Methodology

We would rather be transparent than impressive. Here is exactly what these numbers are, what they are not, and how we keep the site honest.

Data sources

• CDC WONDER — Underlying Cause of Death, the detailed national mortality database. We pulled ages 0–24 by single-year age × injury intent × mechanism for 2018–2020, plus 1999–2020 trend series.
• CDC WISQARS / NVDRS — the injury- and violence-specific views of the same system.
• Agency & clinical guidance — AAP/HealthyChildren, NHTSA, CPSC, NFPA, Poison Control, the 988 Lifeline, SAMHSA, and topic authorities, cited on each page.

Fatality figures are a live CDC WONDER pull (Underlying Cause of Death, D76, 2018–2020) across ages 0–24, plus 1999–2020 trend queries — run through an authenticated browser session because the API blocks scripted access. The ladder ranks mutually-exclusive, intent-primary cause groupings per age band so rungs never double-count; firearms is shown cross-cutting because it spans suicide, homicide, and unintentional deaths.

Life-stage bands

Newborn (under 1), Toddler (1–4), Little kid (5–9), Big kid (10–14), Teen (15–19), and College & new grad (20–24). We split childhood this way because the dominant causes of death change completely at each step — averaging them together hides exactly what a parent needs to know at each age.

How the ladder is built — cause groupings, not double-counting

Each rung is a mutually-exclusive, intent-primary cause grouping: unintentional mechanisms (motor vehicle, drowning, fire, falls, suffocation, poisoning) plus suicide and homicide as their own groups. Because every death lands in exactly one rung, the bars can be compared and summed honestly.

Firearms is shown cross-cutting, never as a rung. A firearm death is already counted under suicide, homicide, or an unintentional shooting. Adding a separate "firearms" rung would double-count those same deaths. So we surface the firearm total as a card — because it is the single most important mechanism to address at older ages — while keeping the ladder math clean.

The household-poisoning cause maps to two different pages by age: cabinet-latch prevention for young children, and the fentanyl/naloxone overdose reality for teens and young adults — the same ICD bucket, two different real-world stories.

Fatal vs. nonfatal — the big limitation

Deaths are clean, public, and comparable, so we start there. But death is the tip of the harm: for most of these hazards, far more children are injured — sometimes permanently — than die. We flag this everywhere and will layer in nonfatal data later.

Suppression & false precision

WONDER suppresses death counts below 10 and flags rates on small numbers as unreliable. We don't back out suppressed cells. Pooling three years and grouping into five-year bands keeps the big causes robust; rare single-age cells that suppress are small enough not to change any ranking.

Classification caveats

Cause-of-death coding is imperfect. SUID mixes SIDS, suffocation, and undetermined causes; many drug deaths are coded "undetermined" intent; suicide and homicide counts depend on death-investigation quality. We report umbrellas and sub-categories separately and note where the line is fuzzy.

Independent verification

The load-bearing statistic on each hazard page was re-checked by a separate research pass instructed to refute it. The verdict — confirmed, partly, refuted, or unverifiable — is published on the page, corrections and all.

A note on the hard topics

Suicide, overdose, and violence pages follow safe-messaging practice: no method detail, a hope-first frame, and crisis resources up top (988; 1-800-222-1222). These are leading causes of death for older youth, so leaving them off would be the dishonest choice.

Editorial firewall

No paid rankings, no sponsored placement, no fear copy, no dark patterns, no claim stronger than the evidence. See About.