← Evaluate statistically based reports

7 Evaluating graphs and visual presentation

A resource booklet always includes at least one graph, and spotting a genuine visual distortion is a fast, high-value evaluation point — separate from the numerical evaluation of the survey itself.

What to check on any graph in a report

  • Does the vertical axis start at zero? A truncated axis exaggerates small differences dramatically.
  • Is the scale uniform throughout, or does it change partway (compressing or stretching part of the range)?
  • If a picture or icon represents a quantity, does its AREA (not just height) scale with the value? Doubling both dimensions of an image quadruples its visual impact.
  • Was the time window shown chosen to favour a particular conclusion (starting or ending at a convenient point)?
  • Does the graph match the claim being made in the accompanying text, or does the text overstate what the graph actually shows?
Exam-style Identifying a truncated axis A bar graph shows a company's sales "soaring" — the two bars go from a y-axis starting at 98 to 102 units. Evaluate this presentation. Identify the axis range: 98 to 102, not starting at zero. Calculate the real percentage change: if sales moved from 99 to 101, that is only a 2% increase. State the effect: because the axis is truncated, the two bars look dramatically different in height even though the underlying change is small — the visual exaggerates the real-world significance of the change. ✓ Naming the specific axis range AND calculating the real percentage change is what turns "the graph looks misleading" into an evaluated, mark-earning observation. ✕

Saying a graph is "misleading" with no specific mechanism named

A bare claim that a graph is "misleading" or "exaggerated" earns little — the answer needs to name the SPECIFIC visual mechanism (a truncated axis, a non-uniform scale, an area-scaled icon) and ideally quantify the real change it is exaggerating.

Practise just this