Monospace or Proportional for Writing Notes? What the Reading Research Actually Found

The most-quoted number against monospace is real: readers were 5% faster with Times than with Courier. It comes from a 1996 print study on an acuity chart, and the same study found Courier won on four other measures. On screen, eye-tracking found no reading-time difference at all.

The folk rule is older than the evidence for it. Monospace is for code, proportional is for prose — you absorbed it from a hundred editors that ship one font for the terminal and another for the document. It sounds like a settled finding. It is mostly a repeated one.

There is real research underneath, and it is more interesting than the rule it supposedly supports: four peer-reviewed papers spanning 1996 to 2022. Read together, they do not say monospace is worse. They say the gap is small, it flips sign under conditions that are easy to meet, and the one modern study that watched people's eyes move across a screen found no reading-time difference.

So the question "monospace or proportional for writing notes" has an honest answer, and it is not a winner. Here is what each study measured.

Is monospace harder to read?

Sometimes, by a little, under conditions worth naming. The strongest version of the received position comes from typographer Matthew Butterick, who writes flatly: "Compared to proportional fonts, monospaced fonts are harder to read" ¹. He is not wrong that a measured penalty exists. He states as a general rule something the data scopes far more narrowly.

Butterick's case deserves quoting at full strength. His verdict on body text: "In standard body text, there are no good reasons to use monospaced fonts. So don't. Use proportional fonts" ¹. He files monospace among typewriter habits to unlearn, listing "Using monospaced fonts rather than proportional fonts" under things you should not do ².

His explanation is historical, and accurate: "Monospaced fonts were invented to suit the mechanical requirements of typewriters. They were not invented to win beauty contests" ¹. True — and also a claim about origin, not legibility. The two get quietly welded together in the folk version.

The measured penalty behind all of this traces to one paper.

The 5% comes from a 1996 print study

In 1996, J. Stephen Mansfield, Gordon Legge and Mark Bane published a study in Investigative Ophthalmology & Visual Science that reported: maximum reading speeds "were 5% faster with Times than with Courier (P < 0.001)" ³. That is the number the folk rule rests on. Its conditions matter more than its size.

The study tested 50 normal subjects and 42 subjects with low vision on the MNREAD Acuity Chart — printed cards, in 1996, built to measure reading acuity rather than simulate a work session ⁴. Times was the proportionally spaced face; Courier the fixed-width one.

Not Courier Bold, and not a screen.

None of that makes the finding wrong. It makes it specific. A 5% difference measured on printed acuity cards thirty years ago is a real result about printed acuity cards thirty years ago. Turning it into a law about the font in your note editor is a longer leap than the sentence admits.

The same study found four things that cut the other way

Four of Mansfield, Legge and Bane's findings favour Courier — the monospace. It won on reading acuity for both groups, and on critical print size. For subjects with low vision it was faster, not slower. And below a certain size the gap did not close: it inverted, dramatically ⁵.

On acuity, scores "obtained with Courier were better than those obtained with Times for both normal (mean difference, 0.05 logMAR, P < 0.001) and subjects with low vision (0.09 logMAR, P < 0.001)" ⁵. The monospace let readers resolve finer text.

The speed result flips sign with the population. For subjects with low vision, "maximum reading speeds were 10% slower with Times than with Courier (P < 0.05)" ⁶ — the 5% penalty becomes a 10% advantage, in the same study.

Then size. Below the critical print size, reading speeds "were substantially slower (by as much as 50%) for Times than for Courier" ⁷.

When text gets small, the proportional face collapses.

The authors' own conclusion is the opposite of the folk summary: "There are small, but significant, advantages of Courier over Times in reading acuity, critical print size, and reading speed for subjects with low vision" ⁸. The folklore quotes one number from a five-number study — the only one pointing its way.

What happened when researchers watched people read on a screen

Luz Rello and Ricardo Baeza-Yates ran the closest thing to a modern test of the rule in 2016: eye-tracking, on screen, in ACM Transactions on Accessible Computing ⁹. On the measure the rule is about — how long reading takes — they found nothing: "We did not find a significant effect of [±Monospace] on Reading Time" (p = 0.065) ¹⁰.

The method was 97 subjects, 48 of them with dyslexia, reading 12 texts in 12 fonts ⁹. A null is not proof of equivalence, and p = 0.065 is a near miss rather than a flat nothing.

But the difference the folk rule promises did not show up.

