys

sampcompR (development version)

Changes in version 0.3.1.2 (2025-03-07)

• Fixed a small bug where multivariate biases got the wrong color, when one type of bias was not present

Changes in version 0.3.1.1 (2025-02-07)

• Fixed a small bug so univariate and bivariate bias functions also work with factor variables

Changes in version 0.3.1 (2025-13-03)

• Added two parameter to plot_uni_compare to adjust the size of the dots and the errorbars

Changes in version 0.3.0 (2025-02-03)

• fix small bug in shapes of plot_uni_compare
• add a new function, to create a heatmap of relative bias for bivariate comparison
• add a new function, to show AAB or AARB in Pearson's r per variable as a table
• add a new function, to calculate missings per variable and display them in a table

Changes in version 0.2.7 (2025-21-01)

• fixed a small error in biv_compare that occurred when sample size was very small and ended in an Error message.
• Added absolute relative bias to biv_compare output

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 (\code{uni_compare}, \code{biv_compare}, \code{multi_compare}.

• The functions are much faster during bootstrapping now.

• The possibility to weight the dataset to the benchmark using \code{raking} and \code{post-stratification}.

sampcompR documentation built on Aug. 8, 2025, 7:35 p.m.