When a small-town statistic changes, read the margin of error

A useful number comes with a label

A claim about a town’s income, housing or commuting patterns can sound exact when it arrives as a single number. If the source is the American Community Survey, the figure is an estimate drawn from a sample.

Its margin of error is part of the information needed to interpret it.

The Census Bureau advises using the decennial census or its Population Estimates Program when the question is population totals. The ACS is especially useful for describing characteristics and how they are distributed.

Before debating a number, identify which program actually produced it.

Five years does not mean the final year alone

An ACS five-year estimate pools responses collected over the whole five-year period. The Census Bureau’s period-estimates explanation says it does not represent one specific day within that period.

For example, a table labeled 2020–2024 combines information gathered from January 1, 2020, through December 31, 2024. Calling it simply the town’s 2024 condition can suggest a precision in timing that the product does not provide.

That pooling is useful because it supports estimates for smaller places and geographic areas. It also means a sudden local change near the end of the period may be mixed with several years of earlier conditions.

Read the period before deciding what a figure says about a recent development.

Make comparisons that the data can support

The Census Bureau recommends comparing one-year estimates with one-year estimates and five-year estimates with five-year estimates. For five-year comparisons, it advises using nonoverlapping periods.

Two successive five-year releases share most of their collection years. Treating the difference between them as a clean one-year rise or fall can therefore mislead readers.

Also check whether the question, population being measured or geographic boundaries changed.

A margin of error describes sampling uncertainty. It is not a claim that the figure is useless, and the mere fact that two published estimates differ does not establish a meaningful change.

The bureau provides guidance and a statistical-testing tool for evaluating comparisons.

A short checklist for the next local chart

When a presentation cites income, ask whether it means household income, family income or per-person income. When it cites housing, check whether the denominator is all housing units, occupied units or a particular group.

Preserve the table identifier so someone else can find the same measure.

For a Gretna comparison, also confirm that the geography is the town rather than a ZIP Code Tabulation Area or a larger county area. Those names can sound interchangeable in conversation while representing different boundaries.

An informative chart gives the estimate, its uncertainty, the collection period and the geography. Those details make a small-town statistic useful enough to question and verify.