Instrumental Variable Estimation under One-Sided Non-Compliance

In Simple Words

Imagine testing whether wearing a new type of safety goggles reduces eye injuries in a factory. If workers randomly offered goggles choose to wear them only half the time, while unoffered workers can never get them, researchers can use the initial random offer as a lever to isolate the pure effect of actually wearing the goggles.

Research question

When evaluating digital decision-support tools under full discretion, frontline staff frequently choose not to inspect algorithmic predictions, creating endogenous selection bias. Standard intent-to-treat estimates dilute actual software effects, while direct comparisons of users versus non-users suffer from self-selection confounding.

Participants

N = 14,977 treatment jobseekers and N = 16,566 control jobseekers across 21 Swiss employment offices.

Study Design

Two-Stage Least Squares (2SLS) Instrumental Variables regression using caseworker randomization status (Z) as an instrument for prediction download behavior (D) under one-sided non-compliance (Imbens & Angrist, 1994).

Results

IV estimates proved that downloading predictions had zero statistically significant impact on short-term caseworker compliance (estimates ranged from -0.05 to +0.03 across specifications, all statistically insignificant).

Takeaway

Instrument actual software usage with randomized assignment status to estimate local average treatment effects when software adoption is voluntary and non-compliance is one-sided.

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.

Statistically significant

A label used when a result passes a chosen statistical threshold, often p < 0.05. It does not show how large or useful an effect is. A result that is not statistically significant does not prove there is no effect.

ExampleA study reports p = 0.03 against a 0.05 threshold. It passes that test, but the effect could still be too small to matter.

American Statistical Association · Statement on p-values ↗Read more about statistically significant →

Related findings

• Catchment Cohort Size Instrument →Methods & evidence

  To measure how school grant size truly affects family spending without getting confused by family choices, researchers look at how many children of school age happen to live in the surrounding area. Since the total grant amount is fixed per school, having more nearby children automatically shrinks the per-child funding without relying on family preferences or behavior.

• Intention-to-Treat (ITT) Instrumentation →Methods & evidence

  When some assigned groups fail to receive a promised intervention, using initial random assignment as an indicator to adjust overall outcomes is like measuring a fitness program's impact based on everyone invited, then mathematically adjusting for those who actually attended the gym.

• Second-Order Peer Instrumentation (G²X Identification) →Methods & evidence

  Imagine trying to prove whether a student's study habits directly improve a classmate's grades. Instead of measuring the classmate directly, researchers look at the background traits of the classmate's friends from another school—people who affect the classmate but have no direct contact with the original student.

• Active Tracking Attrition Mitigation →Methods & evidence

  Dispatching field teams directly to children's homes when they miss school testing operates like a home delivery service tracking down uncollected mail to ensure every package is accounted for.