Untrained readers misread logarithmic axes

What it is

The log axis swaps the comparison semantics from addition to multiplication, while most readers' default reading is linear: equal spacing read as equal difference. Systematic misreadings follow — 10-to-20 and 100-to-200 occupy equal length on the axis (both are doublings), yet readers call it "same growth" or, seeing it flagged, accuse the chart of error; exponential curves straightened by the log lead readers to judge growth by curvature and conclude "flattening later" (the truth is accelerating); the vast absolute differences near the axis end read as "nearly the same." The log axis is not wrong; the default reader is: it is a tool for people who know what they are looking at.

Why it happens

The root is reading-habit transfer: the distance-to-difference conversion trained over a lifetime under linear axes is the default; unless the log labelling is actively noticed, readers keep the old conversion, with error amplifying across decades (the 10⁰-to-10¹ segment understated tenfold against 10³-to-10⁴). Curvature misreading is subtler: growth perception leans on how much a curve bends, and logging removes the cue by straightening the exponential — readers either read linear growth (underestimating) or improvise a trend near the straight line they see. None of this self-reveals: the chart looks "normal," and the error happens in the conversion layer.

Where it stops holding

Misreading risk should not condemn log axes wholesale — research and finance audiences speak logarithm natively, and cross-magnitude data leaves them no alternative. Risk stratifies by audience: cautious in mass communication (or heavily scaffolded), confident in expert circles. Mitigations in ascending strength: prominent "logarithmic scale" labelling (next leaf), raw-value ticks with "equal spacing = tenfold" in the caption, direct value labels at key points, and a companion linear inset. The ultimate mitigation is not using it — when the audience cannot carry it, return to facets.

Applying it

• Before publishing, ask who: for general audiences with uncertain log literacy, deploy the full scaffold — log axis, raw-value ticks, reading hint, linear inset — or switch to facets.
• Templatize the caption: "the horizontal axis is logarithmic; adjacent equally spaced ticks differ tenfold" — put the conversion rule in the title rather than in reader knowledge.
• Verification: give the target-audience sample two axis-reading tasks (ratio comparison across equal segments, judgement of a straightened exponential); high error means the scaffold is insufficient or the log should go.

Related

• Same group: U3.03.1 Log scales suit data spanning orders of magnitude · U3.03.3 The scale type must be prominently labelled
• Nearby: U3.02.5 Facets or log scales usually replace the broken axis · U2.17.2 One question, several viable charts — they differ in reading cost
• Search terms: log scale literacy · misreading · exponential growth perception · graph literacy