causalfragility-lab.r-universe.dev

[causalfragility-lab] NonlinearDiD 0.2.0

haitsubi@msu.edu (Subir Hait) — Thu, 21 May 2026 11:01:54 GMT

Supports staggered difference-in-differences designs with nonlinear outcomes for both panel and repeated cross-section data. Implements estimators for staggered treatment adoption with binary, count, and other nonlinear outcomes, extending Callaway and Sant'Anna (2021) to settings with nonlinear outcome models such as logit, probit, and Poisson. For panel data, units are followed over time and 'idname' identifies repeated observations. For repeated cross-section data, observations are independent within each time period; 'idname' is optional and may identify survey records or households, but the estimator does not require the same units to appear across periods. Repeated cross-section estimation includes pooled quasi-maximum likelihood approaches motivated by Wooldridge (2023) , with optional weighting and clustered inference. Methods also draw on Roth and Sant'Anna (2023) and Sant'Anna and Zhao (2020) .

[causalfragility-lab] AIGovernance 0.1.0

haitsubi@msu.edu (Subir Hait) — Tue, 19 May 2026 14:36:59 GMT

Provides statistical auditing, risk documentation, and reporting tools to support AI governance workflows for employment and hiring decision systems. Implements the EEOC four-fifths adverse impact rule (Equal Employment Opportunity Commission, 1978, ), NYC Local Law 144 bias audit requirements (New York City, 2023, ), and the AI Risk Management Framework checklist items from the National Institute of Standards and Technology (2023, ). Optionally supports EU AI Act high-risk classification (European Parliament and Council, 2024, ). The package does not provide legal advice or certify legal compliance; it is a statistical and documentation support tool.

[cran] rdstagger 0.1.0

haitsubi@msu.edu (Subir Hait) — Tue, 05 May 2026 20:37:06 GMT

Implements a unified framework combining staggered difference-in-differences with regression discontinuity designs and network interference. Extends Callaway and Sant'Anna (2021) to settings where treatment assignment is determined by a running variable crossing a cutoff, adoption timing is heterogeneous across units, and spillover effects operate through a known network structure. Provides group-time average treatment effects (direct and spillover), aggregation schemes, bandwidth selection, and pre-treatment falsification tests.

[causalfragility-lab] achieveGap 0.1.0

haitsubi@msu.edu (Subir Hait) — Tue, 05 May 2026 19:12:23 GMT

Implements a hierarchical penalized spline framework for estimating achievement gap trajectories in longitudinal educational data. The achievement gap between two groups (e.g., low versus high socioeconomic status) is modeled directly as a smooth function of grade while the baseline trajectory is estimated simultaneously within a mixed-effects model. Smoothing parameters are selected using restricted maximum likelihood (REML), and simultaneous confidence bands with correct joint coverage are constructed using posterior simulation. The package also includes functions for simulation-based benchmarking, visualization of gap trajectories, and hypothesis testing for global and grade-specific differences. The modeling framework builds on penalized spline methods (Eilers and Marx, 1996, ) and generalized additive modeling approaches (Wood, 2017, ), with uncertainty quantification following Marra and Wood (2012, ).

[causalfragility-lab] rdstagger 0.1.0

haitsubi@msu.edu (Subir Hait) — Sat, 02 May 2026 22:21:34 GMT

[causalfragility-lab] RobustFlow 0.1.1

haitsubi@msu.edu (Subir Hait) — Wed, 22 Apr 2026 12:43:09 GMT

Provides tools for constructing longitudinal decision paths, quantifying temporal drift, tracking subgroup disparity trajectories, and stress-testing longitudinal conclusions under hidden bias. Implements three signature metrics: the Drift Intensity Index (DII), which measures structural instability in transition dynamics using the Frobenius norm of consecutive transition matrix differences; the Bias Amplification Index (BAI), which quantifies whether group disparities widen or converge over time; and the Temporal Fragility Index (TFI), which estimates the minimum hidden-bias perturbation required to nullify a longitudinal trend conclusion. An interactive 'shiny' application supports exploratory analysis, visualization, and reproducible reporting. Methods are motivated by applications in educational and social science research, including the Early Childhood Longitudinal Study (ECLS). The DII is based on the Frobenius norm as described in Golub and Van Loan (2013, ISBN:9781421407944). The TFI extends the hidden-bias sensitivity framework of Rosenbaum (2002, ISBN:9781441912633). The BAI draws on disparity-trajectory methods discussed in Duncan and Murnane (2011, ISBN:9780871542731).

