Course Introduction
The Regression Discontinuity Design and Impact Evaluation Training Course provides a rigorous and practical framework for estimating causal effects when program participation, policy eligibility, or treatment assignment is determined by a clearly defined cutoff. Participants learn how regression discontinuity designs exploit discontinuities around an assignment threshold to construct credible quasi-experimental comparisons and evaluate the localized impact of interventions.
The course introduces the theoretical foundations of Regression Discontinuity Design (RDD), including treatment assignment mechanisms, running variables, cutoff points, potential outcomes, local treatment effects, and counterfactual reasoning. Participants examine both sharp and fuzzy regression discontinuity designs and learn how institutional rules, eligibility criteria, test scores, income thresholds, geographic boundaries, and other assignment mechanisms can create opportunities for credible impact evaluation.
Participants develop practical skills in preparing and analyzing RDD datasets using appropriate graphical, statistical, and econometric techniques. The training covers bandwidth selection, polynomial specifications, local linear regression, kernel weighting, covariate adjustment, density testing, manipulation checks, and robust inference. Emphasis is placed on understanding how analytical choices influence estimated treatment effects and the validity of causal conclusions.
A major component of the course focuses on identification assumptions and threats to validity. Participants learn how to assess whether individuals or organizations can manipulate the assignment variable around the cutoff, whether other characteristics change discontinuously, and whether the outcome relationship is sufficiently smooth near the threshold. Practical diagnostic techniques help participants distinguish genuine treatment effects from artifacts caused by functional-form choices, bandwidth decisions, sorting, measurement error, or concurrent interventions.
The training also addresses advanced and emerging applications of RDD, including geographic regression discontinuity, spatial discontinuities, multiple cutoffs, discrete running variables, fuzzy assignment mechanisms, heterogeneous treatment effects, and modern robust bias-corrected inference. Participants explore how contemporary statistical software and computational workflows can support visualization, estimation, sensitivity analysis, reproducibility, and large-scale impact evaluations using administrative and observational data.
By the end of the course, participants will be able to design, implement, evaluate, and communicate Regression Discontinuity studies for policy, program, and institutional impact assessment. Participants will understand how to formulate appropriate research questions, prepare suitable datasets, select defensible estimation strategies, conduct validity and robustness checks, interpret local treatment effects, and communicate findings with appropriate attention to assumptions, uncertainty, limitations, and external validity.
Duration
5 days
Who Should Attend
-
Policy analysts involved in evaluating government policies, regulations, eligibility programs, and public interventions.
-
Monitoring and evaluation specialists conducting quantitative impact assessments of development and social programs.
-
Economists and econometricians seeking advanced training in quasi-experimental research designs and causal inference.
-
Data analysts and statisticians working with administrative, longitudinal, survey, education, health, or economic datasets.
-
Academic researchers and postgraduate students conducting empirical policy evaluation and applied econometric research.
-
Development professionals evaluating interventions where eligibility is determined by income, scores, age, geography, or other thresholds.
-
Researchers working with observational data where randomized controlled trials are impractical, costly, or ethically inappropriate.
-
Consultants providing impact evaluation, program assessment, economic research, or evidence-based policy analysis services.
-
Public-sector researchers analyzing policy effects associated with eligibility thresholds, regulatory requirements, or institutional rules.
-
Professionals working in education, healthcare, labor markets, social protection, taxation, public finance, and development evaluation.
-
Quantitative researchers using R, Stata, Python, or other statistical software for applied econometric analysis.
-
Managers and decision-makers who need to understand, interpret, and critically assess Regression Discontinuity impact evaluation findings.
Course Objectives
-
Develop a comprehensive understanding of Regression Discontinuity Design and its role in identifying causal effects from observational and administrative data.
-
Explain running variables, treatment assignment thresholds, potential outcomes, local treatment effects, counterfactuals, and the assumptions underlying RDD.