What did show up was smaller and stranger. Fixation duration, how long the eye rests in one spot, was shorter on the monospace for both groups: 0.22 seconds against 0.26 for readers with dyslexia (p < 0.001) ¹¹, and 0.19 against 0.20 for readers without (p = 0.002) ¹².

Sit with that second number. It is a hundredth of a second, and it holds for readers without dyslexia — not the dyslexia-only effect it usually gets summarized as. The paper says so directly: "What is good for people with dyslexia regarding font types is also good for people without dyslexia" ¹³.

The twist that keeps this honest

Those same readers preferred proportional anyway. Without dyslexia, participants "significantly preferred [−Monospace] fonts (CMU, Garamond, and Times) (x̄ = 3.59, s = 0.98) over the [+Monospace] font (Courier) (x̄ = 2.85, s = 1.30)" (p = 0.003) ¹⁴. They fixated faster on what they liked less. Preference and performance are different measurements.

The paper states the tension in its conclusions without flinching: "However, participants without dyslexia preferred proportional fonts, not monospaced" ¹⁵. Both are true at once, and any version of this post reporting only the flattering half is lying by selection.

The authors then police their own result, which is the most credible thing in the literature. Table XI marks the monospace recommendation with an asterisk and explains why: "Values marked with '*' should be considered as weak recommendations since significant effects were only found for one measure (Fixation Duration) out of the three measures used for objective readability" ¹⁶.

One measure out of three.

The researchers who found the effect call it weak — a lower weight than either side of the internet argument gives it.

The paper is also, quietly, in tension with itself. Its conclusions say monospaced fonts led to better readability for readers with dyslexia via "shorter reading times and fixations" ¹⁵ — but its own test in §5.4 found the reading-time effect not significant ¹⁰, and Table XI marks the recommendation weak on that basis ¹⁶. The test and the table govern. Nobody should write "monospace was read faster," including the people who wrote the paper.

Why the two families behave differently at all

The mechanism is not glyph shape. It is spacing ratio. Timothy Slattery, Mark Yates and Bernhard Angele, publishing in Journal of Experimental Psychology: Applied in 2016, found that "inter-word space needs to be at least 3.5 times the inter-letter spacing for reading to proceed efficiently" ¹⁷. Word boundaries are what the eye is actually hunting for.

That threshold explains the family difference. The same paper found fixed-width fonts "are composed of a larger percentage of 'default' […] inter-letter space than the proportional width fonts, t(9) = 3.397, p < .05" ¹⁸. Monospace spreads letters apart, pushing the word-space ratio toward the floor — the word stops being one clump.

Which is why the most interesting answer came from people who refused the binary. On iA Writer Quattro: "Quattro shares similarities with a proportional typeface. At the same time, it retains a lot of the technical virtues of the classic typewriter fonts using wider gaps between the words…" ¹⁹.

Their design note reads like the Slattery finding restated as craft: "We kept large word spacing and monospaced punctuation" ²⁰. An earlier face, iA Writer Duo, was "a two-spaced typeface that gives a bit more room to W's and M's, based on IBM Plex" ²¹. Widen the word gaps, keep the rhythm, and the ratio problem goes away.

The two families were never the only options.

How to choose your note font

Five moves, in order of how much they matter. None of them is "pick the correct family," because the research does not support one. The evidence supports something less satisfying and more useful: get the size right, notice what you actually prefer, and treat the family as the last and smallest decision rather than the first and largest one.

1. Fix the size before the family. Mansfield's penalty inverts below critical print size, where proportional ran up to 50% slower ⁷. If your text is small, size is the bigger lever.
2. Match the font to the task, not the identity. The measured gaps are small enough that what you are reading matters more than which family it is set in.
3. Treat your preference as data, not a verdict. Rello's readers preferred proportional while fixating faster on monospace ¹⁴. Liking a font is a real reason to use it; it is not a legibility claim.
4. If you try monospace, look at the word gaps. The 3.5× word-to-letter ratio is the mechanism ¹⁷. Faces built with wide word spacing answer the objection.
5. Change it in a month and see. The decision is reversible — the strongest argument against agonizing over it now.

The caveats

This post's spine is a null and a set of small effects — the honest scope is narrow. One on-screen study found no reading-time difference ¹⁰, which is not proof the difference is absent. Effect sizes are hundredths of a second ¹². The 1996 findings are print, on an acuity chart, and do not transfer to a 2026 screen ⁴.

The strongest claim this evidence supports is that the folklore is overstated, not that monospace wins. Butterick's position remains defensible for the case he is actually arguing about, which is standard body text in typeset documents ¹, not the font you write notes in.

