library(MatchIt)
data("lalonde", package = "MatchIt")
form <- treat ~ age + educ + race + married + nodegree + re74 + re75
ps_mod <- glm(form, data = lalonde, family = binomial)
broom::tidy(ps_mod) # coefficients, SE, p-values
exp(coef(ps_mod)) # odds ratios
summary(margins::margins(ps_mod)) # average marginal effects
ll_null <- -ps_mod$null.deviance / 2
1 - as.numeric(logLik(ps_mod)) / ll_null # McFadden pseudo R-squared
Computational Provenance & Reproducibility Record Propensity Score Matching · 2.0.0 · DOI 10.5281/zenodo.23013711
This Computational Provenance Record documents the statistical computing environment, software dependencies, computational provenance, and bibliographic references associated with the RAISINS Propensity Score Matching module. It is intended to support computational reproducibility and software transparency. Detailed statistical methodology, mathematical derivations, and user guidance are provided separately in the official module documentation.
Any issues or updates required please comment here
Go to the App from here
1 Module Metadata
| Parameter | Specification |
|---|---|
| Module | Propensity Score Matching |
| Module Version | 2.0.0 |
| DOI | 10.5281/zenodo.23013711 |
| Document Type | Computational workflow |
| Statistical Engine | R |
| R Version | 4.5.2 |
| Reproducibility | Execution Environment: Posit Connect · GCR · renv-locked |
2 Statistical Dependency Manifest
| Package | Version | Repository | Core Statistical Functions |
|---|---|---|---|
| MatchIt | 4.7.2 |
CRAN | matchit(), match.data(), summary.matchit() |
| optmatch | 0.10.8 |
CRAN | Solver used by matchit(method = "optimal") |
| cobalt | 4.6.2 |
CRAN | love.plot(), bal.plot() |
| marginaleffects | 0.32.0 |
CRAN | avg_comparisons() |
| sandwich | 3.1-1 |
CRAN | Cluster-robust variance used by avg_comparisons(vcov = ~subclass) |
| margins | 0.3.28 |
CRAN | margins() |
| broom | 1.0.13 |
CRAN | tidy() |
| stats | 4.5.2 |
Base R | glm(), lm(), predict(), logLik(), pchisq(), t.test() |
3 Statistical Function Registry
| Analytical Role | Primary Function(s) |
|---|---|
| Propensity-score model | stats::glm(treatment ~ covariates, family = binomial) |
| Coefficients, odds ratios, marginal effects | broom::tidy(), margins::margins() |
| Model fit (log-likelihood, LR chi-square, McFadden R²) | stats::logLik(), stats::pchisq() |
| Matching | MatchIt::matchit(distance = "glm") |
| Matched sample | MatchIt::match.data() |
| Covariate balance before and after matching | MatchIt::summary.matchit() |
| Love plot and propensity-score balance plot | cobalt::love.plot(), cobalt::bal.plot() |
| Treatment effect on the treated (matched sample) | stats::lm() + marginaleffects::avg_comparisons() |
| Welch and paired t-tests (matched sample) | stats::t.test() |
| IPWRA and AIPW (full sample) | stats::glm(), stats::lm(); bootstrap standard errors |
4 Default Methods & Parameters
| Analysis | Step / Parameter | Default Method / Value |
|---|---|---|
| Propensity Score Model |
Model | Logistic regression of the binary treatment on the selected covariates, entered additively; the outcome is never used as a covariate |
| Treatment coding | Exactly two levels required; the second level (in sorted order) is the treated group | |
| Covariates | Qualitative covariates are treated as factors; quantitative covariates as numeric | |
| Matching | Method | Nearest-neighbour (default) or optimal, 1:1, without replacement, distance = "glm" |
| Caliper | Optional, 0.1–0.5 standard deviations of the propensity score; nearest-neighbour matching only | |
| Balance | Standardised mean difference, variance ratio and eCDF statistics before and after matching. |SMD| < 0.1 well balanced; 0.1–0.25 borderline; > 0.25 imbalanced | |
| Diagnostics | Before vs after | McFadden pseudo R², LR chi-square and mean standardised bias, from the original model and from the same model refitted on the matched sample |
| Treatment Effect |
Matched-sample ATT | Linear outcome model with treatment × covariate interactions on the matched sample; ATT from avg_comparisons() over the treated units, with standard errors clustered by matched pair |
| t-tests | Welch two-sample t-test and paired t-test of matched-pair differences, on the matched sample | |
| IPWRA and AIPW | ATE and ATT on the full sample, as robustness checks; propensity scores bounded away from 0 and 1; standard errors from 500 bootstrap resamples, stratified by treatment, with a fixed seed |
5 R Code for Key Analytical Steps
The code blocks below demonstrate the computation behind each reported result using the lalonde dataset supplied with the MatchIt package.
