Quarterly notes on civic experimentation.

Experiment spotlights, practitioner observations, and one pattern from the evidence base worth acting on. No conference announcements. No fundraising asks.

Subscribe →

Previous issues.

Issue 09May 19, 2027
Behavioural Insights TeamBehavioral science

What the Nudge Unit proved.

How a seven-person team in the UK Cabinet Office tested behavioral science at government scale — and what it means for institutional learning.

In 2010, the UK Cabinet Office created a small team with an improbable mandate: apply the insights from behavioral economics to improve government services, and prove the results through randomized co

Read → 9 min read
Issue 08March 24, 2027
Elinor OstromPolycentric governance

Ostrom and the commons.

What the Nobel Prize winner showed about governance that economists missed — and why it matters for civic experimentation.

In 1968, ecologist Garrett Hardin published 'The Tragedy of the Commons,' arguing that shared resources would inevitably be destroyed by individual self-interest. The solution, he claimed, was either

Read → 10 min read
Issue 07February 9, 2027
Job TrainingHuman capital

Why most job training programs fail — and what the ones that work have in common.

The largest RCT in US workforce policy found no effect. A smaller study found 30%. The difference is not the training.

In 1994, Mathematica Policy Research published the results of the largest randomized evaluation of job training ever conducted. Twenty thousand participants across sixteen sites. The program was the J

Read → 10 min read
Issue 06November 12, 2026
Voter EngagementSocial norms

The mailer that changed political science.

What a 2006 Michigan experiment revealed about civic behavior — and why the finding is more complicated than it looks.

In 2006, researchers mailed 344,000 Michigan voters a simple letter. It showed them their own voting history alongside their neighbors' records, with an implicit message: we will check again after the

Read → 9 min read
Issue 05August 13, 2026
FederalismConstitutional Law

Federalism as experiment.

Brandeis, the Tenth Amendment, and why the constitutional structure of the United States is a learning architecture.

In 1932, Justice Louis Brandeis wrote a dissent in a case about ice. The case was New State Ice Co. v. Liebmann, and it concerned whether Oklahoma could require a license to sell ice. Brandeis thought

Read → 11 min read
Issue 04May 8, 2026
SimplificationTax & Revenue

Friction is policy.

What a decade of simplification experiments tells us about administrative burden as a choice.

Across twelve countries and four policy domains, one finding recurs with unusual consistency: when you reduce the steps required to comply with a government process, compliance goes up — sometimes dra

Read → 8 min read
Issue 03February 5, 2026
Public SafetyTargeting

What hot spots policing actually showed.

The most replicated finding in criminology is also the most misapplied.

The Kansas City Gun Experiment (1995) is widely cited as the origin of hot spots policing. Increased patrol in specific high-crime blocks reduced gun crime by 49% in target areas. The finding has been

Read → 10 min read
Issue 02November 6, 2025
Evidence qualityPublication bias

The null result problem.

Why the most important experiments don't get published, and what to do about it.

Roughly 60% of randomized controlled trials in public policy show no statistically significant effect. In published literature, that share drops to about 30%. The gap is not random. Null results are s

Read → 7 min read
Issue 01August 14, 2025
Cash TransfersInternational Development

How conditional cash transfers actually work.

The mechanism behind Mexico's Progresa and what three decades of evidence shows.

When Mexico's Progresa program launched in 1997, it was one of the first large-scale conditional cash transfer programs in the world. Families in extreme poverty received regular payments — on the con

Read → 9 min read

Subscribe

Get the next issue when it drops.

Four issues per year. Each one anchored to a specific pattern from the registry. No filler.

Subscribe by email →