Econometrics 101 - Research Design Worksheet

How to use this worksheet

This worksheet guides you through designing a credible impact evaluation from start to finish. Work through each section systematically, building your research design step by step. The completed worksheet will serve as your evaluation blueprint.

Best Practices:

Involve program implementers and stakeholders in the design process
Be honest about constraints: acknowledge limitations rather than ignore them
Consider multiple identification strategies and choose the most credible
Plan for potential threats to validity and build in robustness checks

1 Program and research question definition

Program description

Describe the development program or intervention you want to evaluate:

Primary research question

State your main causal question clearly and specifically:

Theory of change

Map out the causal pathway from intervention to outcomes:

Policy relevance

Why does this question matter for policy and practice?

2 Outcome measurement strategy

Primary outcomes

List 1-3 most important outcomes and how you'll measure them:

Secondary outcomes

Additional outcomes of interest (spillovers, unintended effects, mechanisms):

Outcome quality checklist

Specific: Outcomes are precisely defined and measurable

Relevant: Outcomes directly relate to program objectives

Feasible: Data collection is realistic given budget and timeline

Pre-specified: Outcomes defined before seeing any results

Balanced: Include both intended and potential unintended effects

3 Identification strategy selection

Assignment mechanism

How are participants selected for the program?

Randomized trial

Random assignment to treatment

Instrumental variables

Valid instrument available

Difference-in-differences

Policy change over time

Regression discontinuity

Threshold-based assignment

Matching/regression

Selection on observables

Method justification

Why is your chosen method the most credible for this context?

Key identifying assumption

State the critical assumption needed for causal interpretation:

5 Sample size and power analysis

Minimum Detectable Effect (MDE)

What's the smallest effect size that would be policy-relevant?

Power calculation parameters

Fill in your power analysis assumptions:

Parameter	Value	Justification
Power (1-β)	____	Typically 0.80 or 0.90
Significance (α)	____	Usually 0.05
Baseline Mean	____	From pilot data or literature
Baseline SD	____	From pilot data or literature
R² (covariates)	____	Variance explained by controls

Required sample size

Based on power calculation:

Power analysis tips

Use conservative assumptions: err on the side of larger samples
Account for clustering if randomization is at group level
Adjust for expected attrition rates
Consider multiple hypothesis testing if many outcomes
Use pilot data or similar studies for parameter estimates

6 Data collection strategy

Baseline data collection

When and how will you collect pre-treatment data?

Follow-up timeline

When will you measure outcomes?

Time Point	Timing After Treatment	Key Outcomes	Rationale
Baseline	Before treatment	All baseline variables	Control for pre-treatment differences
Follow-up 1	____	____	____
Follow-up 2	____	____	____

Data sources

What data will you collect and from where?

Data quality checklist

Reliable: Measurement instruments are validated and consistent

Valid: Measures capture the intended concepts

Timely: Data collection timing aligns with expected treatment effects

Comprehensive: Covers both intended and unintended outcomes

Ethical: Data collection respects participant privacy and consent

7 Analysis plan

Primary specification

Write out your main regression equation:

Control variables

What variables will you include and why?

Robustness checks

How will you test the sensitivity of your results?

Heterogeneity analysis

For which subgroups might effects differ?

Pre-registration requirements

Complete and register your analysis plan BEFORE seeing any outcome data. This prevents data mining and ensures credible results. Include:

Primary and secondary outcome definitions
Exact specification for main analysis
Predetermined subgroup analyses
Multiple testing corrections

8 Implementation and timeline

Project timeline

Map out key milestones:

Phase	Duration	Key Activities	Deliverables
Design & Setup	____	IRB approval, instrument development	Pre-analysis plan
Baseline	____	Data collection, randomization	Baseline report
Implementation	____	Program rollout, monitoring	Implementation report
Follow-up	____	Endline data collection	Clean dataset
Analysis	____	Statistical analysis, report writing	Final report

Budget estimation

Major cost categories and estimates:

Risk management

What could go wrong and how will you address it?

Stakeholder engagement

How will you involve implementers and policymakers?

9 Final quality check

Design quality assessment

Clear Research Question: Specific, policy-relevant causal question

Credible Identification: Plausible strategy for causal inference

Adequate Power: Sufficient sample size to detect meaningful effects

Threat Mitigation: Major validity concerns addressed

Feasible Implementation: Realistic timeline and budget

Ethical Compliance: Meets standards for research with human subjects

Pre-specified Analysis: Clear plan registered before data collection

Overall design assessment

Summarize the strengths and limitations of your design:

Dissemination plan

How will you share results with relevant audiences?

Additional resources

J-PAL Research Resources: Templates and guides for evaluation design
3ie Impact Evaluation Guidelines: Best practices for rigorous evaluation
ImpactMojo Platform: Sample designs and peer review opportunities
Open Science Framework: Pre-registration templates and examples

Research Design Worksheet • ImpactMojo 101 Knowledge Series
Licensed under CC BY-NC-ND 4.0 • Free to use with attribution • www.impactmojo.in

This worksheet is designed to be completed collaboratively with implementers and stakeholders. Save regularly and refer back during implementation.

Research design worksheet