
interpret confidence intervals and margins of error in published reports, convert between confidence levels, and identify errors that a margin of error does not cover.
Instruction
Confidence intervals are everywhere in public data, and they are often misread. This lab is about reading them well.
Five common misreadings.
- “There’s a 95% probability that μ is in this interval.” No. μ is fixed, and the 95% describes how often the method succeeds.
- “95% of the data are in the interval.” No. The interval estimates a parameter, not where individual values fall. The St. Louis temperature interval (52.8-69.6°F) contains only about 23% of daily highs, far fewer than 95%.
- “A 99% interval is better.” It is more often right, but wider. Confidence and precision trade off at a fixed n.
- “The margin of error covers all error.” No. It covers only random sampling variability. Undercoverage, nonresponse, response bias and processing errors are not in it.
- “If two intervals overlap, the groups do not differ.” Not necessarily. The difference between two estimates has its own standard error, √(SE12 + SE22), which is smaller than SE1 + SE2. Intervals can overlap a little while the difference is still significant. Compare groups with an interval for the difference (Lesson 2.4).
Census margins of error. The American Community Survey publishes every estimate on data.census.gov with a margin of error at the 90% confidence level, not 95%. To use an ACS estimate:
- SE = MOE / 1.645.
- A 95% margin = 1.96 × SE = MOE × (1.96/1.645) ≈ MOE × 1.19.
- Illustrative: an estimate of 12.0% with a published MOE of ±1.5 percentage points has SE ≈ 0.91 and a 95% margin of about ±1.8 points (10.2% to 13.8%).
- To compare two areas’ estimates, use SEdiff = √(SE12 + SE22). The Census Bureau’s ACS documentation describes this approach.
Small areas have large ACS margins because their samples are small. A neighborhood estimate with an MOE nearly as big as the estimate itself tells you very little, and a careful report says so.
Polls and the lead. Illustrative: a poll of 1,068 likely voters reports Candidate A 52%, Candidate B 48%, margin of error ±3 points. Is A ahead? The ±3 applies to each candidate’s share. The lead (4 points) is a difference between two shares from the same sample, and when one goes up the other goes down. For shares near 50%, the margin of error for the lead is about twice the reported margin: here about ±6 points. A 4-point lead is well within it, so the poll alone cannot say who leads the population.
Estimates are about populations, not individuals. An interval for the average commute in a county says nothing precise about your commute.
The discourse task (MP.3). Bring one published estimate with a margin of error, such as an ACS table for your county, a poll, or a CDC survey estimate. State the parameter, the confidence level, and what the interval means; name one error the margin does not cover; and judge one comparison the source makes, or that a reader might make. Then reread your critique against the list of misreadings above and correct any you committed (or talk it through with a parent or guardian and record what they questioned).
📖 Read for this lesson
The reading opens over the lesson; close it to come back to exactly where you were. The checks below draw on it.
Formative check
Work through these before moving on. They are not graded — they tell you whether the standard below has landed yet.
An ACS estimate published with a margin of error of ±1.5 points has a 95% margin of error of about ±1.8 points.
A poll reports 52% vs 48% with a margin of error of ±3 points. What is the approximate margin of error for the 4-point lead?
95% confidence
Tap to revealThe method captures the parameter in about 95% of samples.
Wider interval
Tap to revealHigher confidence or smaller n.
Margin of error excludes
Tap to revealBias from coverage, nonresponse, wording, processing.
ACS MOE
Tap to reveal90% level; SE = MOE/1.645.
Compare two estimates
Tap to revealUse SE of the difference: √(SE₁² + SE₂²).
Record (or write) a 90-second critique of a published estimate with a margin of error (for example the ACS broadband table for Missouri and twelve of its counties hosted in this lesson’s reading, where every estimate has the Census Bureau’s own margin, or another poll or ACS table you find). State the parameter and confidence level, convert the margin to 95% if needed, name one nonsampling error, and judge one comparison. Then add a closing paragraph that names the misreading from the list a hasty reader would be most likely to make about your chosen estimate, and corrects it.
How confident are you that you can read published estimates and margins of error correctly and spot common misinterpretations?
Practice
Work these on paper or in your notebook, then open Check your answer. Aim for all of Fluency and Application; try at least one Challenge.
Printable version: this unit’s practice workbook (PDF)
Fluency
Build speed and accuracy with the core skill.
- An ACS estimate has a published margin of error of ±2.4 points (90% confidence). Find the SE and the 95% margin of error.
Check your answer
Answer: SE = 2.4/1.645 ≈ 1.459; 95% margin ≈ 2.86 points. - True or false: a 95% confidence interval for mean daily high contains about 95% of individual days’ highs.
Check your answer
Answer: False.It estimates the mean; individual days vary far more. - True or false: for the same data, a 99% interval is wider than a 95% interval.
Check your answer
Answer: True. - Name two sources of error that a survey’s margin of error does not include.
Check your answer
Answer: Any two of: undercoverage, nonresponse, response or wording bias, measurement or processing errors. - A poll of 1,068 reports 52% vs 48% with a margin of error of ±3 points. About what is the margin of error for the 4-point lead?
