margin of error
The statistical uncertainty range around a poll estimate, here roughly plus or minus 3 points.
As defined in David Brady on the 2012 US Election
Where it comes up · 4
- Doug Rivers on polling 7/21/20085 claims
- The Search for the Truth with the Anonymous Master2 claims
- How to Think So Clearly People Assume You’re A Genius1 claim
- David Brady on the 2012 US Election1 claim
Most important claims · 9 of 9Search all claims for “margin of error” →
- Calling a 46-40 poll lead (±3% margin of error) a 'statistical dead heat' or 'statistical tie' is incorrect: the best estimate is a 6-point lead, not zero, and while the 95% confidence standard means you can't formally declare a winner, it is far more likely the lead is real than that the true gap is zero, since extreme errors are unlikely.Doug Rivers on polling 7/21/2008
- Of about 1,300 polls Rivers examined from the 2008 presidential primaries, fewer than 50 reported anything other than the margin of error for a simple random sample with no weighting—meaning the reported margins of error are systematically misleading, which Rivers calls scandalous.Doug Rivers on polling 7/21/2008
- Weighting creates two distinct problems: it roughly doubles the real variability compared to a simple random sample, and it leaves residual skews (unobserved or uncorrected biases) that do NOT shrink as sample size grows, so conventional sampling-error estimates tell you how the procedure varies sample-to-sample but not how systematically wrong the procedure is.Doug Rivers on polling 7/21/2008
- Weighting underrepresented groups by large factors (5-10x for some groups) drastically inflates variability: with a weight of 10 in a sample of 1,000, a single miscoded respondent shifts the whole sample by 1%, so a single error can destroy a nominal 3% margin of error.Doug Rivers on polling 7/21/2008
- Analysis is dangerous in complex systems like markets because it is easy to miss a single variable that matters, and it is easy to tell persuasive stories, so analysis tends to be wrong and needs margins for error and stress testing.The Search for the Truth with the Anonymous Master
- A truly random sample of about 1,000 people gives a margin of error of about plus or minus 3% with 95% confidence regardless of total population size, so a tiny fraction (one in 300,000 Americans) suffices; quadrupling to ~4,000 tightens it to about 1% at 99% confidence.Doug Rivers on polling 7/21/2008
- Because analysis is fragile to a single missed link, an analyst should build in margins for error so the conclusion survives small mistakes, stress-test the reasoning openly to locate the weak link, and then validate it by testing, since testing checks every link in a chain at once.The Search for the Truth with the Anonymous Master
- Confidence is a byproduct of scope—to increase confidence in a conclusion, widen the scope and increase the margin of error rather than attempting to increase precision.How to Think So Clearly People Assume You’re A Genius
- Across the agreed-upon swing states (Colorado, Nevada, Missouri, Iowa, Wisconsin, Indiana, Florida, Virginia, North Carolina, New Hampshire, Ohio), Obama holds a lead right at the margin of error (±3 points), while losing in Indiana, North Carolina, and likely Florida, with a surprising 4.5-5 point lead in Ohio.David Brady on the 2012 US Election