[causalfragility-lab] EpiNova 0.1.0

haitsubi@msu.edu (Subir Hait) — Tue, 21 Apr 2026 22:47:01 GMT

An extended epidemiological modelling framework that goes beyond the classical SIR (Susceptible-Infectious-Recovered) model. Supports SEIR (Susceptible-Exposed-Infectious-Recovered), SEIRD (Susceptible-Exposed-Infectious-Recovered-Deceased), SVEIRD (Susceptible-Vaccinated-Exposed-Infectious-Recovered-Deceased), and age-stratified compartmental models with flexible intervention functions (spline-based, Gaussian process, or user-defined). Inference is available via maximum likelihood or sequential Monte Carlo (SMC, also known as particle filtering) with no external binary dependencies. Includes a dependency-free real-time effective reproduction number (Rt) estimator, spatial multi-patch models with gravity-model mobility, ensemble forecasting via Bayesian model averaging (BMA), and proper scoring rules including CRPS (Continuous Ranked Probability Score), coverage, and MAE (Mean Absolute Error) for forecast evaluation. Methods follow Anderson and May (1991, ISBN:9780198545996), Doucet, de Freitas, and Gordon (2001) , Cori et al. (2013) , and Gneiting and Raftery (2007) .

[causalfragility-lab] hlmLab 0.1.0

haitsubi@msu.edu (Subir Hait) — Tue, 21 Apr 2026 04:07:56 GMT

Provides functions for visualization and decomposition in hierarchical linear models (HLM) for applications in education, psychology, and the social sciences. Includes variance decomposition for two-level and three-level data structures following Snijders and Bosker (2012, ISBN:9781849202015), intraclass correlation (ICC) estimation and design effect computation as described in Shrout and Fleiss (1979) , and contextual effect decomposition via the Mundlak (1978) specification distinguishing within- and between-cluster components. Supports visualization of random slopes and cross-level interactions following Hofmann and Gavin (1998) and Hamaker and Muthen (2020) . Multilevel models are estimated using 'lme4' (Bates et al., 2015 ). An optional 'Shiny' application enables interactive exploration of model components and parameter variation. The implementation follows the multilevel modeling framework of Raudenbush and Bryk (2002, ISBN:9780761919049).

[causalfragility-lab] aiDIF 0.1.0

haitsubi@msu.edu (Subir Hait) — Mon, 20 Apr 2026 20:41:21 GMT

Detects and quantifies differential item functioning (DIF) in AI-scored educational and psychological assessments. Provides a fully self-contained robust DIF engine (M-estimation via iteratively re-weighted least squares with the bi-square loss) alongside the novel Differential AI Scoring Bias (DASB) test, which detects item-level scoring shifts that differ across subgroups when comparing human and AI scoring conditions. Includes simulation utilities, anchor weight diagnostics, and an AI-effect classification framework.

[causalfragility-lab] socialdrift 0.1.0

haitsubi@msu.edu (Subir Hait) — Sun, 19 Apr 2026 03:36:59 GMT

Tools for constructing, auditing, and visualizing temporal social interaction networks from event-log data. Supports graph construction from raw user-to-user interaction logs, longitudinal tracking of network structure, community dynamics, user role trajectories, and concentration of engagement over time. Designed for computational social science, platform analytics, and digital community health monitoring. Includes four longitudinal audit indices: the Network Drift Index ('NDI'), Community Fragmentation Index ('CFI'), Visibility Concentration Index ('VCI'), and Role Mobility Index ('RMI'). 'NDI', 'CFI', 'VCI', and 'RMI' are purpose-built composite scores for longitudinal platform auditing.