-
Distinguish sharp and fuzzy Regression Discontinuity Designs and determine when each approach is appropriate for specific policy evaluation settings.
-
Prepare and structure RDD datasets by correctly identifying cutoff values, treatment status, running variables, outcomes, covariates, and relevant analytical samples.
-
Apply appropriate graphical and econometric techniques to investigate discontinuities in outcomes and assess treatment effects around policy thresholds.
-
Select and justify appropriate bandwidths, functional forms, kernel specifications, polynomial orders, and estimation strategies for reliable RDD analysis.
-
Diagnose threats to identification including manipulation of the running variable, sorting around the cutoff, measurement error, discrete assignment variables, and concurrent changes.
-
Conduct robustness, sensitivity, placebo, density, covariate continuity, and specification checks to assess the credibility of estimated Regression Discontinuity effects.
-
Apply advanced RDD techniques to fuzzy designs, geographic discontinuities, multiple cutoffs, heterogeneous effects, and other complex impact evaluation settings.
-
Communicate RDD findings clearly through statistical outputs, visualizations, technical reports, policy briefs, and reproducible analytical workflows that acknowledge limitations and uncertainty.
Comprehensive Course Outline
Module 1: Foundations of Impact Evaluation and Causal Inference
-
Understanding causal inference, counterfactual outcomes, treatment effects, and the fundamental challenges of estimating program and policy impacts.
-
Comparing experimental and quasi-experimental evaluation strategies and identifying situations where Regression Discontinuity can provide credible causal evidence.
-
Translating policy eligibility rules, institutional thresholds, geographic boundaries, and assignment mechanisms into testable impact evaluation research questions.
-
Examining local treatment effects, internal validity, external validity, generalizability, and the role of research design in credible policy evaluation.
Module 2: Core Principles of Regression Discontinuity Design
-
Understanding the running variable, treatment assignment rule, cutoff point, treatment status, outcome variable, and local comparison underlying an RDD.
-
Examining the intuition behind comparing observations immediately above and below a predetermined treatment assignment threshold.
-
Distinguishing sharp and fuzzy RDD designs based on whether crossing the cutoff deterministically determines treatment participation.
-
Understanding the assumptions required for identification and interpreting RDD estimates as local causal effects around the treatment threshold.
Module 3: RDD Data Preparation and Exploratory Analysis
-
Structuring RDD datasets with running variables, treatment indicators, outcomes, covariates, cutoff values, and appropriate observation windows.
-
Conducting descriptive analysis to understand distributions, sample sizes, treatment assignment patterns, missing data, and observations around the threshold.
-
Creating informative scatterplots, binned means, fitted relationships, and discontinuity visualizations to explore potential treatment effects.
-
Identifying data quality problems including measurement error, duplicated records, inconsistent eligibility rules, missing observations, and unusual observations near the cutoff.
Module 4: Estimation Strategies for Regression Discontinuity
-
Applying local linear and local polynomial regression methods to estimate treatment effects around the assignment threshold.
-
Understanding bandwidth selection, kernel weighting, polynomial specifications, and the implications of using different functional forms.
-
Incorporating covariates appropriately while preserving the underlying identification strategy and avoiding adjustment for post-treatment variables.
-
Comparing alternative estimation specifications and interpreting coefficient estimates, confidence intervals, standard errors, and local treatment effects.
Module 5: Identification Assumptions and Validity Diagnostics
-
Assessing whether individuals, organizations, or institutions can manipulate the running variable to influence treatment eligibility around the cutoff.
-
Conducting density and sorting diagnostics to identify unusual changes in the distribution of observations immediately around the assignment threshold.
-
Testing continuity of predetermined covariates and other characteristics to investigate whether comparable units exist on both sides of the cutoff.
-
Evaluating whether other policies, institutional changes, eligibility rules, or external shocks occur simultaneously with the treatment threshold.
Module 6: Bandwidth Selection and Robust Statistical Inference
-
Understanding the bias-variance trade-off associated with narrow and wide bandwidth choices and their implications for RDD estimation.
