Age bands or rating tiers should keep their natural order even if sorting by value looked tidier

What it is

Categories with a recognised inherent order — age bands, satisfaction levels, sizes, rating tiers — must keep that order on the category axis, however "prettier" a value sort might look. Re-sorting ordered categories by value smashes the semantic ladder: the "which step" readers infer from position no longer tracks the grade order, and the stair-step structure (monotonic patterns across grades) is replaced by an artificial zigzag — "satisfaction rises with age" turns into an unreadable sawtooth.

Why it happens

For ordered categories, position is itself an information channel: readers assume left-to-right means low-to-high grade and read stair-steps, breakpoints, and monotonicity from it. After a value re-sort, the position channel is hijacked by the value channel — position's meaning silently switches from "grade" to "size," but readers are not told and keep reading grades, producing a distorted false staircase. The damage happens at the semantic layer, not the visual layer, so it never raises an "ugly chart" alarm; it takes effect as "readers draw wrong conclusions." That is exactly why it is more dangerous than scrambling.

Where it stops holding

An ordered-category chart can have a legitimate value-sorted twin — when the task genuinely is ranking rather than reading the distribution across grades ("which age band is largest?"). But that is a different chart: the grade-order version answers distribution questions, the value-sorted version answers extremum questions; present both, each with its sort basis labelled. One chart carries one order; mixing them, or quietly switching the basis, is the most error-prone form.

Applying it

• Maintain a list of ordered categories (grades, age bands, sizes, frequency tiers); any category axis hitting the list is barred from value sorting and keeps its intrinsic order.
• When a value ranking is genuinely needed, produce a separate chart titled "sorted by count," distinct from the grade-order version.
• Verification: ask readers to describe "the pattern across grades"; a description contradicting the data (a monotonic series read as zigzag) means the order was destroyed — restore the intrinsic order.

Related

• Same group: U5.04.1 Category-axis order is a designer's choice, not a property of the data · U5.04.2 Value sorting makes rank and extremes readable without search · U5.04.4 Side-by-side charts need a shared order for cross-chart comparison · U5.04.5 State the sorting basis in the axis title or caption
• Nearby: U5.04.1 Category-axis order is a designer's choice, not a property of the data · U4.02.2 Rainbow palettes manufacture false boundaries
• Search terms: ordinal categories · semantic order · ordinal axis sorting