RDD design and diagnostics checklist

Examine assignment, manipulation, continuity and other simultaneous schemes. Diagnostic tests alone do not prove identification assumptions.

When to use this worksheet and what completion involves Fictional completed example

EvalCompass Worksheet

RDD design and diagnostics checklist

Method: Regression Discontinuity Design

Examine assignment, manipulation, continuity and other simultaneous schemes. Diagnostic tests alone do not prove identification assumptions.

Basic information

Enter evaluation basics

Completed example for this section (fictional)

This fictional example is for explanation. Adapt it to your own program. Your entries will not change.

Program name: Score-based training assignment
Analyst: Analyst A (fictional)
Data sources: Rules, scores, receipt and use records (fictional)
Creation date: July 1, 2026 (fictional)

Reason for this entry: Assignment records are needed alongside outcomes.

Program name

Analyst

Data sources

Creation date

Writing tips and explanation for this section

About this section

Enter evaluation basics

Assignment records are needed alongside outcomes.

View the method explanation

Writing tips

Follow each field’s heading to record the program name, focus, period, and other details. Leave undecided matters clearly marked as unresolved.

Read the full fictional example

Identify cutoff rules

Specify rules determining intervention receipt

Completed example for this section (fictional)

This fictional example is for explanation. Adapt it to your own program. Your entries will not change.

Running variable and units: Pre-evaluation score in points
Cutoff value: 60 points
Intervention above or below cutoff: Training assigned at 60 or above
Exceptions, if any: No exceptions in illustration; verify in actual scheme
Sharp or fuzzy RDD: Sharp RDD arithmetic illustration

Reason for this entry: A convenient fictional scheme, not a real one.

Running variable and units

Cutoff value

Intervention above or below cutoff

Exceptions, if any

Sharp or fuzzy RDD

Writing tips and explanation for this section

About this section

Specify rules determining intervention receipt

A convenient fictional scheme, not a real one.

View the method explanation

Writing tips

Follow each field’s heading to record the program name, focus, period, and other details. Leave undecided matters clearly marked as unresolved.

Read the full fictional example

Identification-assumption assessment plan

Enter diagnostic and institutional/operational checks and limitations

Completed example for this section (fictional)

This fictional example is for explanation. Adapt it to your own program. Your entries will not change.

Scroll horizontally to view the table

Diagnostic/check	Method	Judgment criteria
Score manipulation and other institutional jumps	Check density, pre-traits and assignment rules	Investigate generation if anomalies arise; non-significance does not establish safety

Reason for this entry: Combine tests with institutional understanding.

Scroll horizontally to view the table

Diagnostic/check	Method	Judgment criteria

Writing tips and explanation for this section

About this section

Enter diagnostic and institutional/operational checks and limitations

Combine tests with institutional understanding.

View the method explanation

Writing tips

Keep the information in each row aligned. Scroll wide tables horizontally to see every column.

Read the full fictional example

Estimation settings

Enter analytical rationale

Completed example for this section (fictional)

This fictional example is for explanation. Adapt it to your own program. Your entries will not change.

Bandwidth selection, such as optimal-bandwidth algorithms: Consider data-driven bandwidth selection
Selected bandwidth: Not calculated
Polynomial order/local method and rationale: Local linear as initial candidate
Sample within bandwidth: Unknown; illustrated estimates do not establish sample count

Reason for this entry: Do not infer bandwidth or sample size from fictional outcomes.

Bandwidth selection, such as optimal-bandwidth algorithms

Selected bandwidth

Polynomial order/local method and rationale

Sample within bandwidth

Writing tips and explanation for this section

About this section

Enter analytical rationale

Do not infer bandwidth or sample size from fictional outcomes.

View the method explanation

Writing tips

Follow each field’s heading to record the program name, focus, period, and other details. Leave undecided matters clearly marked as unresolved.

Read the full fictional example

Design checklist

Completed example for this section (fictional)

This fictional example is for explanation. Adapt it to your own program. Your entries will not change.

☐ Not checked: Documented running variable/cutoff and checked exceptions and receipt probabilities
☐ Not checked: Examined manipulation with density diagnostics and recorded context/limits
☐ Not checked: Examined pre-covariate continuity, stating tests cannot prove assumptions
☐ Not checked: Recorded bandwidth selection and rationale
☐ Not checked: Checked stability across bandwidths
☐ Not checked: Recorded need/results for placebo diagnostics; non-significance alone does not establish validity
☑ Checked: Specified sharp/fuzzy estimand and cutoff locality

Reason for this entry: Only the policy to describe local effects is settled. Diagnostics and robustness remain undone.

Documented running variable/cutoff and checked exceptions and receipt probabilities
Examined manipulation with density diagnostics and recorded context/limits
Examined pre-covariate continuity, stating tests cannot prove assumptions
Recorded bandwidth selection and rationale
Checked stability across bandwidths
Recorded need/results for placebo diagnostics; non-significance alone does not establish validity
Specified sharp/fuzzy estimand and cutoff locality

Writing tips and explanation for this section

About this section

Only the policy to describe local effects is settled. Diagnostics and robustness remain undone.

View the method explanation

Writing tips

Read each item and check those you have confirmed. Do not use the fictional example’s checkbox states as findings about your own program.