-
Applying data-driven bandwidth selection methods while considering practical context, sample size, functional form, and local identification.
-
Understanding robust bias-corrected inference and appropriate confidence intervals for modern Regression Discontinuity estimation.
-
Conducting sensitivity analyses across bandwidths, kernels, polynomial specifications, and estimation procedures to evaluate result stability.
Module 7: Fuzzy, Geographic, and Advanced RDD Designs
-
Applying fuzzy Regression Discontinuity methods when crossing the threshold changes treatment probability rather than perfectly determining treatment receipt.
-
Understanding geographic Regression Discontinuity designs where treatment assignment changes across spatial boundaries, jurisdictions, or administrative regions.
-
Examining multiple-cutoff designs, discrete running variables, repeated thresholds, and institutional settings with complex treatment assignment mechanisms.
-
Addressing practical challenges involving imperfect compliance, clustered observations, spatial dependence, geographic sorting, and limited observations near boundaries.
Module 8: Heterogeneous and Dynamic Treatment Effects
-
Investigating whether treatment effects differ across demographic, socioeconomic, geographic, institutional, or baseline outcome characteristics.
-
Examining treatment effect heterogeneity around different portions of the running-variable distribution and across relevant subgroups.
-
Evaluating settings where policy effects develop over time and determining whether the available data support meaningful dynamic impact analysis.
-
Using interaction specifications and subgroup analyses carefully while considering statistical power, multiple comparisons, and the underlying identification assumptions.
Module 9: Emerging Methods, Technology, and Advanced Impact Evaluation
-
Exploring modern RDD estimators, robust inference procedures, automated bandwidth selection, and computational approaches for complex evaluation datasets.
-
Integrating RDD with complementary quasi-experimental approaches such as Difference-in-Differences, matching, synthetic control, and instrumental-variable methods.
-
Using statistical programming and reproducible workflows to automate diagnostics, generate evaluation figures, document analytical decisions, and manage large administrative datasets.
-
Examining responsible applications of machine learning and AI-assisted analytics for exploratory analysis, heterogeneity investigation, visualization, and reproducible policy evaluation.
Module 10: Applied RDD Impact Evaluation and Reporting
-
Developing a complete Regression Discontinuity evaluation from research question formulation and data preparation through estimation and validity assessment.
-
Applying RDD techniques to practical policy cases involving education, healthcare, taxation, social protection, labor markets, public services, and development programs.
-
Producing technical reports, policy briefs, statistical tables, visualizations, and presentations that clearly communicate local treatment effects and uncertainty.
-
Documenting assumptions, bandwidth choices, specifications, robustness checks, limitations, and external validity considerations for transparent and reproducible evaluation studies.
Training Approach
This course will be delivered by our skilled trainers who have vast knowledge and experience as expert professionals in the fields. The course is taught in English and through a mix of theory, practical activities, group discussion and case studies. Course manuals and additional training materials will be provided to the participants upon completion of the training.
Tailor-Made Course
This course can also be tailor-made to meet organization requirement. For further inquiries, please contact us on: Email: training@upskilldevelopment.com Tel: +254 721 331 808
Training Venue
The training will be held at our Upskill Training Centre. We also offer training for a group (at a discount of 10% to 50%) at requested location all over the world. The Onsite course fee covers the course tuition, training materials, two break refreshments, buffet lunch, airport transfers, Upskill gift package, and guided tour.
Visa application, travel expenses, dinners, accommodation, insurance, and other personal expenses are catered by the participant
Certification
Participants will be issued with Upskill certificate upon completion of this course.
Airport Pickup and Accommodation
Airport pickup and accommodation is arranged upon request. For booking contact our Training Coordinator through Email: training@upskilldevelopment.com, +254 721 331 808
Terms of Payment:
Unless otherwise agreed between the two parties’ payment of the course fee should be done 3 working days before commencement of the training so as to enable us to prepare better.
