# Data choices — us-measles-2026 story

**Source.** The CDC files behind the [Measles Cases and Outbreaks](https://www.cdc.gov/measles/data-research/index.html) page, as of 5 November 2026 (see PROVENANCE.md). Every charted and quoted number comes from these files or from sentences in the archived page; nothing is adjusted, smoothed or imputed.

**What a case is.** A confirmed measles case notified to CDC by a state or local health department by noon on the Thursday before the update. Probable cases are excluded. CDC assigns each case to a year by the MMWR (epidemiological) week of rash onset, so a rash in the last days of December can count in either year. The national total includes cases in international visitors (18 in 2026); the state file counts only US residents (5,164 in 2026), by reporting jurisdiction, with New York City separate from New York State. CDC labels 2025 and 2026 preliminary.

**Provisional and incomplete.** 2026 is the year to date. The snapshot was published on Friday 6 November 2026 with reports through noon on Thursday 5 November; MMWR week 44 runs to 7 November, so it is partial, and earlier weeks can also be revised. In the weekly file the most recent weeks are the least complete: reports arrive after rash onset. The post treats the latest two weeks (43 and 44, ending 31 October and 7 November) as still being reported, excludes them from the "about 220 a week" average (weeks ending 5 September to 24 October; mean of 8 weeks, rounded) and shades only those two weeks on the chart. Two weeks is a judgment, not a CDC rule; CDC does not publish a reporting-lag estimate on this page.

**Annual series.** `MeaslesCasesHistory.json` gives 1962–2026 (charted as reported cases: case definitions and surveillance changed over the period, and CDC's older publications, such as the historical table in [MMWR vol. 46, no. 54](https://www.cdc.gov/mmwr/PDF/wk/mm4654.pdf), give slightly different figures for some early years; 1991's 9,643 matches); the separate 1985-on file agrees year for year (checked). Reporting in the 1960s was far from complete (CDC estimates 3 to 4 million infections a year before the vaccine, against 481,530 reported in 1962), so early years are a floor, not a measure of infections. "Most since 1991" means no year from 1992 to 2025 had as many confirmed cases as 2026 has so far; 1991 had 9,643. 2025's 2,289 was itself the highest since 1991 (1992 had 2,126). The comparison ignores population growth: the US population is larger now than in 1991, so the rate per person in 2026 is further below 1991's than the counts suggest.

**Weekly series.** Weeks run Sunday to Saturday. The MMWR year of a week is the year that contains its Wednesday; with that rule, weekly sums equal the annual totals for every year 2022–2026 (checked), including 2025's 53rd week (28 December 2025 to 3 January 2026). Week numbers are counted from the first week of each MMWR year. "3.1 times" compares cumulative totals at week 44 of each year (to 7 November 2026 and to 1 November 2025).

**States.** The top-four share (60%) is of resident cases. CDC reports 47 jurisdictions with cases, matching the count of non-zero rows here.

**Summary figures.** Age groups, vaccination status, hospitalisations and deaths come from CDC's summary file (`measles_hosp.json`) as published; "58% under 20" adds the under-5 and 5–19 groups (908 + 2,104 of 5,182). "96% unvaccinated or unknown" is CDC's combined category; it is not the share known to be unvaccinated. Deaths are not quoted: the data file gives 7 confirmed deaths in 2026, but the archived page's own HTML shows a hard-coded "4*" in the 2026 deaths cell (the 2025 cell is filled by script), with a note that the number is subject to change and that CDC is developing a standard case definition for measles deaths with CSTE. The two disagree, so the post leaves deaths out.

**Outbreaks.** CDC defines an outbreak as three or more related cases; states may count outbreaks differently, and CDC may fold a state's cluster into an earlier outbreak. Counts (43 new in 2026, 48 in 2025) and the outbreak-associated share are quoted from the page text.

**Kindergarten coverage.** State estimates from CDC SchoolVaxView, school years 2009–10 to 2025–26. Coverage measures documented completion of the kindergarten MMR series, generally two doses, rather than exemption status; exemptions can cover one or more vaccines and are not the complement of MMR coverage ([CDC definitions](https://www.cdc.gov/schoolvaxview/about/index.html)). The estimates concern kindergarten pupils, not whole-state populations. The 2019–20 estimates exclude Alaska, Delaware and DC (not reported); 2025–26 excludes Kansas and Montana. Houston and New York City are excluded as separate programmes. Medians are rounded half up after removing floating-point noise (2019–20: 94.55, shown as 94.6); no state has an estimate for 2010–11 in this file. Survey methods vary by state and year (census, random sample, voluntary response), and the number of reporting jurisdictions varies (48 states in 2019–20, DC not reporting; 48 states plus DC in 2025–26). The chart and the median use the 50 states and DC; Houston and New York City, which appear as separate immunisation programmes, are kept in the CSV (`state=false`) but excluded. The median is of state estimates, unweighted, so it differs from CDC's national figure (95.2% in 2019–20, 92.4% in 2025–26), which the post quotes separately. "Below 95%" counts estimates strictly under 95.0.

**What the data cannot show.** Case counts, state totals and state-wide coverage cannot show why cases rose, or whether one chain of transmission lasted 12 months, which is what PAHO's elimination review assesses. Explanations in the post are attributed to CDC and PAHO.

**Size.** All source files are archived here: about 400 KB in total.