[causalfragility-lab] MultiSpline 0.2.0

haitsubi@msu.edu (Subir Hait) — Thu, 16 Apr 2026 04:04:47 GMT

Provides a unified framework for fitting, predicting, and interpreting nonlinear relationships in single-level, multilevel, and longitudinal regression models. Flexible functional forms are supported using natural cubic splines ('splines'), B-splines ('splines'), and GAM smooths ('mgcv'). Supports two-way and nested clustering via 'lme4', automatic knot selection by AIC or BIC, multilevel R-squared decomposition (Nakagawa-Schielzeth marginal and conditional R-squared with level-specific variance partitioning), a postestimation suite returning first and second derivatives with confidence bands, turning points and inflection regions, and a model comparison workflow contrasting linear, polynomial, and spline fits by AIC, BIC, and likelihood-ratio tests. Cluster heterogeneity in nonlinear effects is supported via random-slope spline terms.

[causalfragility-lab] DecisionDrift 0.1.0

haitsubi@msu.edu (Subir Hait) — Sun, 12 Apr 2026 21:56:38 GMT

Tools for detecting, decomposing, and stress-testing temporal drift in repeated binary decision systems. Complements the 'decisionpaths' package by shifting focus from path construction to system-level change over time. Implements five core analytic modules: (1) prevalence drift — did the overall decision rate change over time?; (2) transition drift — did the probability of switching or persisting change?; (3) entropy and stability trends — did path complexity evolve?; (4) group-differential drift — did the system drift differently across subgroups?; (5) change-point and regime-shift detection — did the system change abruptly after a policy or model update? Additionally provides a robustness module for testing stability of drift conclusions across analytic choices, and a sensitivity module for probing vulnerability to data problems including missingness, miscoding, and threshold shifts. Defines four original drift indices: the Decision Drift Index (DDI), Transition Drift Index (TDI), Group Differential Drift (GDD), and Cumulative Drift Burden (CDB). Applications include algorithmic audit, AI governance, education, health, and organisational research.

[causalfragility-lab] statAPA 0.1.0

haitsubi@msu.edu (Subir Hait) — Sat, 11 Apr 2026 19:58:04 GMT

Produces publication-ready statistical tables and figures formatted according to the 7th edition of the American Psychological Association (APA) style guidelines. Supports descriptive statistics, t-tests, z-tests, chi-square tests, Analysis of Variance (ANOVA), Analysis of Covariance (ANCOVA), two-way ANOVA with simple effects, Multivariate Analysis of Variance (MANOVA), robust and cluster-robust regression using Heteroscedasticity-Consistent (HC) standard errors, post-hoc pairwise comparisons, homoskedasticity and heteroscedasticity diagnostics including the Non-Constant Variance (NCV) test, proportion tests, and multilevel mixed-effects models with intraclass correlation coefficients (ICC) and model-comparison tables. Output can be directed to the console, Microsoft Word (via 'officer' and 'flextable'), or LaTeX. For APA style guidelines see American Psychological Association (2020, ISBN:978-1-4338-3216-1).

[causalfragility-lab] RobustMediate 0.1.1

haitsubi@msu.edu (Subir Hait) — Thu, 09 Apr 2026 16:48:46 GMT

Provides tools for causal mediation analysis with continuous treatments using inverse probability weighting (IPW). Estimates natural direct and indirect effects over a user-defined treatment grid and supports flexible dose-response mediation analysis. Includes diagnostic procedures for assessing covariate balance in both treatment and mediator models using standardized mean differences. Implements pathway-specific extensions of the impact threshold for a confounding variable (ITCV; Frank, 2000 ) adapted to mediation settings. Provides joint sensitivity analysis combining E-values (VanderWeele and Ding, 2017 ) and violations of sequential ignorability (Imai, Keele, and Yamamoto, 2010 ). Additional utilities include visualization of dose-response mediation functions, robustness profiles, fragility summaries, and formatted outputs for applied research. Supports clustered data structures and multiple outcome families.