5.1 Propensity-Score Model
5.2 Matching and Covariate Balance
m.out <- matchit(form, data = lalonde,
method = "nearest", # or "optimal"
distance = "glm", ratio = 1, replace = FALSE,
caliper = 0.2) # optional; nearest-neighbour only
summary(m.out)$nn # matched / unmatched counts
summary(m.out)$sum.all # balance before matching
summary(m.out)$sum.matched # balance after matching
cobalt::love.plot(m.out, stars = "std", abs = TRUE)
cobalt::bal.plot(m.out, var.name = "distance", which = "both")5.3 Treatment Effect on the Matched Sample
library(marginaleffects)
m.dat <- match.data(m.out)
fit <- lm(re78 ~ treat * (age + educ + race + married + nodegree + re74 + re75),
data = m.dat, weights = weights)
avg_comparisons(fit, variables = "treat",
newdata = subset(m.dat, treat == 1),
wts = "weights", vcov = ~subclass) # ATT
t.test(re78 ~ treat, data = m.dat) # Welch t-test5.4 IPWRA and AIPW (Full Sample)
X <- lalonde[, c("age", "educ", "race", "married", "nodegree", "re74", "re75")]
A <- lalonde$treat; Y <- lalonde$re78
e <- fitted(glm(A ~ ., data = cbind(A = A, X), family = binomial))
# IPWRA, ATT: treated weighted 1, controls weighted e/(1-e)
f1 <- lm(Y ~ ., data = cbind(Y = Y, X)[A == 1, ])
f0 <- lm(Y ~ ., data = cbind(Y = Y, X)[A == 0, ], weights = (e / (1 - e))[A == 0])
mean((predict(f1, X) - predict(f0, X))[A == 1])
# AIPW, ATE
m1 <- predict(lm(Y ~ ., data = cbind(Y = Y, X)[A == 1, ]), X)
m0 <- predict(lm(Y ~ ., data = cbind(Y = Y, X)[A == 0, ]), X)
mean(A * (Y - m1) / e - (1 - A) * (Y - m0) / (1 - e) + (m1 - m0))Explore the entire Propensity Score Matching module in preview mode using our demo datasets. To submit suggestions or report a workflow issue, please use the discussion section below, or visit the official RAISINS website.
6 RAISINS Native Statistical Framework
RAISINS uses R for all its statistical computations. Every package used to generate major results is listed and demonstrated with examples, so results can be reproduced independently. These results are then organized and formatted on the RAISINS website along with visualisation to make them easier to use and interpret. RAISINS also has its own custom-built statistical tools for managing workflows, validating results, and generating reports. Details of these are not fully covered here, they’re shared with outside researchers only on request, and are subject to licensing terms.
7 Package References
R Core Team. (2025). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/
Ho, D., Imai, K., King, G., & Stuart, E. A. (2011). MatchIt: Nonparametric Preprocessing for Parametric Causal Inference. Journal of Statistical Software, 42(8), 1-28. https://doi.org/10.18637/jss.v042.i08
Hansen, B. B., & Klopfer, S. O. (2006). Optimal full matching and related designs via network flows. Journal of Computational and Graphical Statistics, 15(3), 609-627. https://doi.org/10.1198/106186006X137047
Greifer, N. (2024). cobalt: Covariate Balance Tables and Plots (R package version 4.6.2). https://doi.org/10.32614/CRAN.package.cobalt
Arel-Bundock, V., Greifer, N., & Heiss, A. (2024). How to Interpret Statistical Models Using marginaleffects for R and Python. Journal of Statistical Software, 111(9), 1-32. https://doi.org/10.18637/jss.v111.i09
Zeileis, A., Köll, S., & Graham, N. (2020). Various Versatile Variances: An Object-Oriented Implementation of Clustered Covariances in R. Journal of Statistical Software, 95(1), 1-36. https://doi.org/10.18637/jss.v095.i01
Leeper, T. J. (2024). margins: Marginal Effects for Model Objects (R package version 0.3.28). https://doi.org/10.32614/CRAN.package.margins
Robinson, D., Hayes, A., & Couch, S. (2025). broom: Convert Statistical Objects into Tidy Tibbles (R package version 1.0.13). https://CRAN.R-project.org/package=broom
Rosenbaum, P. R., & Rubin, D. B. (1983). The Central Role of the Propensity Score in Observational Studies for Causal Effects. Biometrika, 70(1), 41-55. https://doi.org/10.1093/biomet/70.1.41
Austin, P. C. (2011). An Introduction to Propensity Score Methods for Reducing the Effects of Confounding in Observational Studies. Multivariate Behavioral Research, 46(3), 399-424. https://doi.org/10.1080/00273171.2011.568786
Robins, J. M., Rotnitzky, A., & Zhao, L. P. (1994). Estimation of Regression Coefficients When Some Regressors Are Not Always Observed. Journal of the American Statistical Association, 89(427), 846-866. https://doi.org/10.1080/01621459.1994.10476818
Allaire, J. J., Xie, Y., Dervieux, C., McPherson, J., Luraschi, J., Ushey, K., Atkins, A., Wickham, H., Cheng, J., Chang, W., & Iannone, R. (2026). rmarkdown: Dynamic Documents for R (R package version 2.31). https://github.com/rstudio/rmarkdown