Our sources

We only publish numbers we can trace to a primary source: the original study, official statistics, or the organization's own reporting. Secondary retellings, influencer polls, and unattributed "studies show" claims are rejected at the fact-gate.

Primary sources currently cited

• National Center for Family & Marriage Research (NCFMR), Bowling Green State University — Family Profile FP-24-04 (2022)NCFMR analysis of IPUMS-USA American Community Survey 2022
• Pew Research Center — "8 facts about divorce, marriage and remarriage in the United States" (2025)Pew Research Center short read, published 16 October 2025; states the widowed/divorced distinction when explaining gender gaps in remarriage
• Pew Research Center — "The Demographics of Remarriage" (Chapter 2 of Four-in-Ten Couples Are Saying "I Do," Again) (2013)Pew Research Center analysis of 2013 American Community Survey data (report published Nov 2014)
• U.S. Census Bureau — "Marriage, Divorce, Widowhood Remain Prevalent Among Older Populations" (2016)2016 American Community Survey (published 2021)

What we rejected, and why

This list matters as much as the one above. These are figures that circulate widely in writing about widowhood and that we could not trace to a primary source, so they do not appear anywhere on this site.

• "There are ~11.5 million (or 11 million / 14 million / 15 million) widows and widowers in the United States." — Could not be read on a primary source in this run. data.census.gov table S1201/B12001 is JavaScript-rendered and the Census data API now requires an API key, so no absolute national count was actually verified. Every circulating version of this number traces to aggregator/roundup pages. Use the NCFMR and Census percentage figures above instead of a headcount.
• "The median/average age of widowhood in the US is 59 (or 55, or 60)." — Repeatedly cited online but not found on Census, NCHS, or NCFMR in this run. NCFMR's widowhood profile (FP-20-23) explicitly does not report a median age at widowhood. Reject until a primary source is located. Safe rephrase: cite the Census finding that ~70% of the recently widowed are 65+ (entry wid-recent-age-concentration-2016).
• "X% of widows/widowers remarry" or "widows remarry within Y years on average." — No primary source found that isolates the widowed from the divorced. Pew's 64%/52% figure combines both groups. Do not convert it into a widow-specific rate. Safe rephrase: "Pew Research Center reports that widows are less likely than divorced adults to remarry" (no number).
• "Two-thirds of Americans who divorce remarry (68% of men vs 64% of women)." — Real Pew 2025 figure, but it describes DIVORCED Americans only. Out of scope for a widowhood site and highly likely to be misapplied to widows. Not approved for use here.
• Any grief-timeline, loneliness-prevalence, online-dating-usage, or dating-platform-pricing statistic. — Explicitly out of scope for this run — nothing was researched or verified. Numbers on these topics must not appear in drafts until a separate fact-gate run approves them.
• "Pew Research Center notes that widowed adults are less likely than divorced adults to remarry" — used on /dating-a-widower/the-hard-parts/ to indicate what a widower is likely to want long term. — Wording drift from the source. Pew's sentence says "widows", inside an explanation of why previously married WOMEN remarry less than men. Broadening it to "widowed adults" and applying it to a widower is not supported. Fix: attribute with the year and Pew's own subject ("Pew Research Center reported in 2025 that widows are less likely than divorced adults to remarry"), or drop the clause on a page about a man.
• "Pew separately notes that widowed adults are less likely than divorced adults to remarry" — recurrence on /dating-a-widower/why-people-say-never-again/ (re-check 2026-08-13). — Same wording drift as the /the-hard-parts/ entry, and the sentence also carries no year. Pew (16 Oct 2025) writes "widows", i.e. women. Required fix: "Pew Research Center wrote in 2025 that widows are less likely than divorced adults to remarry, and that women are much more likely than men to be widowed" — and the page must not read that back as a prediction about widowers.
• "Widowed men in the older age brackets are, in the plainest terms, outnumbered" — on /dating-a-widower/why-people-say-never-again/, derived from NCFMR 29.5%/10.3%. — Derived count comparison. NCFMR publishes shares within each sex, not a headcount of widowed men vs widowed women. The direction is plausible but the site does not publish a verified count, so the inference must not be stated as fact. Safe rephrase: state the two NCFMR shares and stop, or say "a much smaller share of older men than older women is widowed".
• Number-free frequency claims on /dating-a-widower/when-a-widower-pulls-away/ (re-check 2026-08-13): "it is more often overwhelm than a decision" (directAnswer), "the first is the rarer explanation and the second the ordinary one", "the other common cause", "a common time for a widower to pull back", "what has often happened", "a common version", "Many people describe the second relationship as an addition rather than a replacement" (FAQ). — Each is a prevalence claim with the digits removed, and nothing in approved_stats.json covers withdrawal behaviour or how widowers describe a second relationship; no study was located for either. Fix: mark explicitly as the editors' reading (the page already does this once — "In the editors' reading it is capacity, not a verdict") or rewrite without the frequency, e.g. "One version people find hardest to read is the retreat that follows a good evening" and "A second relationship does not have to be a replacement".
• Universals/absolutes on the same page: "He has heard this from everyone since the funeral" (body + FAQ), "no amount of patience converts one into the other", "Nobody can answer it, including him." — Population-wide absolutes that cannot be sourced or falsified. Safe rephrase: "he has almost certainly heard it already"; "patience on its own is not what changes this pattern"; "it is not a question he can answer either".
• Clinical statements on the same page, unsourced: "Grief is an experience, not a diagnosis"; the threshold "prolonged grief that has not shifted at all over many months"; the warning-sign list (not eating or sleeping for weeks, cannot work, drinking through it); and "The retreats get shorter over months" as a description of ordinary grief. — Prolonged grief disorder IS a diagnosis in DSM-5-TR, so the first clause is inaccurate as written; the duration threshold and the warning-sign list are clinical criteria presented with no authority behind them; the retreat trajectory is an unsourced prognosis. Fix: run a separate fact-gate pass for a clinical source (DSM-5-TR, NIMH, or NHS bereavement guidance) and cite it in-sentence, or drop the diagnostic framing and the month threshold and keep only non-quantified referral advice (his doctor, a bereavement service, and in the U.S. the 988 Suicide & Crisis Lifeline rather than an unnamed "local crisis line").
• "A widowed man of that age often has no one in his social circle who is in the same position" — attached to the NCFMR 29.5%/10.3% figure on /when-a-widower-pulls-away/. — Source extension. NCFMR measures the share of people who are widowed, not the composition of anyone's friendships. Safe rephrase that the data does support: "far fewer men his age are in that position than women his age are".
• "in the same figures it falls far more often on women than on men" — /dating-a-widow/what-widowed-means-in-a-relationship/ (re-check 2026-08-13), attached to the NCFMR and Census figures cited just above it. — Same family as the 'outnumbered' rejection: an INCIDENCE claim ('falls far more often on') attributed to sources that publish neither incidence nor counts. NCFMR publishes the share currently widowed within each sex; the Census 71%/69% pair is an AGE distribution of the newly widowed and, read straight, shows men and women alike. Two acceptable fixes: (a) share language — "a far higher share of older women than older men is widowed"; or (b) attribute the direction to the source that actually publishes it — "Pew Research Center wrote in 2025 that women are much more likely than men to be widowed".

The rules

1. Named source in the same sentence as the number, with the year.
2. Primary source only — we link the origin, not an article about it.
3. Dated verification: every statistic carries its verification date on the statistics page.
4. Unverifiable → unpublished. No exceptions.