A separate legend forces the eye to travel back and forth for every single mark it identifies

What it is

A legend parks the colour-to-category mapping outside the chart, so every identification forces a round-trip: look at the mark, memorise its colour, travel to the legend, scan the entries, travel back. The trip is never free — each gaze shift carries a fixation-switching cost, and the total multiplies across categories and identifications. The farther and longer the legend, the more every lookup costs.

Why it happens

The round-trip cannot be skipped because visual working memory holds only three or four candidate features: readers cannot carry the whole legend in their head while scanning the chart, so every unfamiliar colour must be verified on the spot. Identification that could run in parallel is thus chopped into a serial five-step loop, and the cost of "one glance" becomes "one journey." Eye-tracking studies capture this cost through fixation counts and travel paths: under legend conditions the gaze shuttles back and forth, while direct labels keep the gaze inside the plot — this comparison is the standard instrument of legend-cost research.

Studying it

The standard paradigm is a two-condition contrast: the same chart rendered with a legend versus direct labels, readers performing category naming or value reading, with task time, error rate, and gaze-transition counts recorded. The independent variable is label position (legend / line end / beside point); the dependent variables are round-trip frequency and completion time. Mind the ecological validity: first-reading in the lab amplifies legend cost, whereas practised users pay less — conclusions should separate first-encounter cost from long-term use cost.

Where it stops holding

The round-trip cost decays with familiarity: daily users of the same dashboard have memorised the mapping and the legend is furniture; with only one or two categories the trip is barely felt. Interactive hover can bypass part of the journey, but it adds an action and resolves one point at a time. The cost is structural — it does not vanish because "users will get used to it"; habit merely moves the cost into memory.

Applying it

• Default every new chart to direct labels; enable a legend only where labels will not fit.
• Count legend entries against marked elements: large products (many categories, many marks) mark the highest round-trip cost and the strongest case for labelling.
• Verification: have new readers name every series on the legend-free version; a perfect score means labelling succeeded — whatever they cannot name is still legend-dependent.

Related

• Same group: U5.01.2 Direct labelling eliminates the round-trip · U5.01.3 Legend order should match the graphic order
• Nearby: U5.02.1 Legend-to-chart distance sets the cost of each round-trip · U5.02.4 Too many legend entries degrade matching into a linear search
• Search terms: direct labelling · legend cost · eye tracking visualization