Inference for Released Plug-in Sampling Synthetic Dataset

Considering the singly imputed synthetic data generated via plug-in sampling under the multivariate normal model, draws inference procedures including the generalized variance, the sphericity test, the test for independence between two subsets of variables, and the test for the regression of one set of variables on the other. For more details see Klein et al. (2021) .

PSinference

PSinference provides exact finite-sample inferential procedures for singly and multiply released plug-in sampling (PS) synthetic datasets under a multivariate normal model. The key insight is simple: an analyst who receives $M$ independent synthetic datasets $V_1, \ldots, V_M$ of size $n$ naturally treats all released data as a whole by stacking them into a single dataset of size $Mn$. This stacking is statistically justified — the $Mn$ rows are conditionally i.i.d. given the original data — and immediately extends the exact procedures of Klein et al. (2021) to arbitrary $M \geq 1$ via the substitution $n \to Mn$.

This work was supported by the Fundação para a Ciência e a Tecnologia (FCT, Portugal) under projects UID/00297/2025 and UID/PRR/00297/2025 (NOVAMath).

Installation

You can install the stable version from CRAN.

install.packages('PSinference', dependencies = TRUE)

You can install the development version from Github

# install.packages("remotes")
remotes::install_github("ricardomourarpm/PSinference")

Quick Start

r library(PSinference) data(brittany_soil_ps)

Generate 5 synthetic releases (stacked)

set.seed(42) V <- simSynthData(brittany_soil_ps, M = 5)

Sphericity test

res <- sphericity_test(V, M = 5) print(res) plot(res)

Or use the unified wrapper

ps_test(V, M = 5, test = "independence", part = 4L)

M = 1 recovers Klein et al. (2021)

V1 <- simSynthData(brittany_soil_ps) ps_test(V1, M = 1, test = "sphericity")

To cite package PSinference in publications use:

Augusto V, Norouzirad M, Fonseca M, Moura R (202). PSinference: Inference for Released Plug-in Sampling Synthetic Dataset. R package version 1.0.0, https://cran.r-project.org/package=PSinference.

A BibTeX entry for LaTeX users is

@Manual{PSinference, title = {PSinference: Inference for Released Plug-in Sampling Synthetic Dataset}, author = {Vítor Augusto and Mina Norouzirad and Miguel Fonseca and Ricardo Moura}, year = {2026}, note = {R package version 1.0.0}, url = {https://cran.r-project.org/package=PSinference} }

References

Klein, M., Moura, R., and Sinha, B. (2021). Multivariate normal inference based on singly imputed synthetic data under plug-in sampling. Sankhya B, 83, 273--287.

License

This package is free and open source software, licensed under GPL-3.