The low-vision and dyslexia findings ^{6 11} are reported as research populations, not recommendations. If you have a vision condition or a reading difficulty, an optometrist or specialist knows things this literature does not.

This is not medical advice.

And the folk rule has no basis in the specification it appeals to. The CSS Fonts Module Level 3, a W3C Recommendation dated 20 September 2018, defines the family by geometry alone: "The sole criterion of a monospace font is that all glyphs have the same fixed width. This is often used to render samples of computer code" ²². "Often used" is an observation about habit. "Monospace is for code" is a description that got promoted to a rule.

Frequently asked questions

The short version of every answer below: the measured differences between monospace and proportional are small, they depend on size and reader, and they point in different directions depending on what you measure. No study here crowns a family. The most-cited number favours proportional; the study it comes from favours monospace four other ways.

Are monospaced fonts easier to read?

Not measurably, and not harder either, on the best on-screen evidence. Rello and Baeza-Yates's 2016 eye-tracking study found no significant monospace effect on reading time (p = 0.065) ¹⁰. Fixations were slightly shorter on monospace for both reader groups, by about a hundredth of a second ¹² — an effect the authors themselves call weak ¹⁶.

Is monospace harder to read?

Slightly, in one 1996 print measurement, under conditions worth naming. Reading speeds ran 5% faster with Times than Courier on the MNREAD acuity chart ³. In that same study Courier read better at small sizes and for subjects with low vision ^{5 6}. On screen, the reading-time difference did not appear at all ¹⁰.

Should I write in monospace?

If you like it, yes. The measured differences are small, mixed in direction, and dependent on size and reader ^{5 10}. Readers without dyslexia in the 2016 study preferred proportional while fixating faster on monospace ¹⁴ — preference and performance disagreed inside one experiment. Pick the one you want to look at, and keep it changeable.

Why do writers use monospace fonts?

Rhythm, restraint, and history — plus a spacing argument that turns out to be real. iA designed Quattro to keep "the technical virtues of the classic typewriter fonts using wider gaps between the words" ¹⁹. Wide word spacing tracks the 3.5× word-to-letter ratio reading research identifies as the efficiency threshold ¹⁷.

Does monospace help readers without dyslexia too?

Directionally, by a tiny margin. The fixation-duration effect held for readers without dyslexia (p = 0.002), not only for readers with it ¹², and the paper states that what helps one group helps the other ¹³. The margin is roughly 0.01 seconds, and its authors label the recommendation weak ¹⁶. This is not medical advice.

What's the best font for taking notes?

There is no single answer, and the largest study says so. Wallace and colleagues found that reading speeds "increased by 35% when comparing fastest and slowest fonts without affecting reading comprehension" ²³ — then concluded: "High WPM variability across fonts suggests that one font does not fit all" ²⁴. The best font is the one you measure on yourself.

Monospaced fonts: what are the pros and cons for the reader?

Pros: better measured acuity, better performance at small sizes, and a large advantage below critical print size in the 1996 print data ^{5 7}. Cons: a 5% speed cost at comfortable print sizes ³, and most readers without dyslexia simply prefer proportional ¹⁴. On screen, reading time showed no difference ¹⁰.

The decision you never have to encode

The largest study here ends on individuation, not a winner. Wallace, Bylinskii, Dobres and colleagues, in ACM Transactions on Computer-Human Interaction in 2022, measured a 35% reading-speed spread between fastest and slowest fonts with no comprehension cost ²³, and concluded: "High WPM variability across fonts suggests that one font does not fit all" ²⁴.

The variance is between readers, not between families. Which is another way of saying the answer was never going to arrive from a study — it arrives from you, reading. That variance runs through the wider question of how a typeface shapes thinking, covered in The Font on Your Screen Is Shaping How You Think.

That advice is only useful if you can act on it, and you can, because the decision is not stored anywhere. A note written in plain Markdown carries no typeface. The width lives in the renderer's font choice, never in the bytes — the spec is explicit that a monospace font is defined by fixed width, and nothing else ²².

The font is a view, not the file, exactly like the line length you read at and the theme you read in.

So the folk rule loses twice. It is not in the research, and it is not in the file. Pick the font you want to look at this month, and change your mind next month, because you never encoded the decision.

Your notes are plain text rendered by a reader you control — mnmnote.com keeps the words in open Markdown, so the typeface stays a choice you can take back.