What field experiments are and why they matter for evidence-based policy · History: from agricultural RCTs to development economics and business experiments · Causal inference: the fundamental problem and how randomisation solves it · Types: RCTs, natural experiments, quasi-experiments, lab-in-field, and A/B tests · Ethics: informed consent, IRB approval, and do-no-harm principles

Formulating testable research questions and causal hypotheses · Defining treatment arms, control groups, and comparison conditions · Randomisation strategies: simple, stratified, cluster, and matched-pairs · Unit of randomisation: individual, household, firm, community, or school · Pre-analysis plans: components, purpose, and registering on AEA Registry or OSF

Type I and Type II errors: alpha, beta, and the consequences of each · Power calculations: the four determinants of adequate sample size · Minimum Detectable Effect (MDE): defining meaningful effect sizes · Intracluster Correlation Coefficient (ICC) in cluster randomised trials · Software: G*Power, Stata, R (pwr package), and online calculators

Designing survey instruments: validated scales and custom measurement tools · Baseline, midline, and endline data collection protocols · Avoiding measurement bias: social desirability, recall, and interviewer effects · Digital data collection: SurveyCTO, ODK, and KoboToolbox · Data quality assurance: back-checks, logic checks, field audits, outlier detection

Attrition: differential vs random, Lee bounds, and Horowitz-Manski bounds · Spillover effects and contamination: detection and the SUTVA assumption · Non-compliance: Intent-to-Treat (ITT) vs Local Average Treatment Effect (LATE/IV) · Multiple hypothesis testing: FWER, FDR, and pre-specified outcomes · Hawthorne effect and experimenter demand: minimising and documenting

Designing experiments inside companies: HR, operations, and management · Getting management and ethics board buy-in for internal experiments · Testing training programmes, incentives, nudges, and management practices · Protecting commercial confidentiality and proprietary data ethically · IGL Masterclass case studies: real firm-level RCTs and lessons learned

Measuring entrepreneur outcomes: revenue, employment, survival, and capability · Testing business support, mentorship, and accelerator interventions · Innovation experiments: product, process, market, and organisational change · Working with incubators, innovation hubs, and government enterprise programmes · IGL evidence review: global evidence on what works for entrepreneurs

AI-assisted survey design: adaptive questionnaires and intelligent skip logic · Machine learning for heterogeneous treatment effects (HTE): Causal Forests and GRF · NLP for qualitative data: automated coding, sentiment analysis, topic modelling · AI-enhanced data quality: automated back-checks and anomaly detection · Ethical considerations: transparency, algorithmic bias, and AI in human subjects research

Translating experimental results into policy-relevant language · Writing policy briefs: structure, evidence hierarchy, and communicating uncertainty · Effect sizes, confidence intervals, and what ‘significant’ means for policy · Building sustainable research partnerships with governments and firms · Navigating political and institutional constraints in applied research

Academic publishing: journal selection, submission, peer review, and revision · Pre-registration and open science: reproducibility and transparency · Writing the results section: tables, figures, and interpreting coefficients · Conference presentations: structuring a 15-minute research talk · Policy dissemination: J-PAL bulletins, VoxDev, IGC blogs, and media

Choose a real-world research question with policy or business relevance · Write a complete pre-analysis plan (PAP) to AEA Registry standard · Design the full experiment: randomisation, power calculation, instruments, timeline · Live canvas feedback: present design to panel for structured critique · Final submission: revised design incorporating all panel recommendations

Design, power, and implement RCTs to the highest global standards · Collect rigorous data and apply ITT and LATE estimation · Navigate attrition, spillovers, and the five main threats to validity · Communicate findings to policymakers, firms, and academic audiences · Achieve a credential grounded in IGL, J-PAL, World Bank DIME, and AEA standards

Complete all 11 modules, submit a complete pre-analysis plan (PAP) registered on the AEA Registry or OSF, and present the experiment design to a review panel (minimum 75%).

Impact Evaluator, Development Economist, Research Manager (NGO/Government), J-PAL/IGL Research Affiliate, Policy Analyst, Academic Researcher. Average starting salary: $55,000–$95,000 USD.