Signal the structure

Claim: the structure of a text is visible before it is read: headings that name their content, lists for parallel items, tables for comparisons, bold for the defined term.

What it asks

  • Headings say what the section holds, not that a section starts ("Results of the timing test", not "Results").
  • Three or more parallel items are a list; two or more items compared on the same attributes are a table.
  • Short blocks, with the key word first: readers scan the first lines and the left edge most.
  • One way to mark each meaning, kept across the whole document.

Why

  • Signalling raises retention, g = 0.53, and transfer, 0.33, over 103 studies (Schneider 2018, measured). The cues it pools are organisational signals, colour and labels; the evidence here names no effect for lists, or for bold that marks a definition.
  • Prior knowledge: whether readers with little of it gain most is contested. One meta-analysis says so (Richter 2016), and Schneider 2018 found no important moderator.
  • Tables suit looking up exact values, and graphs spatial tasks (Vessey 1991, a review of cognitive fit). That supports a table for values compared on the same attributes.
  • Lists and bold keywords are practice, argued from Miller 1956 and anecdote (Moran 2016).
  • Scanning: most web users scan, and a scannable page raised measured usability by 47% on one test site (Nielsen 1997, company research).

Limits

Nielsen Norman Group reports that readers often miss content folded behind a toggle (Wang 2023). That source is a company article below the quality bar: an example, not evidence. Fold only what is optional, or what must stay hidden until the reader asks for it: the answer to a retrieval prompt, for one.