Why Averages Lie — Visual Guide without getting fooled.
A visual checklist for reading viral charts: find the denominator, inspect the distribution, separate rates from counts, and name the caveat before repeating the claim.
Fig. 1 — A single summary number can miss the shape that changes the story.
chartedtruths.com/guides/averages-lie
01 What the chart directly shows
02 What the denominator changes
03 When the average hides the distribution
04 Rate, count, and base-rate traps
05 Correlation, causation, and implied claims
06 The pre-share checklist
The mental model
Every chart is an argument about what counts.
The first job is not to decide whether a chart is “good” or “bad.” The first job is to translate it into a sentence precise enough to test. What population is being counted? What is the unit of analysis? What time window is included? Which denominator would make the headline smaller or larger?
Once those questions are clear, the chart becomes easier to read. The visual may still be useful, but it can no longer smuggle in a broader claim than the data supports.
A good chart should make the comparison clear, name the source, show the denominator, and include the caveat that changes how far the claim can go.
Side note · the three-sentence test
Before sharing a chart, write three sentences: what it directly shows, what it does not prove, and what denominator would change the interpretation. If those are hard to write, the chart needs a slower read.
DenominatorWho is included, and who was left out?
DistributionDoes the average hide two different groups?
Base rateIs the event rare even after the example feels vivid?
CaveatWhat would make the claim narrower or less confident?
Field guide
Use the checklist on the actual claim.
The Data Literacy Cheat Sheet is the starter map for Charted Truths. It teaches the small set of questions that protect a viewer from most viral chart mistakes: what is the denominator, what is the unit of analysis, is the average hiding the distribution, does the chart show a rate or a count, and does the claim imply causation when the data only supports association.
Use it before believing a chart, sharing a chart, or turning a chart into an argument. The point is not to become cynical. The point is to become harder to fool.
1. Name the claim before judging the chart
Write the claim as a plain sentence first: "This chart shows that X is higher than Y for Z group during T period." If you cannot fill in X, Y, Z, and T, you are not ready to argue from the chart.
2. Find the denominator
The denominator is the group underneath the percentage. A chart that says "40 percent" is incomplete until you know 40 percent of what: users, accounts, survey respondents, households, games, arrests, claims, or visits.
3. Check whether the average hides the shape
Averages are useful when the group is compact. They become dangerous when the group has two peaks, extreme outliers, or a long tail. If the distribution is unavailable, use the narrower sentence: "The average moved" rather than "most people experienced this."
Reading traps
Counts, rates, and causal claims.
4. Separate counts from rates
Counts answer how many. Rates answer how often relative to the population or opportunity. A city with more people can have more events and still have a lower rate. A rare event can double and remain rare.
If the chart shows...
Ask for...
Why it matters
Total events
Events per relevant population
Bigger groups naturally produce bigger counts
Percent change
Starting value and ending value
A huge percentage can come from a tiny base
Ranking
Margin between ranks
Adjacent ranks may be statistically indistinguishable
One vivid example
Base rate
Vivid examples can overwhelm rare-event math
5. Watch for implied causation
Two lines moving together do not prove one caused the other. Ask whether the chart rules out seasonality, selection effects, policy changes, measurement changes, or a third factor driving both lines.
6. Do the pre-share check
What the chart directly shows.
What population or denominator it uses.
What caveat would change the interpretation.
Whether it shows a count, rate, average, distribution, or relationship.
Whether the headline claims more than the chart proves.
Keep going
Where this guide came from
Why Dating Apps Feel Broken — Explained With Data chartedtruths.com/library
Source discipline
Charted Truths treats sources, denominators, and caveats as part of the product. Every published guide should trace back to the video folder, source manifest, chart spec, and review notes.
Next step
Patterns Visualized Foundations — Learn the statistical ideas behind the charts you see every day — without getting buried in equations.
Charted Truths