Check your answer
Answer: About ±6 points.For shares near 50%, roughly twice the reported margin. - Two intervals overlap slightly. Does that prove the two population values are equal?
Check your answer
Answer: No. Compare using an interval for the difference.
Application
Use the skill in context. Show your reasoning.
- (Illustrative ACS-style estimates, 90% margins) County A: 14.2% ± 1.8; County B: 11.5% ± 2.1. Is the difference statistically significant at the 90% level?
Check your answer
Answer: The difference is 2.7 points. Its 90% margin is √(1.8² + 2.1²) ≈ 2.77. Since 2.7 < 2.77, the interval for the difference includes 0: not significant at 90%.Margins of independent estimates combine as the square root of the sum of squares. - A news story says: ‘Support is 47% with a margin of error of 3 points, so fewer than half of voters support the measure.’ Evaluate the claim.
Check your answer
Answer: Not justified: the interval is about 44% to 50%, which includes values up to 50%, and the margin covers only sampling error. - A small-area ACS estimate reads 8% ± 7 points. What should a careful report say about it?
Check your answer
Answer: That it is very imprecise (the 90% interval runs from about 1% to 15%) and should not be used for fine comparisons; combine areas or use five-year estimates if possible.
Challenge
Stretch problems. Expect to think before you write.
- Explain why the margin of error for the lead in a single poll is about twice the margin for each candidate’s share, when the two shares are near 50%.
Check your answer
Answer: The lead is p₁ − p₂ from the same sample, and the shares move in opposite directions (when one rises, the other tends to fall), so their variability adds. Var(p̂₁ − p̂₂) = [p₁ + p₂ − (p₁ − p₂)²]/n ≈ 1/n when the shares are near 0.5, which is four times 0.25/n, so the SD and the margin are about twice as large.
Review
Keep earlier skills sharp.
- In a random sample of 500, 373 said yes. Construct a 99% confidence interval for the population proportion.
Check your answer
Answer: p̂ = 0.746; margin ≈ 0.050; interval ≈ (0.696, 0.796). We are 99% confident that the proportion of the whole population who would say yes is between 0.696 and 0.796.z* ≈ 2.576. Conditions: a random sample; the sample is less than 10% of the population; successes and failures (373 and 127) are both at least 10. - (Lesson 2.2) Why do you round a required sample size up, never down?
Check your answer
Answer: Rounding down would give a margin of error slightly larger than required.
The Census Bureau’s Own Margins: Broadband Subscriptions in Thirteen Missouri Areas
U.S. Census Bureau, American Community Survey 2023 1-year estimates, detailed table B28002 (lines 1 and 4) for Missouri and twelve of its counties and county-equivalents, from the ACS Summary File (table-based format, acsdt1y2023-b28002.dat); the claims to judge are SLS original framings; retrieved 2026-11-10 · Public domain (works of the United States government), with SLS original worked comparisons
Why this reading. The discourse task says to bring an estimate with a margin of error, “such as an ACS table for your county”. Here is one: the Census Bureau’s 2023 American Community Survey count of households with a broadband subscription, for Missouri and twelve of its counties and county-equivalents, with the Bureau’s own published margins of error. It is exactly the object the lesson teaches you to read: a 90% margin, an estimate that is a count divided by a count, and a comparison you may be tempted to make by eye. Three comparisons are worked, and three claims are set up to judge.
The survey
The American Community Survey samples about 3.5 million addresses a year across the United States and asks, among many other things, whether the household has an internet subscription and of what type. The Bureau publishes each estimate with a margin of error at the 90% confidence level. The figures below are from detailed table B28002, Presence and Types of Internet Subscriptions in Household, 2023 1-year estimates, line 1 (all households) and line 4 (households with a broadband subscription of any type), taken from the ACS Summary File. One-year estimates exist only for areas of 65,000 or more people. Retrieved 2026-11-10.
