Axis starts at zero
“Race in dead heat”
y-axis: 0 → 60
H018
How can the same two numbers tell two completely different stories?
Keep the numbers fixed. Change only where the vertical axis begins.
Axis starts wherever you drag it
“Candidate A surges to commanding lead”
y-axis: 48 → 51.5
starts at 48
The real numbers
50.9% vs 49.1% — a 1.8-point gap
VS.
What the truncated chart shows
A's bar looks 2.6× as tall
The numbers never change. The visual story does.
The gap is always 1.8 percentage points. With the axis beginning at
48, Candidate A’s visible bar becomes
2.6× as tall as Candidate B’s.
Try it somewhere else
A cereal ad shows Brand A's sugar bar twice as tall as Brand B's. The chart's axis starts at 10 grams.
Before believing Brand A has double the sugar, what should you check?
Why does moving the axis change the story?
The rule: a bar's length is what our eyes read as its value, so a
bar chart makes an implicit promise that its axis starts at zero. Break that promise
— start the axis at 48 instead of 0 — and a 1.8-point gap gets stretched
until one bar looks more than twice as tall as the other, even though neither number
moved. This is one of the most common chart manipulations in the wild — cable
news graphics, earnings decks, product comparisons — and the tell is the same
every time: check where the axis starts before you read the bars.
The same instinct — pick the framing that tells the story you want — shows up whenever a single summary number stands in for a messier reality, the way one enormous income can drag an “average” income far away from what any typical person actually earns.
Related exhibits
Sources: illustrative polling values, Candidate A 50.9%, Candidate B 49.1%.
Both charts display exactly the same values; only the vertical-axis baseline changes.