Line charts suit trends over ordered variables

What it is

Line charts suit questions about how a value changes over an ordered domain: the horizontal axis may represent time, distance, dose, or explicitly sequenced stages, while the vertical axis represents magnitude. The connected path supports judgments of trend, turns, and rate of change. Time is ordered and usually has measurable intervals; an ordinal variable supplies order without necessarily supplying comparable distance. “Suit trends” therefore does not make every ordered category a valid rate axis, nor does it mean the intervals between observations were continuously observed.

Why it happens

Connections organize adjacent observations into one sequence, letting readers follow direction and local change without retaining and subtracting every value. Slope depends jointly on vertical change and horizontal interval; steepness represents rate only when both distances are meaningful. Segments also invite interpolation and an impression of process continuity. That inference makes trends legible but can exceed the evidence. Observed, estimated, interpolated, and missing values must remain distinct data states.

Studying it

Compare lines, unconnected dots, and bars for the same sequence, separating overview direction, turn localization, point read-off, rate comparison, and gap detection. Manipulate sampling, horizontal spacing, point density, aspect ratio, noise, and segment treatment; measure error, response time, confidence, and interpretations of the process between observations. For ordinal axes, test ordering and slope judgments separately to detect whether equal visual slots are being mistaken for equal quantitative intervals. Evidence for faster trend recognition does not establish better exact-value reading, and results transfer only to the tested task and scale.

Where it stops holding

Two observations can express a direction or difference, but cannot by themselves support a complex shape or robust trend; the evidence needed depends on sampling, noise, and the question rather than a fixed count. Step changes, event counts, and interval aggregates may call for step lines, points, or bars. Smoothing, cycles, and forecasts introduce model assumptions and must not masquerade as raw observations. A nominal axis, or evenly spaced dates with unequal real intervals, can make path shape an artifact of arrangement.

Applying it

• Record whether the horizontal variable is temporal, metrically continuous, or ordinal only. Call slope a rate only when interval magnitude is meaningful.
• Place observations at true horizontal intervals and distinguish observations, interpolation, forecasts, and missingness. Break paths across long gaps or definition changes.
• Add direct labels or details on demand, plus a keyboard-accessible data table, series summary, and state descriptions so trend is not available only as line shape.
• Validate overview, point-value, rate, and gap tasks separately. Change the mapping or explanation if readers treat ordinal slots or connections as a continuously observed process.

Related

• Same group: U2.03.2 Never connect a categorical axis with lines · U2.03.3 Beyond a few lines, tracking collapses
• Adjacent: U1.13.2 Connectivity makes a line read as one continuous process · U3.06.1 Uneven sampling plotted at even spacing distorts rates of change
• Search terms: line chart · ordered domain · temporal trend · ordinal scale · interpolation