Package: regsensitivity 0.2.0

Mike Nguyen

regsensitivity: Regression Sensitivity Analysis for Omitted Variable Bias

Sensitivity analysis for linear regression coefficients in the presence of omitted variables. Implements the identified-set and breakdown-point analyses of Diegert, Masten and Poirier (2026) <doi:10.48550/arXiv.2206.02303>, Oster (2019) <doi:10.1080/07350015.2016.1227711> and Masten and Poirier (2026) <doi:10.1257/aer.20230242>. Provides bounds on regression coefficients under relaxations of the no-omitted-variables assumption indexed by interpretable sensitivity parameters, the breakdown point of a hypothesis about the long-regression coefficient, and plotting helpers.

Authors:Mike Nguyen [aut, cre, cph]

regsensitivity_0.2.0.tar.gz
regsensitivity_0.2.0.zip(r-4.7-any)regsensitivity_0.2.0.zip(r-4.6-any)
regsensitivity_0.2.0.tar.gz(r-4.7-any)regsensitivity_0.2.0.tar.gz(r-4.6-any)
regsensitivity_0.2.0.tgz(r-4.6-emscripten)
manual.pdf |manual.html✨
DESCRIPTION |NEWS
card.svg |card.png
regsensitivity/json (API)

# Install 'regsensitivity' in R:
install.packages('regsensitivity', repos = c('https://cran.r-universe.dev', 'https://cloud.r-project.org'))

Bug tracker:https://github.com/mikenguyen13/regsensitivity/issues

Pkgdown/docs site:https://mikenguyen13.github.io

Datasets:
  • bfg2020 - Bazzi, Fiszbein and Gebresilasse (2020) "Frontier Culture" replication data

On CRAN:

Conda:

3.48 score 20 exports 18 dependencies

Last updated from:7b8985efb6. Checks:6 OK. Indexed: yes.

TargetResultTimeFilesSyslog
linux-develOK220
source / vignettesOK342
linux-releaseOK215
windows-develOK114
windows-releaseOK159
wasm-releaseOK124

Exports:bnd_eqbnd_lbbnd_ubcalibrate_partial_r2calibrate_rhoregsen_bootregsen_boundsregsen_breakdownregsen_coresregsen_exploreregsen_glanceregsen_multiregsen_summaryregsen_tableregsen_tidyregsensitivityscale_color_regsenscale_colour_regsenscale_fill_regsentheme_regsen

Dependencies:clicpp11farverggplot2gluegtableisobandlabelinglifecyclenloptrR6RColorBrewerrlangS7scalesvctrsviridisLitewithr

Coming from the Stata package
Command map | The Stata vignette, translated | Default summary | Bounds at a single cbar | Bounds over a grid of cbar | Restricting rxbar as well | Breakdown frontier | Breakdown for a one-sided hypothesis | Bounding the unobservable's effect on the outcome | Oster (2019) | Oster breakdown over rmax | Coverage | Beyond the Stata package
Rendered fromstata-migration.Rmdusingknitr::rmarkdown

Last update: 2026-10-20
Started: 2026-10-20

Publication-ready plots and tables
Figures | Reading an unbounded identified set | Reference lines | Sizing for a journal | Matching your manuscript's font | Colour, greyscale and accessibility | Overriding anything | Tables | The tidy route | The convenience route | LaTeX for a journal | Infinite bounds | Feeding modelsummary
Rendered fromplots-and-tables.Rmdusingknitr::rmarkdown

Last update: 2026-10-20
Started: 2026-10-20

Reading the output: interpretation, edge cases and decisions
The numbers, and what each one means | The decision path | Case 1: a robust conclusion | Case 2: a fragile conclusion | Case 3: a breakdown point of zero | Case 4: the set is unbounded early -- a weak calibration set | Case 5: which covariates go in compare | Case 6: what cbar does, and when it matters | Case 7: bounding the effect on the outcome too | Case 8: Oster's $\delta$, and why a large one is not enough | Case 9: the bootstrap interval | Case 10: several treatments at once | Common mistakes | A reporting checklist | References
Rendered frominterpreting-results.Rmdusingknitr::rmarkdown

Last update: 2026-10-20
Started: 2026-10-20

regsensitivity: a tour
The setup | The three sensitivity parameters | DMP (2026) bounds | Constrained rybar | DMP breakdown | The frontier in the other direction | Asserting endogenous controls | Oster (2019) bounds | Oster, |delta| <= d (bound) | Oster breakdown | Using several cores | Summary call | References
Rendered fromregsensitivity.Rmdusingknitr::rmarkdown

Last update: 2026-10-20
Started: 2026-10-20

Replication: Diegert, Masten and Poirier (2026)
Table 1, Panel C, column (5): breakdown points | Table 3: correlations between observed covariates | Table 4, column (5): rho_k calibration | Figure 1: Sensitivity analysis for Republican Vote Share | Left panel: bounds on $\beta_{long}$ as a function of $\bar r_X$, $\bar c = 1$ | Right panel: the breakdown frontier in $(\bar r_X, \bar r_Y)$ space | Overlaying the calibration values | Figure 2: not reproducible from the bundled data | Figure 3: moving the state fixed effects into the calibration set | Bonus: identified set under $\bar r_Y < \infty$
Rendered fromdmp2022-replication.Rmdusingknitr::rmarkdown

Last update: 2026-10-20
Started: 2026-10-20

Stylized examples from Masten and Poirier (2026)
A stylized data-generating process | Figure 1: Oster's identified set $\mathcal{B}I(\delta, R{long})$ | Figure 2: cumulative identified set $\delta \le \bar d$ | Table 1: breakdown points | Effect of a maxovb constraint | Meta-analysis tables (paper Tables 3-4)
Rendered frommp2022-stylized.Rmdusingknitr::rmarkdown

Last update: 2026-10-20
Started: 2026-10-20