How to read a confidence interval
Each PLACES estimate has a 95% confidence interval. On this site the interval is the band in each row of the report card. The estimate is the bar in the band. The two numbers of the interval are in the row as text.
What the interval shows
CDC makes the interval from 1,000 runs of its model:
Monte Carlo simulation is used to generate 1,000 simulated datasets for the point estimate, the final estimates are reported as the mean and 95% confidence interval (the 2.5th, 97.5th percentiles) over 1,000 draws.
Centers for Disease Control and Prevention, PLACES Methodology (page dated 7 Dec 2025, read on 2 Oct 2026)
The interval shows how sure the model is of its estimate. If the model is right about the area, the share is probably in the range of the band.
Why small areas have wide bands
The smaller the areas are, the broader the confidence intervals of an estimate will be.
Centers for Disease Control and Prevention, PLACES Frequently Asked Questions (page dated 11 Dec 2025, read on 2 Oct 2026)
A county with a million adults can have a band of one or two points. A ZIP Code with a few hundred adults can have a band of ten points or more. A page for an area with fewer than 500 residents has a note that says so.
The overlap rule
First, it is not adequate to simply compare point estimates. The confidence intervals should also be considered, and some are very broad.
Centers for Disease Control and Prevention, PLACES Frequently Asked Questions (page dated 11 Dec 2025, read on 2 Oct 2026)
Before the site says that an area is higher or lower than its county or the nation, the two bands must not overlap at all. The whole band of the area must be above or below the whole band of the other.
When the bands overlap, the site says similar. This is true even when the two estimates are far apart.
A state figure has no band. Against a state figure, the band of the area must not include the state figure. The words are above and below. Otherwise the word is in line.
The rule is stricter than a formal test of difference. It can call two areas similar when a test would separate them. It does not call two areas different when the bands do not support it.
The lists give no rank
The lists of counties, places and ZIP Codes are in name order or in ZIP Code order. The estimate columns have no sort. The bands of many rows overlap, so the order of two estimates does not show which area is higher. CDC says:
Because these are modeled and not direct estimates, the data should not be used for ranking the overall health of any county, place, census tract, or ZCTA.
Centers for Disease Control and Prevention, PLACES Frequently Asked Questions (page dated 11 Dec 2025, read on 2 Oct 2026)
PLACES does not provide a weighted composite score for the included counties, places, census tracts, or ZCTAs.
Centers for Disease Control and Prevention, PLACES Frequently Asked Questions (page dated 11 Dec 2025, read on 2 Oct 2026)
Releases
A page can show the estimate and the band of the same area in each release. CDC says about the intervals:
Beginning with the 2023 release, the assumption in the Monte Carlo simulation approach was changed so that the random error for the random effects varied only within counties. As a result, the estimated CIs may be wider than previous releases.
Centers for Disease Control and Prevention, PLACES Frequently Asked Questions (page dated 11 Dec 2025, read on 2 Oct 2026)
The releases also use different BRFSS years. The site shows them as a record. It gives no comparison between two releases.
Next: what the figures are for.
Last checked against the CDC pages: