Lottery-Based Instrumental Variable Identification Strategy

In Simple Words

Imagine trying to test if wearing running shoes makes people faster, but fast people are naturally more likely to buy running shoes. Using a lottery to hand out running shoes completely breaks that confounding connection, ensuring the only difference between the two groups is pure chance.

Research question

Unobserved youth characteristics (such as motivation, conscientiousness, or family background) confound observational comparisons between employed and unemployed youth, making unadjusted comparisons unreliable.

Participants

N = 163,447 applicants randomized across hundreds of local provider lotteries in New York City (2005–2008).

Study Design

Two-Stage Least Squares (2SLS) model using randomly assigned lottery win status as an instrument for actual SYEP program participation, conditioning on provider-year-age lottery pool fixed effects.

Results

First-stage F-statistics exceed standard weak-instrument thresholds, showing that winning a lottery increases SYEP participation by over 50 percentage points, isolating clean local average treatment effects (LATE) free from selection bias.

Takeaway

1. Utilize administrative randomized selection lotteries as instruments when evaluating oversubscribed public interventions to isolate unconfounded causal parameters.

Read alongside the research

Plain-English glossary

Original explanations, informed by the references below. Examples are illustrations, not data from the studies above. The research text stays unchanged.

Percentage points

The difference between two percentages.

ExampleA rise from 40% to 50% is 10 percentage points. Relative to the starting value, that is a 25% increase.

Office for National Statistics · Percentages ↗Read more about percentage points →

Related findings

• Community Incentive and Distance Instrumental Variable Method →Economics (general)

  Imagine trying to determine if eating at a local diner makes people happy, but happier people naturally choose to eat out more often. To isolate the real effect, a researcher gives randomly assigned discount coupons to various neighborhoods; comparing neighborhood happiness based solely on coupon density reveals the pure effect of diner visits.

• Instrumental Variables Correction for Management Measurement Error →Economics (general)

  Imagine trying to measure how much a medicine helps sick patients by looking at regular hospital records. Because doctors give the medicine primarily when patients get extremely ill, simple statistical comparisons make the medicine look weak or useless.

• Administrative Social Security Data Linkage for Earnings Measurement →Economics (general)

  Instead of asking individuals to remember and report exact yearly income numbers from a decade ago, researchers obtain signed consent to link survey profiles directly to official government tax records. This creates an unalterable, objective earnings ledger covering years before and after a financial event.

• Agnostic Machine Learning Heterogeneity Detection in Field Trials →Economics (general)

  Evaluating a program using only average results is like measuring the average temperature of a room with one frozen wall and one burning wall and concluding the room is comfortable. Machine learning acts like a thermal camera, splitting the room into sections to reveal the extreme cold and extreme heat hiding behind the average.