GoWin Tools
Tools
โ† Social Media Character Counter

Social Media Character Counter ยท 6 min read

The Science of Text Length: What Research Says About Readability

Readability research stretches back to the 1920s. Here is what decades of studies on sentence length, paragraph length, line length, and cognitive load tell us about how to write text that people can actually process.

Why Length Matters

Text length โ€” at the sentence, paragraph, and line level โ€” directly affects how easily readers process written content. This is not a stylistic preference; it reflects cognitive constraints in working memory, eye movement patterns, and the way the visual system handles text. Getting these lengths right improves comprehension and reduces the effort required to read.

Sentence Length: The Flesch-Kincaid Foundation

The most influential readability research came from Rudolf Flesch, an Austrian-born linguist whose 1948 paper introduced the Flesch Reading Ease Score. Flesch found that two variables predicted readability more than any other: average sentence length and average number of syllables per word. Longer sentences and more complex words correlated with lower comprehension in populations he studied.

The Flesch-Kincaid Grade Level formula, developed by Kincaid et al. (1975) for the US Navy, refined this into the formula used in Microsoft Word's readability statistics today:

Grade Level = 0.39 ร— (words/sentences) + 11.8 ร— (syllables/words) โˆ’ 15.59

Research-derived guidance on sentence length:

Average Sentence LengthReadability
8 wordsVery easy (tabloid newspapers)
14 wordsEasy (popular fiction)
17 wordsStandard (news reporting)
21 wordsDifficult (quality press)
25+ wordsVery difficult (academic prose)

The implication: for general-audience online content, sentence lengths averaging 14โ€“17 words are optimal. This does not mean every sentence should be exactly 15 words โ€” varied sentence length keeps writing engaging. But the average across a piece should stay in this range for non-specialist audiences.

Paragraph Length: Working Memory and Chunking

Working memory โ€” the cognitive system that holds information in mind during processing โ€” has a limited capacity, famously estimated by Miller (1956) as "7 plus or minus 2" items. A paragraph functions as a working memory chunk: the reader holds the paragraph's central idea in mind while processing each sentence, then releases it and moves to the next.

Research on paragraph length suggests:

  • Print (books): paragraphs of 150โ€“250 words are well-tolerated
  • Web/screen: paragraphs of 40โ€“70 words are optimal; longer paragraphs are more likely to be skipped
  • Mobile screens: 30โ€“50 word paragraphs; long paragraphs fill entire screens and feel intimidating

The reason web and mobile paragraphs should be shorter than print paragraphs is environmental, not cognitive. On a screen, a long unbroken block of text appears visually heavy and is harder to scan. Readers who skim content (which Nielsen's eye-tracking studies confirm most web users do) skip long paragraphs and jump to shorter, more scannable sections.

Line Length: The Optimal CPL

The length of a single line of text โ€” measured in characters per line (CPL) or the CSS unit ch โ€” affects reading comfort and speed significantly. The research consensus for optimal line length:

  • Print (books, newspapers): 45โ€“75 characters per line
  • Web (single column): 50โ€“75 characters per line
  • Web (maximum): 85โ€“90 characters before comprehension drops

The reason for the upper limit: when lines are too long, readers struggle to find the beginning of the next line after a saccade. The eye has to travel further to reset and is more likely to land on the wrong line. This causes more regressions and comprehension errors.

The reason for the lower limit: very short lines (under 30 CPL) force too many line returns. Each line return interrupts the reading rhythm and requires the eye to find the new line start. Very narrow columns (common in newspaper multi-column formats) work because the narrow columns are surrounded by adjacent columns that provide visual context for line-return tracking.

The CSS property max-width: 65ch โ€” constraining a text container to 65 characters wide โ€” is a common implementation of the optimal line length for web typography.

Headline and Title Length

Research on headline readability comes largely from journalism and digital marketing, where headline performance can be measured empirically through click-through rates:

  • 6 words: the "sweet spot" identified in multiple eye-tracking studies โ€” readers often read only the first 3 and last 3 words of a headline
  • 8โ€“12 words: Google search result headline length, constrained by search result display
  • 60โ€“70 characters: the maximum before Google truncates a title tag in search results
  • Social sharing: Twitter/X shares display 100โ€“280 characters; Facebook up to 100 characters prominently

The Inverted Pyramid: What Goes First

Beyond the mechanics of length, research on web reading โ€” most extensively by Jakob Nielsen's NN/g group โ€” confirms that web readers do not read top to bottom. They scan, jump to subheadings, and read the most important information first (if it is placed first). This is the journalistic "inverted pyramid" principle: lead with the conclusion, follow with supporting detail.

The character-count implications: the first 100โ€“150 characters of any block of text (especially on mobile, where text is truncated in previews) must contain the essential message. Writing that builds to its point over several sentences will have its payoff skipped by most web readers.

Count characters for any platform โ†’

References

  1. Flesch, R. (1948). A new readability yardstick. Journal of Applied Psychology, 32(3), 221โ€“233.
  2. Kincaid, J.P., et al. (1975). Derivation of new readability formulas for Navy enlisted personnel. Naval Technical Training Command.
  3. DuBay, W.H. (2004). The principles of readability. Impact Information.
  4. Nielsen, J., & Pernice, K. (2010). Eyetracking Web Usability. New Riders.
  5. Baddeley, A.D. (1986). Working Memory. Oxford University Press.