dat <- read.csv("Organic_Adoption.csv")
dat$Education <- as.factor(dat$Education)
form <- Adoption ~ Farm_Size_ha + Rainfall_mm + Farmer_Age +
Education + Extension_Contact + Credit_Access
ps_mod <- glm(form, data = dat, family = binomial)
Computational Provenance & Reproducibility Record Propensity Score Matching · 2.0.0 · DOI pending
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 | pending |
| 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() |
| cobalt | 4.6.2 |
CRAN | love.plot(), bal.plot(), bal.tab(), summary.matchit() |
| broom | 1.0.13 |
CRAN | tidy() |
| margins | 0.3.28 |
CRAN | margins() |
| optmatch | 0.10.8 |
CRAN | backend for method = "optimal"/"full" matching |
| stats | 4.5.2 |
Base R | glm(), predict(), logLik(), pchisq() |
3 Statistical Function Registry
| Analytical Role | Primary Function(s) |
|---|---|
| Propensity-score estimation | stats::glm(treatment ~ covariates, family = binomial) |
| Matching | MatchIt::matchit(distance = "glm", ...) |
| Matched analytic sample | MatchIt::match.data() |
| Covariate balance tables (before/after) | summary.matchit()$sum.all, summary.matchit()$sum.matched (via cobalt) |
| Standardized mean difference (Love) plot | cobalt::love.plot() |
| Covariate density / eCDF overlap plot | cobalt::bal.plot() |
| Propensity-score coefficient table | broom::tidy() on the fitted glm() object |
| Average marginal effects | margins::margins() |
4 Default Methods & Parameters
| Analysis | Step / Parameter | Default Method / Value |
|---|---|---|
| Propensity Score Model |
Method | Logistic regression of the binary treatment indicator on the selected covariates, via stats::glm(family = binomial) |
| Predictors | All user-selected qualitative and quantitative covariates entered additively (no interaction terms by default); the outcome variable is never included as a predictor | |
| Distance measure | distance = "glm": the fitted propensity score itself (predicted probability of treatment) is used as the distance metric for matching |
|
| Variable requirements | Treatment column must have exactly two levels; rows with a covariate that is constant (a single level) across the whole sample cause the model fit to be rejected with an on-screen error | |
| Matching | Method | User-selectable: nearest-neighbor (default), optimal, full, subclassification, exact, or coarsened exact matching, via MatchIt::matchit() |
| Ratio | 1:1 matching (one control matched to each treated unit) | |
| Caliper | Optional, user-specified between 0.1 and 0.5 (in standard-deviation units of the propensity score); when set, a treated unit is left unmatched rather than matched to a control outside the caliper. No caliper is applied if left blank | |
| Balance Assessment |
Metric | Standardized mean difference (SMD) of each covariate between treated and control groups, computed before and after matching via summary.matchit() |
| Practical thresholds | |SMD| < 0.1: well balanced; 0.1-0.25: borderline; > 0.25: meaningfully imbalanced (the RA-One assistant and the app narrative apply these same thresholds) | |
| Model Diagnostics |
Model fit | Log-likelihood, likelihood-ratio chi-square test, and McFadden's pseudo R-squared, reported for the propensity-score model both before matching (original fit) and after matching (the same model refit, weighted, on the matched sample) |
| Mean standardized bias | Mean of the absolute standardized mean differences across all covariates, reported before and after matching as a single summary diagnostic | |
| Plots | Before/After distribution | Kernel density of the estimated propensity score by treatment group, before matching (from the fitted model) and after matching (on match.data()) |
| Love plot, jitter plot, balance plot | Rendered directly from the MatchIt match object via cobalt::love.plot(), plot(m.out, type = "jitter"), and cobalt::bal.plot() respectively - no statistics are recomputed outside these functions |
5 R Code for Key Analytical Steps
The code blocks below demonstrate the exact computation behind each reported result, using the built-in Organic_Adoption dataset bundled with the module (a farm-level observational dataset on organic-farming adoption, covariates, and yield).
5.1 Propensity-Score Model
5.2 Matching
library(MatchIt)
m.out <- matchit(
form, data = dat,
method = "nearest", # user-selectable
distance = "glm",
ratio = 1,
caliper = 0.2, # optional, user-specified
link = "logit",
replace = FALSE
)
summary(m.out)$nn # matched / unmatched / discarded counts5.3 Covariate Balance (Standardized Mean Difference)
library(cobalt)
bal <- summary(m.out)
bal$sum.all[, "Std. Mean Diff."] # before matching
bal$sum.matched[, "Std. Mean Diff."] # after matching
love.plot(m.out, stars = "std", abs = TRUE) # Love plot
bal.plot(m.out, var.name = "distance", which = "both") # balance plot5.4 Propensity-Score Coefficient Table & Marginal Effects
library(broom); library(margins)
broom::tidy(ps_mod) # coefficient, SE, p-value, odds ratio = exp(estimate)
summary(margins::margins(ps_mod)) # average marginal effect of each covariateExplore 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. (2011). MatchIt: Nonparametric Preprocessing for Parametric Causal Inference. Journal of Statistical Software, 42(8), 1-28. https://doi.org/10.18637/jss.v042.i08
Greifer, N. (2024). cobalt: Covariate Balance Tables and Plots (R package version 4.6.2). https://CRAN.R-project.org/package=cobalt
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