Flagship Course • Free Forever

Designing What Works: Development Interventions from Model to Scale

From Model to Principle to Scale

The missing craft between policy and evaluation: how to design a development intervention that actually works — and survives contact with scale. It dissects the models proven at population scale (the graduation approach, cash transfers, community health workers, self-help groups, teaching at the right level), extracts the design principles behind them, and confronts the hard part: why proven interventions collapse when you multiply them. Grounded in South Asian delivery systems — NRLM, ASHA, DBT, Pratham, BRAC.

Proven-at-Scale Models Indian Cases Interactive Lexicon
14 Comprehensive Modules
Models → Principles → Scale
South Asia Focus
Design Judgement
Intervention Design Lexicon Course Papers Soon AI Study Companion Soon

Why a Whole Course on Intervention Design?

The development sector knows how to diagnose problems and how to evaluate results. The craft in between — taking a problem and a theory of change and designing a concrete intervention that works, and then keeps working at scale — is mostly learned by apprenticeship, if at all. This course makes that craft explicit. It teaches design the way it is actually earned: by dissecting the handful of interventions that reached population scale, extracting the principles that made them work, and studying why so many proven pilots collapse on the way up.

Where this sits. It is the missing middle of your curriculum. Public Policy covers how policy is made upstream; MEL and Causal Inference cover how you measure whether something worked downstream. This course is about the design decision in between — the intervention itself. It builds on the foundations in Theory of Change 101, Logframe 101, and Cost Effectiveness 101.

Models before opinions

Design principles are extracted from interventions that actually reached scale — the graduation approach, conditional cash transfers, ASHA, self-help groups, teaching at the right level — not from templates. You learn the craft by taking apart what worked.

Grounded in delivery

Every principle is tested against the systems that actually deliver in South Asia: NRLM and its SHG federations, the ASHA cadre, Aadhaar-linked DBT rails, Pratham's classrooms. Design that ignores the delivery chain is design that fails at scale.

The scaling question

The intellectual core is the hardest problem in the sector: why interventions that work in a trial break when you multiply them. The Voltage Effect, PDIA, and the science of scaling — and the three real pathways from pilot to population.

"The ideas that scale are not always the best ideas. They are the ones that keep their voltage as they grow." — after John A. List, The Voltage Effect (2022)

Assess Yourself — Intervention Design

Six auto-graded questions on the core ideas of the course — models, principles, and the scaling problem. Pick an answer and check it; each explains the reasoning. Nothing is stored and there's no sign-in.

1What most clearly distinguishes an intervention from a policy?Foundations
2In the BRAC graduation approach, the sequenced bundle outperformed its individual components mainly because:The Proven Models
3Moving from broad self-targeting to a tight proxy-means test (PMT) most directly changes:Design Principles
4In List’s Voltage Effect, one core reason an intervention loses its effect at scale is that:The Scaling Question
5Problem-Driven Iterative Adaptation (PDIA) is best summarised as:Design Principles
6A cost-effectiveness comparison used to choose between two interventions should be built on:Design Principles

Papers & Resources

A short, opinionated reading list. The books anchor the way of thinking; the evidence sources are the field's reference points for what works and what scales.

Core books

Banerjee & Duflo, Poor Economics (2011). John A. List, The Voltage Effect (2022). Andrews, Pritchett & Woolcock, Building State Capability (2017, free online). Karlan & Appel, More Than Good Intentions (2011). Pritchett, The Rebirth of Education (2013).

Evidence & scaling

J-PAL policy insights and the Scaling initiative; 3ie evidence & gap maps; GiveWell cost-effectiveness analyses. The BRAC graduation six-country RCT (Banerjee et al., Science 2015) and Progresa/Oportunidades evaluations are the canonical case files.

Sibling courses

Sits between Public Policy (upstream) and MEL / Causal Inference (downstream). Feeds from Theory of Change 101, Logframe 101, Cost Effectiveness 101, and Community Development 101.