[causalfragility-lab] MLCausal 0.1.0

haitsubi@msu.edu (Subir Hait) — Wed, 08 Apr 2026 20:16:35 GMT

Provides an end-to-end workflow for estimating average treatment effects in clustered (multilevel) observational data. Core functionality includes cluster-aware propensity score estimation using fixed effects and Mundlak-style specifications, inverse probability weighting, within-cluster nearest-neighbor matching, covariate balance diagnostics at both individual and cluster-mean levels, outcome regression with cluster-robust standard errors, propensity score overlap visualization, and tipping-point sensitivity analysis for omitted cluster-level confounding.

[causalfragility-lab] AIBias 0.1.1

haitsubi@msu.edu (Subir Hait) — Sat, 04 Apr 2026 19:12:05 GMT

Provides tools for detecting, quantifying, and visualizing algorithmic bias as a longitudinal process in repeated decision systems. Existing fairness metrics treat bias as a single-period snapshot; this package operationalizes the view that bias in sequential systems must be measured over time. Implements group-specific decision-rate trajectories, standardized disparity measures analogous to the standardized mean difference (Cohen, 1988, ISBN:0-8058-0283-5), cumulative bias burden, Markov-based transition disparity (recovery and retention gaps), and a dynamic amplification index that quantifies whether prior decisions compound current group inequality. The amplification framework extends longitudinal causal inference ideas from Robins (1986) and the sequential decision-process perspective in the fairness literature (see ) to the audit setting. Covariate-adjusted trajectories are estimated via logistic regression, generalized additive models (Wood, 2017, ), or generalized linear mixed models (Bates, 2015, ). Uncertainty quantification uses the cluster bootstrap (Cameron, 2008, ).

[causalfragility-lab] mlmoderator 0.2.1

haitsubi@msu.edu (Subir Hait) — Fri, 03 Apr 2026 22:46:26 GMT

Provides a unified workflow for probing, plotting, and assessing the robustness of cross-level interaction effects in two-level mixed-effects models fitted with 'lme4' (Bates et al., 2015) . Implements simple slopes analysis following Aiken and West (1991, ISBN:9780761907121), Johnson-Neyman intervals following Johnson and Fay (1950) and Bauer and Curran (2005) , and grand- or group-mean centering as described in Enders and Tofighi (2007) . Includes a slope variance decomposition that separates fixed-effect uncertainty from random-slope variance (tau11), a contour surface plot of predicted outcomes over the full predictor-by-moderator space, and robustness diagnostics comprising intraclass correlation coefficient shift analysis and leave-one-cluster-out (LOCO) stability checks. Designed for researchers in education, psychology, biostatistics, epidemiology, organizational science, and other fields where outcomes are clustered within higher-level units.

[causalfragility-lab] drmeta 0.1.0

haitsubi@msu.edu (Subir Hait) — Wed, 01 Apr 2026 22:46:53 GMT

Implements Design-Robust Meta-Analysis (DR-Meta), a variance-function random-effects framework in which between-study heterogeneity is modelled as a function of a study-level design robustness index, allowing heterogeneity to depend systematically on study quality or design strength rather than being treated as a single nuisance parameter. The package provides profiled restricted maximum likelihood (REML) estimation of the overall effect and variance-function parameters, study-specific weights, heterogeneity diagnostics (tau-squared, I-squared), influence and leave-one-out analysis, and graphical tools including forest plots and influence plots. The DR-Meta framework nests classical fixed-effects and standard random-effects meta-analysis as special cases, making it a strict generalisation of existing approaches.