Read the full fictional example

Before writing, check when to use the worksheet and what completion involves

RDD design and diagnostics checklist is a worksheet for applying Regression Discontinuity Design in meetings or individual work. For the overall duration of the method, use “Illustrative planning durations, not common standards: preparation 1–3 weeks, analysis/diagnostics 1–2 weeks” as a reference, adjusting for scale and resources. Fill in what you know provisionally and leave unresolved matters as questions to check next.

When this worksheet is useful

Eligibility determined by available numerical score, income or age cutoffs
Retrospective education selection, income-based social benefits or subsidy programs with clear cutoffs
Receipt or probability changes at a usable threshold

Participants and preparation

1–2 econometric/statistical specialists collaborating with authorized administrative/score-data managers

Bring the materials used for decisions, existing data, and differing stakeholder views to help complete the worksheet.

Suggested order

1. Basic informationEnter evaluation basics
2. Identify cutoff rulesSpecify rules determining intervention receipt
3. Identification-assumption assessment planEnter diagnostic and institutional/operational checks and limitations
4. Estimation settingsEnter analytical rationale
5. Design checklist

Signs the worksheet is ready for review

RDD plot: running variable/outcome scatter and local polynomial fits
Discontinuity estimates, standard errors and confidence intervals
Robustness: bandwidths, placebo cutoffs, etc.

Cautions for review

Check manipulation and other cutoff-linked schemes using institutional documents, implementation, distributions and pre-traits.
Do not choose bandwidth to fit findings. Show rationale, multiple settings and uncertainty.
Do not generalize local effects as whole-program effects.

Guides to read before starting

Read a completed fictional example (your entries will not change)

Fictional example; not actual results

A fictional scheme assigns training at scores of 60 or above to illustrate sharp RDD estimates.

Basic information

Program name: Score-based training assignment
Analyst: Analyst A (fictional)
Data sources: Rules, scores, receipt and use records (fictional)
Creation date: July 1, 2026 (fictional)

Reason for this entry: Assignment records are needed alongside outcomes.

Identify cutoff rules

Running variable and units: Pre-evaluation score in points
Cutoff value: 60 points
Intervention above or below cutoff: Training assigned at 60 or above
Exceptions, if any: No exceptions in illustration; verify in actual scheme
Sharp or fuzzy RDD: Sharp RDD arithmetic illustration

Reason for this entry: A convenient fictional scheme, not a real one.

Identification-assumption assessment plan

Scroll horizontally to view the table

Diagnostic/check	Method	Judgment criteria
Score manipulation and other institutional jumps	Check density, pre-traits and assignment rules	Investigate generation if anomalies arise; non-significance does not establish safety

Reason for this entry: Combine tests with institutional understanding.

Estimation settings

Bandwidth selection, such as optimal-bandwidth algorithms: Consider data-driven bandwidth selection
Selected bandwidth: Not calculated
Polynomial order/local method and rationale: Local linear as initial candidate
Sample within bandwidth: Unknown; illustrated estimates do not establish sample count

Reason for this entry: Do not infer bandwidth or sample size from fictional outcomes.

Design checklist

☐ Not checked: Documented running variable/cutoff and checked exceptions and receipt probabilities
☐ Not checked: Examined manipulation with density diagnostics and recorded context/limits
☐ Not checked: Examined pre-covariate continuity, stating tests cannot prove assumptions
☐ Not checked: Recorded bandwidth selection and rationale
☐ Not checked: Checked stability across bandwidths
☐ Not checked: Recorded need/results for placebo diagnostics; non-significance alone does not establish validity
☑ Checked: Specified sharp/fuzzy estimand and cutoff locality

Reason for this entry: Only the policy to describe local effects is settled. Diagnostics and robustness remain undone.

Common errors in judgment

Simply compare everyone scoring 60 or above with everyone scoring below 60

How to revise: Estimate functions locally on both sides

Reason: Distant scores may have very different preconditions.

Ignore exceptions and report sharp RDD

How to revise: Check actual receipt-probability jumps before selecting design

Reason: Meaning depends on correspondence between assignment and receipt.

Editorial and checking information

Content updated: 2026-09-13Source checked: 2026-09-13 (includes AI cross-checking)No documented human reviewIncludes fictional examplesEditorial policy and scope of checking Cite this page Report a correction

References and checked sections

Development Research in Practice: Research design for impact evaluation
World Bank DIME Analytics
Section consulted
Regression discontinuity
How it relates to the text
Check threshold assignment mechanisms and interpret effects locally.
Source checked: 2026-09-13 (includes AI cross-checking)

How to cite this page

EvalCompass Editorial Team. “RDD design and diagnostics checklist”. EvalCompass, updated 2026-09-13. https://evalcompass.jun-nakatani.com/en/worksheets/rdd/

Add the date you accessed the page. Text created by this site is available under CC BY 4.0. Quotations, external materials, and bundled fonts have their own terms of use. When reusing content, credit the source, link to the license, and indicate any changes.

Suggest a correction or improvement to this page

Change history

2026-09-13 — Checked sources for definitions and conditions and added fictional examples, judgment rationales and completed entries for every field. Revised unsupported historical or uptake claims. Reviewed procedures and cautions, adjusting fixed counts, thresholds and institutional generalizations. Preserved field positions, types and counts while revising fixed conditions and thresholds.