control group
The trial participants who do not receive the active treatment, used as a baseline to measure a vaccine's or drug's effect.
As defined in John Cochrane on the Pandemic
Where it comes up · 15
- Nature, Nurture, and Identical Twins (with David Bessis)4 claims
- Chris Blattman on Sweatshops3 claims
- Austin Frakt on Medicaid and the Oregon Medicaid Study2 claims
- Chris Blattman on Cash, Poverty, and Development2 claims
- 318 | Edward Miguel on the Developing Practice of Development Economics2 claims
- Designing Better Decisions1 claim
- Future of Finance1 claim
- Making Your 80,000 Hours Count (with Benjamin Todd)1 claim
and 7 more
Most important claims · 10 of 23Search all claims for “control group” →
- AI value propositions must be validated with a feedback loop that covers edge cases and measures incremental lift against a baseline, which requires experimental design such as a control or random group, because product managers often fail to specify this.Future of Finance
- The Scared Straight program, which took youthful offenders to tour prisons to deter crime, appeared to work because participants committed fewer crimes afterward, but compared to a control group that never went through it the participants actually committed more crimes, showing that wanting to make a difference and even apparent before-after improvement is not enough without counterfactual comparison.Making Your 80,000 Hours Count (with Benjamin Todd)
- Bouchard's study began by collecting a fraternal-twin control group as the wise standard procedure, but the 1990 Science paper stated they decided not to publish it due to space constraints and small sample size—both implausible excuses, since reporting a single correlation coefficient takes minimal space and a nearly 2x sample difference yields only about a 40% precision change, making the omission scientifically unacceptable.Nature, Nurture, and Identical Twins (with David Bessis)
- The study persuaded factory owners with large unskilled workforces to randomize hiring: among ~150 qualified applicants for ~50 unskilled openings, 50 were offered the factory job, 50 were offered a $300 cash grant plus five days of business consulting, and 50 formed a pure control group, all followed over a year to compare income, health, and career trajectories.Chris Blattman on Sweatshops
- Merely being offered and briefly working a factory job (about an extra 10 weeks of exposure on average) doubled the rate of self-reported serious health problems after a year, from 4% in the control group to 8% in the factory group, attributed to particulate matter and chemical exposure rather than stress, since reported anxiety and depression did not differ between groups.Chris Blattman on Sweatshops
- Oregon's decision to allocate roughly 30,000 Medicaid expansion slots out of 90,000 applicants by random lottery inadvertently created a randomized controlled trial of Medicaid access, which Harvard and MIT researchers seized upon to study its effects.Austin Frakt on Medicaid and the Oregon Medicaid Study
- Being offered a factory job raised the probability of still being in any factory job a year later by only about 12 percentage points, because 20% of the control group found factory jobs elsewhere on their own, so the experiment mainly compared more factory exposure (average 20 weeks) to less (average 10 weeks) rather than factory work versus no factory work.Chris Blattman on Sweatshops
- Because most people in standard vaccine trials are careful and few get exposed, requiring researchers to wait for tiny numbers of natural infections makes trials slow; allowing healthy young volunteers to deliberately expose themselves in challenge trials would have produced better data far more quickly, but the FDA refused on ethical grounds.John Cochrane on the Pandemic
- The known way to debias the twin natural experiment, traceable to Falconer's 1960 textbook, is to use a control group of fraternal twins separated under the same conditions; one then takes the identical-twin correlation, subtracts the fraternal-twin correlation, and multiplies by two to approximate heritability—so the methodology was fully available in 1960.Nature, Nurture, and Identical Twins (with David Bessis)
- Although the lottery randomized many thousands, the analyzable study ended with roughly 6,000 in the control group and 6,000 in the treatment group (those permitted to apply), of whom only about 25%—around 1,500 people—actually enrolled in Medicaid, which is the small group whose effects the latest analysis estimates.Austin Frakt on Medicaid and the Oregon Medicaid Study