[causalfragility-lab] CEDMr 0.1.0

haitsubi@msu.edu (Subir Hait) — Wed, 01 Apr 2026 14:24:27 GMT

Provides tools for implementing the Capability-Ecological Developmental Model (CEDM) in longitudinal and multilevel data. The package supports estimation and interpretation of models examining how socioeconomic status (SES), health indicators, and contextual factors jointly relate to academic outcomes. Functionality includes: (1) classification of ecological capability regimes (amplifying, neutral, compensatory); (2) estimation of moderated multilevel models with higher-order interaction terms; (3) causal mediation analysis using doubly robust estimation; (4) random-effects within-between (REWB) decomposition; (5) nonlinear moderation using restricted cubic splines; (6) clustering of longitudinal health trajectories; and (7) sensitivity analysis using the impact threshold for a confounding variable (ITCV) and robustness-to-replacement (RIR) measures. The package is designed for use with general longitudinal multilevel datasets.

[causalfragility-lab] confoundvis 0.1.0

haitsubi@msu.edu (Subir Hait) — Mon, 30 Mar 2026 16:52:37 GMT

Provides visualization tools for sensitivity analysis to unmeasured confounding in observational studies. Includes contour-based sensitivity plots, robustness curves, and benchmark-oriented graphics that help researchers assess how strong omitted confounding would need to be to attenuate, invalidate, or reverse estimated effects. Supports regression-based sensitivity analysis frameworks, including impact threshold approaches (Frank, 2000, ), partial R-squared methods (Cinelli and Hazlett, 2020, ), and E-value style metrics (VanderWeele and Ding, 2017, ). Emphasizes clear, interpretable, and publication-ready graphical summaries for transparent reporting of causal sensitivity analyses across the social, behavioral, health, and educational sciences.

[causalfragility-lab] metaLong 0.1.0

haitsubi@msu.edu (Subir Hait) — Wed, 25 Mar 2026 19:47:03 GMT

Tools for longitudinal meta-analysis where studies contribute effect sizes at multiple follow-up time points. Implements robust variance estimation (RVE) with Tipton small-sample corrections following Hedges, Tipton, and Johnson (2010) and Tipton (2015) , time-varying sensitivity analysis via the Impact Threshold for a Confounding Variable (ITCV) following Frank (2000) , benchmark calibration of the ITCV threshold against observed study-level covariates, spline-based nonlinear time-trend modeling with a nonlinearity test, and leave-k-out fragility analysis across the follow-up trajectory. Designed for researchers synthesising evidence from studies with repeated outcome measurement in education, psychology, health, and the social sciences.

[causalfragility-lab] CausalSpline 0.1.0

haitsubi@msu.edu (Subir Hait) — Sat, 21 Mar 2026 12:37:42 GMT

Estimates nonlinear causal dose-response functions for continuous treatments using spline-based methods under standard causal assumptions (unconfoundedness / ignorability). Implements three identification strategies: Inverse Probability Weighting (IPW) via the generalised propensity score (GPS), G-computation (outcome regression), and a doubly-robust combination. Natural cubic splines and B-splines are supported for both the exposure-response curve f(T) and the propensity nuisance model. Pointwise confidence bands are obtained via the sandwich estimator or nonparametric bootstrap. Also provides fragility diagnostics including pointwise curvature-based fragility, uncertainty-normalised fragility, and regional integration over user-defined treatment intervals. Builds on the framework of Hirano and Imbens (2004) for continuous treatments and extends it to fully nonparametric spline estimation.

[causalfragility-lab] decisionpaths 0.1.0

haitsubi@msu.edu (Subir Hait) — Sat, 14 Mar 2026 15:39:13 GMT

Tools for constructing and auditing longitudinal decision paths from panel data. Implements a decision infrastructure framework for representing institutional AI systems as generators of time-ordered binary decision sequences. Provides functions to build path objects from panel data, summarise per-unit descriptors (dosage, switching rate, onset, duration, longest run), compute the Decision Reliability Index (DRI) following Cronbach (1951) , estimate Shannon decision-path entropy following Shannon (1948) , classify systems by infrastructure type (static, periodic, continuous, human-in-the-loop), and evaluate subgroup disparities in decision exposure and stability. Applications include education, policy, health, and organisational research.