Research notes · Updated 6 September 2026
What is the safest state in the US?
The answer depends on the measure
There is no single safest state across every kind of risk. For a crime-data comparison, choose an offense, a reporting year and comparable coverage first. Safety Atlas's current selected city profiles do not establish a comprehensive national state ranking.
Why we do not name a nationwide winner here
An earlier version named Vermont first in the Safety Atlas index. That statement was not supported by a consistent statewide comparison in the current public data, so we have withdrawn it. A median of selected city scores, a state's reported offense rate and a composite state score are different measures. None should be substituted for another.
The Vermont directory remains useful for its published jurisdiction profiles, with their source years and limitations. The Vermont sample lets you inspect a report format. Neither link is evidence that Vermont is the safest state in the country.
Choose the question before choosing a ranking
| Your question | Evidence to compare | What it cannot establish |
|---|---|---|
| Which state reports less violent crime? | The same violent-offense definition, year and coverage basis for each state | Property crime, road safety or an individual person's risk |
| Which place reports less property crime? | Comparable offense-specific rates and reporting participation | Whether every neighborhood has the same pattern |
| Where should I research a move? | Local geography, source dates, reported offenses and other needs you identify | A universal recommendation from a single crime score |
A repeatable state comparison
- Select a complete reporting year. Keep release date separate from the year incidents occurred. A 2026 edition can describe earlier reporting.
- Choose the same measure. Record offense definitions and whether figures are agency-reported counts or estimates. Avoid combining overlapping categories.
- Check participation. Record missing agencies and reporting months. Missing values are not reported zeros; agency coverage and population coverage are different denominators.
- Check the population. Rates per 100,000 require a population matched to the reported geography and period. A city's population cannot normalize a county count.
- Keep the comparison cohort. State how many places qualify and why others are excluded. Do not label a selected group as a complete statewide ranking.
- Retain source and retrieval date. A reader should be able to identify the same release, definition and calculation.
Begin with the FBI Crime Data Explorer and its release notes. The FBI's NIBRS overview explains incident reporting and how it differs from summary totals. Our NIBRS and UCR guide explains the terminology used in these checks.
How to use Safety Atlas profiles
Start at the state and city directory, open a profile, and read its offense table alongside the score. The current scored set is identified on the page. Keep its source year and population with any number you cite, and consult the methodology before comparing adjusted scores with unadjusted rates.
For a report or procurement decision, use the crime-data evidence checklist. It helps distinguish an auditable comparison from an unsupported safety label.
Frequently asked questions
Does a low crime rate mean no crime?
No. It summarizes reported incidents for a defined geography and period. It does not predict future incidents or describe every resident's experience.
Why can two rankings disagree?
They may use different years, offenses, weights, populations, coverage rules or non-crime measures. Compare those definitions before comparing rank numbers.
Can a few city profiles establish a state ranking?
No. Selected profiles do not represent every jurisdiction. Their scores can support comparisons within the disclosed set, not a comprehensive state conclusion.