Two claims often framed as political theater are, in fact, statistical milestones: real median household income reached a modern-series record in 2025, and the 2025 homicide rate fell to a level last seen in the mid-1950s. The evidence is stronger than the noise—if you respect the definitions, the series, and what each measure can and cannot say.
At a Glance
- Real median household income was $87,460 in 2025—the highest since the Census series began in 1967.
- The national homicide rate in 2025 was about 4.1 per 100,000 people, tying mid-1950s lows in the FBI record.
- Both facts are narrowly defined: “median” is not “everyone,” and a homicide rate is not a count of murders.
- Distributional gaps persist: lower earners saw weaker gains even as the median set a record.
- Historical comparisons in crime are credible at the rate level but still hinge on changing data coverage and methods across eras.
What the numbers actually say—and why the definitions matter
Median household income is the income right at the midpoint—half of households earn more, half earn less—after adjusting for inflation with a standard price index. In 2025, the Census Bureau reports that figure at $87,460, the highest level recorded since the modern series begins in 1967. Treating that as anything other than a records-based statement about the median misunderstands the measure. It does not imply that every household is thriving, that the mean is at a record, or that distributional gaps have closed; it says the midpoint moved up to a new high in the available series. That framing is accurate and well-supported by the primary data custodian itself.
Homicide statistics require a parallel discipline. The FBI’s key public-safety benchmark is a rate—offenses per 100,000 population—because populations change over time. In 2025, major-wire recaps of the FBI release place the homicide rate at about 4.1 per 100,000, tying 1955 and 1956 and described as the lowest in roughly three-quarters of a century. That is a substantial public-safety improvement by national standards and a level that aligns with mid-century benchmarks in the Bureau’s own historical tables.
How we got here: the series behind the claims
The income figure rests on the CPS ASEC survey and long-standing Census procedures for inflation adjustment and imputation; the “record” qualifier is bound to the 1967–present continuity of that series. Because Census anchors the claim in its own historical table, the comparison set is clear and auditable. The core trend over the last decade includes a pandemic-era shock, subsequent recovery, and then enough real wage and employment momentum by 2025 to lift the median above prior peaks. The rate-based homicide trend shows a sharp rise beginning in 2020, followed by an unusually steep decline that, according to national compilations from the FBI, returned the country to mid-1950s levels by 2025. These are not cherry-picked anecdotes; they are full-series results that survive standard scrutiny.
Context complicates—but does not contradict—these toplines. Median income can set a record while inequality widens, because the median is insensitive to gains at the very top; distributional tables often show stronger long-run growth at the 90th percentile than at the 10th. Several outlets, summarizing the same Census release, highlighted that lower earners saw weaker or statistically insignificant gains even as the midpoint rose—an important nuance for understanding lived experience alongside the headline.
Common misreadings: where rhetoric outruns the metric
There are two predictable pitfalls. First, translating “record median household income” into “record incomes for everyone” is incorrect; the pluralizing move implies a universality the median does not claim. The measured gain is real, but it is about the middle, not the entire distribution. Second, reporting the “lowest homicide rate in decades” as “the fewest murders” invites confusion between a rate and an absolute count; with a much larger population today than in 1955, today’s count can be higher even while the rate is comparable or lower. Experts and careful fact-checks have repeatedly corrected these semantic drifts without disputing the underlying figures.
A third, subtler error is to push historicity beyond what the data can bear. Saying “lowest in 75 years” aligns with the FBI’s time series that modernizes around 1960 and makes cautious linkages to earlier decades. Claiming precision across 125 years, by contrast, presumes uniform comparability across eras that used different definitions, coverage, and data systems. The safer statement is the one the FBI-linked reporting makes: the 2025 rate ties mid-1950s lows in its own record and is exceptionally low by any modern standard.
Mechanics and comparability: why these measures are trusted
Census income statistics are built on large-sample surveys with transparent deflators and revisions. They are not immune to sampling error or methodological updates, but they are the accepted baseline for incomes at scale and over time. When Census says the 2025 real median household income is a series-high since 1967, that is the authoritative keeper of the series rendering a verdict. Secondary coverage that repeats the figure and the “record” label is functionally an echo of the primary source, not a contested interpretation.
The FBI’s Uniform Crime Reporting program, increasingly fed by the incident-level NIBRS system, aggregates from thousands of law-enforcement agencies. Participation gaps and definitional changes are real operational issues, but the homicide rate is among the most robust crime indicators: homicides are comparatively well-reported, less subject to definitional gray zones than other offenses, and tracked consistently for national trend work. When the national rate prints at roughly 4.1 per 100,000, analysts can say—with appropriate caveats about historical comparability—that the country is at or near its modern-era floor.
“He tells the unvarnished truth.” 😂 Mike, the Washington Post documented 30,573 false or misleading Trump claims in his FIRST four years alone. Fact-checkers are literally correcting claims from the UN speech you’re praising TODAY. This is also a man with 34 felony convictions…
— CS (@CRS195610) September 22, 2026
What the milestones mean—and what they do not
For economic well-being, a record median is a strong signpost of broad improvement after a turbulent period; it typically coincides with tight labor markets, rising real wages for at least the middle of the distribution, and household balance sheets that, on average, are sturdier than a year earlier. For public safety, a homicide rate at mid-century lows demonstrates that the pandemic-era spike was not a new permanent plateau and that prevention, policing, and community dynamics can move lethal violence sharply in either direction.
But neither milestone resolves the distributional or structural debates. Households below the median can still face stagnant wages, high housing costs, or regional shocks. Communities with historically high homicide burdens may not experience national averages. Policymaking that treats a national median or national rate as the lived reality for every group will miss these heterogeneities. The right reading of the data is twin-track: acknowledge the national achievements—and then interrogate who is still left out and why.
How to read future claims on the same terrain
Apply three tests. First, insist on the exact metric: median, mean, percentile, or rate per 100,000. Second, anchor to the series owner when available: Census for income, FBI for crimes. Third, contextualize the comparison window: “record since 1967” for income, “tied with mid-1950s in the FBI record” for homicide. If a claim passes those tests, it is likely sound even if political rhetoric around it is exuberant. If it swaps the metric, stretches the historical window beyond the series’ comparability, or confuses rates with counts, demand a rewrite.
Sources:
census.gov, wbiw.com, finance.yahoo.com, cbsnews.com, axios.com
