Skip to contents

vazul

vazul is an R package for analyis blinding in research contexts. It offers two main approaches to anonymize data while preserving analytical validity: masking (replacing values with anonymous labels) and scrambling (randomizing the order of existing values).

Analysis Blinding Approaches

Masking replaces original values with anonymous labels, completely hiding the original information:

treatment <- c("control", "treatment", "control")
mask_labels(treatment)
#> "masked_group_01" "masked_group_02" "masked_group_01"

Scrambling preserves all original values but randomizes their order:

scramble_values(treatment) 
#> "treatment" "control" "control"  # Same values, different order

Installation

You can install the released version of vazul from CRAN with:

Or the development version from GitHub with:

remotes::install_github("nthun/vazul")

Functions

Masking Functions

Replace categorical values with anonymous labels to completely hide original information.

mask_labels() - Mask vector values 

library(vazul)

# Basic masking
treatment <- c("control", "treatment", "control", "treatment")
set.seed(123)
mask_labels(treatment)
#> "masked_group_01" "masked_group_02" "masked_group_01" "masked_group_02"

# Custom prefix
mask_labels(treatment, prefix = "group_")
#> "group_01" "group_02" "group_01" "group_02"

mask_variables() - Mask data frame columns 

df <- data.frame(
  condition = c("A", "B", "A", "B"),
  treatment = c("ctrl", "test", "ctrl", "test"),
  score = c(85, 92, 78, 88)
)

# Mask multiple columns
mask_variables(df, c("condition", "treatment"))

# Use tidyselect helpers
mask_variables(df, where(is.character))

The .across_variables parameter allows for consistent masking across multiple columns (e.g., longitudinal data in wide format).

df <- data.frame(
  wave_1 = c("A", "B", "A"),
  wave_2 = c("B", "A", "B"),
  score = c(10, 20, 30)
)

# Mask across variables consistently
mask_variables(df, starts_with("wave_"), .across_variables = TRUE)

Scrambling Functions

Randomize the order of values while preserving the original data content.

scramble_values() - Scramble vector order 

# Numeric data
set.seed(123) 
scramble_values(1:5)
#> [1] 3 2 5 4 1

# Categorical data
scramble_values(c("A", "B", "C", "A", "B"))
#> [1] "B" "A" "C" "B" "A"

scramble_variables() - Scramble data frame columns 

df <- data.frame(x = 1:6, group = rep(c("A", "B"), each = 3))

# Scramble across entire column
scramble_variables(df, "x")

# Scramble within groups
scramble_variables(df, "x", .groups = "group")

# Using dplyr grouping
library(dplyr)
df |> group_by(group) |> scramble_variables("x")

Row-wise scrambling: Use .byrow = TRUE to shuffle values within each row across the selected columns. 

df_items <- data.frame(
  item1 = c(1, 4, 7),
  item2 = c(2, 5, 8), 
  item3 = c(3, 6, 9)
)

# Shuffles values horizontally within each row
scramble_variables(df_items, item1:item3, .byrow = TRUE)

Datasets

The vazul package includes two research datasets for demonstration and practice.

MARP Dataset

The marp dataset contains cross-national survey data on religiosity: 10,535 participants across 24 countries studying religiosity and well-being (Hoogeveen, Sarafoglou, van Elk, & Wagenmakers, 2022a).

Williams Dataset

The williams dataset contains experimental data from the large-scale direct replication of the stereotyping study by Williams, Sng, and Neuberg (2016), released by Holzmeister et al. (2024).

Dataset References

  • Hoogeveen, S., Sarafoglou, A., van Elk, M., & Wagenmakers, E.-J. (2022a). A many-analysts approach to the relation between religiosity and well-being: The dataset. PsyArXiv. https://doi.org/10.31234/osf.io/dpex6
  • Holzmeister, F., Camerer, C., Chen, Y., Dreber, A., Hoogeveen, S., Huber, J., … Waldén, V. (2024, November 19). Data and Analysis. Retrieved from osf.io/47drs
  • Williams, K. E. G., Sng, O., & Neuberg, S. L. (2016). Ecology-driven stereotypes override race stereotypes. Proceedings of the National Academy of Sciences, 113(2), 310–315. https://doi.org/10.1073/pnas.1519401113

Explanation of the package name

Vazul was a Hungarian prince in the 11. century. He was blinded by the king to become unfit for the throne. More info: https://en.wikipedia.org/wiki/Vazul

Documentation

Authors

  • Tamás Nagy - Package author and maintainer
  • Márton Kovács - Author
  • Alexandra Sarafoglou - Data contributor and author

Citation

Nagy, T., Kovács, M., & Sarafoglou, A. (2026). vazul: An R package for analysis blinding. Zenodo. https://doi.org/10.5281/zenodo.18269711

License

MIT License - see LICENSE file for details.