Fluency and cognitive load

Processing fluency

How a visual looks affects what people think of its content. When something is easy to read, people are more likely to trust it and accept its message without question. This cuts both ways: a smooth, familiar design can make readers overlook mistakes or misleading elements.

The research suggests:

  • Readability supports accurate interpretation. Legible colours, clear contrast and a plain, legible typeface make information easier to process, and more trusted.
  • Ease biases judgements of truth. People read the feeling of ease as a sign that something is true or familiar (Reber & Schwarz, 1999). So reserve a polished finish for content you are sure of.
  • Disfluency can trigger scrutiny. Text that is harder to read can prompt more careful thinking and help readers notice inconsistencies (Song & Schwarz, 2008). Later replications of disfluency effects have been mixed, and making visuals hard to read on purpose is rarely a good idea.
  • Style sets the tone first. Typeface and colour convey emotional tone before the content is read (Hazlett et al., 2013). A playful style on a serious topic sends mixed signals.
  • Hard-to-pronounce feels risky. Unfamiliar or hard-to-pronounce labels are judged as riskier and less familiar (Song & Schwarz, 2009). Choose plain terms for legends, categories and metrics.
  • Small differences do little. Subtle differences between similar fonts or colours rarely have an effect. Make emphasis deliberate and clear.

Cognitive load

Readers spend mental effort interpreting legends, axes, scales, colour schemes and unfamiliar chart types. If the insight is trivial, that effort feels wasted and erodes trust. If the insight is valuable, the effort feels justified.

  • Extraneous load is effort spent on presentation instead of meaning: clutter, decoration, ambiguous labels, 3D, unnecessarily complex chart types. Remove it.
  • Intrinsic load comes from the complexity of the data itself. You can’t remove it, but you can manage it: add context, split it into panels, use small multiples or progressive disclosure.
  • Germane load is the effort spent understanding and remembering the message. Leave room for it by keeping extraneous load low.

What counts as extraneous depends on the reader. Novices need explanation and simplification, while the same explanation slows experts down (the expertise reversal effect).

What this means for design:

  • Design for the audience, not the analyst. Explanatory charts must guide readers who don’t know the data.
  • Guide the eye. Use titles, annotations, thresholds and hierarchy to state the takeaway, rather than leaving the reader to find it.
  • Structure visuals like a story, matching the order and grouping the reader expects.
  • Prefer conventional forms: bars for comparison, lines for trends. Novelty costs effort.

  • There are no universal rules, only trade-offs. Think of each choice as a point on a spectrum of complexity, familiarity and density (Sibinga & Waldron, 2021), not as a binary.

Sources: Reber & Schwarz (1999); Song & Schwarz (2008, 2009); Hazlett et al. (2013); Sibinga & Waldron (2021); Szafir & Szafir (2016); data.europa.eu, Data communication. See references.