LogicalDUAL-USE

Denominator Neglect

What it is

Presenting a count or a numerator — “300 deaths”, “nine winning tickets” — without the total it comes out of, or choosing a ratio format because the audience will weigh the top number and ignore the bottom one.

How it works

Numerators are concrete: people, crashes, winning tickets. Denominators are abstractions: the millions of flights, doses or draws in which nothing happened. Reyna and Brainerd's fuzzy-trace account holds that people reason from the gist — the vivid instances — and neglect the denominator that gives them their weight. The effect is measurable and strange. Denes-Raj and Epstein found that many people preferred to draw from a bowl with 9 winning jellybeans among 100 rather than 1 among 10, knowing the odds were worse, because nine chances felt like more. Yamagishi found that a cancer described as killing 1,286 people in every 10,000 was rated more dangerous than one killing 24.14 in every 100, though the second rate is nearly twice the first. The persuasion move follows: report the numerator alone, or pick the format whose numerator is largest. The same lever runs in reverse when the denominator is chosen to be enormous — “only 0.001 percent” — so that real harm vanishes. The corrective is a fixed reference class: how many out of how many, on the same base for every comparison.

Real-world examples

  • •Tesla's quarterly vehicle-safety reports compared crashes per mile with Autopilot engaged against crashes per mile for all US driving; researchers pointed out that Autopilot is used mostly on highways, where crashes per mile are far lower for every car, so the two numbers had different denominators and the comparison did not show what it appeared to show.
  • •In 2021 counts of deaths reported to the US VAERS system after COVID-19 vaccination circulated as “deaths from the vaccine”; the reports were unverified, the denominator was hundreds of millions of doses, and the background number of deaths expected among that many people in any given week was never shown beside the count.
  • •Opponents of nuclear power cite death counts from Chernobyl and Fukushima; Markandya and Wilkinson's 2007 comparison in The Lancet placed deaths per unit of electricity generated from nuclear far below coal, oil and gas, once the denominator of energy produced is included. The raw count and the rate point in opposite directions.
  • •In 2018 the White House publicized counts of crimes charged against undocumented immigrants in Texas; peer-reviewed rate-based work, including Light and Miller's 2018 study in Criminology, found lower violent-crime rates for undocumented immigrants than for native-born residents. Counts without the population denominator cannot settle the question either way.
  • •Airline safety coverage that reports “the deadliest year since 2014” without departures flown; the denominator of tens of millions of flights is what makes the rate, and the rate is what a traveler needs.

Ethical guidelines

Where the line is

Reporting counts and frequencies is legitimate when the denominator is stated and held constant across comparisons, and natural-frequency formats are often the clearest choice; it becomes manipulation when the denominator is dropped, mismatched between the two sides of a comparison, or selected after the fact so that the top number reads as large or small.

  • ●Never publish a numerator without its denominator, and never a denominator chosen after the fact to make the numerator look large or small.
  • ●Keep one reference class across a comparison: per 1,000 people, per million miles, per 100,000 doses — not one base for your side and another for theirs.
  • ●When using natural frequencies (a legitimate and clear format), fix the denominator and state it in every figure.
  • ●If you know the audience will read the count as a rate, supplying the count alone is a choice to let them.

How to defend against it

  • ►Ask “out of how many?” for every count, and “compared with what base rate?” for every rate. Refuse to update until you have the answer.
  • ►Convert competing claims to the same denominator before comparing them; a difference that disappears on a common base was a difference in bases.
  • ►Watch for mixed formats in one document — “1 in 10” beside “9 out of 100” — and rewrite them on a single scale.
  • ►When a number is presented as tiny (“0.001 percent”), ask how many people that is; when it is presented as huge (“30,000 cases”), ask what fraction that is.
  • ►For safety claims, ask whether the two groups being compared were doing the same thing under the same conditions — the denominator is not just a number but a population and an exposure.

From the Defense Playbook

Ask for the Denominatorseconds

Whenever you are shown a count ("4,000 complaints", "12 deaths", "9 out of 10 dentists"), ask "out of how many?", because a numerator on its own cannot tell you whether something is common, rare, rising, or falling.

Every playbook entry states how strong its evidence is and when not to use it. Browse the full playbook.

References

  1. Reyna, V. F., & Brainerd, C. J. (2008). Numeracy, ratio bias, and denominator neglect in judgments of risk and probability. Learning and Individual Differences, 18(1), 89-107
    The fuzzy-trace account of denominator neglect and its relation to numeracy.
  2. Denes-Raj, V., & Epstein, S. (1994). Conflict between intuitive and rational processing: When people behave against their better judgment. Journal of Personality and Social Psychology, 66(5), 819-829
    The jellybean experiments: participants chose the bowl with more winning beans and worse odds.
  3. Yamagishi, K. (1997). When a 12.86% mortality is more dangerous than 24.14%: Implications for risk communication. Applied Cognitive Psychology, 11(6), 495-506
    Risks expressed with larger numerators over larger denominators were rated as more dangerous than higher rates with smaller numerators.
  4. Markandya, A., & Wilkinson, P. (2007). Electricity generation and health. The Lancet, 370(9591), 979-990
    Deaths per unit of electricity generated by fuel source, the denominator that reverses the raw-count comparison for nuclear power.
  5. Light, M. T., & Miller, T. (2018). Does undocumented immigration increase violent crime?. Criminology, 56(2), 370-401
    Rate-based evidence on undocumented immigration and violent crime, contrasted with count-only claims.
Last reviewed
Suggest a correction

Detect Denominator Neglect in any text

Paste any message, email, or article into our free Manipulation Detector to see if Denominator Neglect or other techniques are being used on you.

Try the Detector FreeCreate Free Account