Table 1. Households with a broadband subscription, ACS 2023 1-year
| Area | Households | MOE (90%) | With broadband | MOE (90%) | Percent | MOE of percent (90%) | MOE at 95% |
|---|---|---|---|---|---|---|---|
| Missouri (state) | 2,556,271 | ±10,804 | 2,319,607 | ±14,203 | 90.7% | ±0.40 | ±0.48 |
| Jackson County | 310,868 | ±3,159 | 290,926 | ±3,598 | 93.6% | ±0.66 | ±0.79 |
| St. Louis County | 419,348 | ±3,213 | 394,648 | ±4,620 | 94.1% | ±0.83 | ±0.99 |
| St. Louis city | 150,404 | ±2,942 | 136,373 | ±3,576 | 90.7% | ±1.58 | ±1.89 |
| St. Charles County | 163,272 | ±1,908 | 155,874 | ±2,706 | 95.5% | ±1.23 | ±1.46 |
| Greene County | 132,232 | ±1,656 | 115,952 | ±2,542 | 87.7% | ±1.58 | ±1.88 |
| Clay County | 103,995 | ±1,736 | 94,469 | ±2,348 | 90.8% | ±1.67 | ±1.99 |
| Jefferson County | 88,836 | ±1,556 | 82,463 | ±1,949 | 92.8% | ±1.47 | ±1.76 |
| Boone County | 76,762 | ±1,573 | 70,452 | ±2,251 | 91.8% | ±2.25 | ±2.68 |
| Jasper County | 52,000 | ±1,093 | 46,490 | ±1,593 | 89.4% | ±2.42 | ±2.88 |
| Christian County | 35,654 | ±851 | 31,843 | ±1,453 | 89.3% | ±3.47 | ±4.14 |
| Platte County | 44,751 | ±880 | 42,035 | ±1,099 | 93.9% | ±1.62 | ±1.93 |
| St. Francois County | 25,489 | ±1,026 | 22,481 | ±1,136 | 88.2% | ±2.69 | ±3.21 |
The Bureau publishes the two counts and their margins; the percent and its margin are derived. For a proportion p = Y/X whose numerator is a subset of its denominator, the ACS documentation gives MOE(p) = (1/X)√(MOEY2 − p2MOEX2). For Greene County: p = 115,952/132,232 = 0.8769; MOE = (1/132,232)√(2,5422 − 0.87692 × 1,6562) = 0.01578, or ±1.58 points. The last column converts to 95% by multiplying by 1.96/1.645 = 1.19.
Three comparisons, worked
The standard error of each estimate is MOE/1.645; the SE of a difference between two areas is √(SE12 + SE22); the 90% margin on the difference is 1.645 times that.
| Comparison | Difference (points) | SE of difference | 90% margin on difference | Verdict |
|---|---|---|---|---|
| Jackson minus Greene | 5.9 | √(0.4012 + 0.9592) = 1.040 | ±1.71 | difference exceeds its margin |
| St. Louis minus St. Louis city | 3.4 | √(0.5062 + 0.9632) = 1.088 | ±1.79 | difference exceeds its margin |
| Jasper minus Christian | 0.1 | √(1.4712 + 2.1112) = 2.573 | ±4.23 | not distinguishable |
Look at the Jasper–Christian row against the naive method. Jasper’s 90% interval is 87.0 to 91.8; Christian’s is 85.8 to 92.8. They overlap almost completely, and the difference is also not significant, so here the two methods agree. They do not always: the lesson’s fifth misreading is that overlapping intervals prove no difference, and the difference’s own margin, √(SE12 + SE22), is always smaller than SE1 + SE2, which is what overlap tests. Try the St. Louis County–St. Louis city row: the two 90% intervals (93.3 to 94.9 and 89.1 to 92.3) do not overlap, and the difference of 3.4 points is larger than its own margin of ±1.79.
Claims to judge (SLS original framings of these real figures)
A. “St. Louis city trails the county around it in broadband: 90.7% of households against 94.1%.” The arithmetic holds: the 3.4-point gap is about 3.2 standard errors. Now ask what a margin of error does not cover. The ACS counts households, not people, and excludes group quarters (dormitories, care homes, shelters). A city with many one-person households, many renters who move often, or many residents in group quarters can differ from the county in ways the subscription question does not capture, and a household may “have a subscription” that it cannot afford to keep.
B. “Only 88% of Greene County households have broadband. Springfield’s county has the worst connectivity in Missouri.” Greene’s estimate is 87.7% ± 1.58, and the gap to the state, 3.1 points, is about 3.1 standard errors: real beyond doubt. But “worst in the state” compares Greene with 114 other jurisdictions (113 counties and St. Louis city), 98 of which are not in the 1-year file at all, and the 16 that are have margins from under one point to about 3.5 points. Buchanan County (St. Joseph), also in the 1-year file, has an estimate of 87.2%, below Greene’s, and its margin is not small relative to the gap. A ranking built from a pile of estimates, each with its own margin, is a much weaker claim than any one comparison.
C. “The margin of error is ±0.40, so the statewide figure of 90.7% is known almost exactly.” Yes for sampling error: with about 2.6 million households and a large sample, the interval is narrow. But the 9.3% without broadband is roughly 236,664 households, and the ACS’s own nonresponse follow-up, imputation of missing answers and the wording of the question all bear on that figure in ways the ±0.40 does not include. A tiny margin of error describes precision, not accuracy.
Bring one of these to the discourse lab, or bring your own county’s row. To retrieve a county not printed here, search data.census.gov for table B28002 and choose the 2023 1-year or, for counties under 65,000 people, the 5-year estimate, whose margins will be different again.
Sources: U.S. Census Bureau, ACS table B28002 for Missouri counties (data.census.gov) (data.census.gov); U.S. Census Bureau, ACS Summary File (the file used here: 2023 1-year, acsdt1y2023-b28002.dat) (www2.census.gov); U.S. Census Bureau, ACS Accuracy of the Data (2023), including the derived-estimate formulas (census.gov); U.S. Census Bureau, guidance on using ACS estimates and margins of error (census.gov)