Reading speed changes with the text because the work of reading changes. English-speaking adults average about 260 words per minute on fiction and about 238 on non-fiction (Brysbaert, 2019), and slow further on technical material, for which no reliable universal figure exists. The practical answer is to keep more than one pace in mind and choose between them for each text.
Fiction and non-fiction: the measured gap
The figures used here come from Brysbaert (2019), a meta-analysis of 190 studies with 18,573 participants. For adults reading English silently, it reports:
| Text type | Average | Typical range |
|---|---|---|
| Non-fiction | 238 WPM | about 175–300 WPM |
| Fiction | 260 WPM | about 200–320 WPM |
Two things stand out. The averages differ by 22 WPM, and the ranges overlap almost entirely. Many people read non-fiction faster than other people read fiction. The gap is a difference between averages, not a rule about any one reader.
Why fiction is faster: mostly word length
The explanation Brysbaert (2019) gives is less about subject matter than it sounds. Non-fiction tends to use longer words. A minute of reading covers a roughly similar amount of text either way, but when the words are longer, fewer of them fit into it. Because speed is counted in words, the count comes out lower.
This is a useful reminder about what WPM measures. A “word” in a novel full of dialogue is, on average, a shorter thing than a “word” in a policy report. Part of the fiction advantage is the unit, not the reader.
What the gap means for planning
Over a short text the difference is small. Over a book it adds up. A 90,000-word novel takes about 5 hr 46 min at 260 WPM. A 90,000-word non-fiction book takes about 6 hr 18 min at 238 WPM. That is 32 minutes more for the same word count, before allowing for any passages that need slower reading. Our guide on how long it takes to read a book works through more of these cases.
Technical text: there is no verified universal WPM
You will see specific figures quoted for technical, legal, scientific or academic reading. Treat them with caution. The sources this site relies on do not report an average reading rate for “technical text” as a category, and this guide does not offer one.
There is a good reason such a figure is hard to pin down. “Technical” covers a software manual, a tax regulation, a medical paper and a mathematical proof. The same document can be easy for a specialist and close to unreadable for a newcomer. An average over all of that would describe almost no one.
What the research does support is a qualitative account of why these texts are slower.
Unfamiliar words take longer
Just & Carpenter (1980) tracked readers’ eyes and found that people spend longer on words they encounter rarely. Technical writing is dense with such words: terms of art, abbreviations, names of methods and compounds. Each one costs extra time, and the cost is highest for a reader new to the field. As the terms become familiar, that cost shrinks, which is why specialists read their own literature much faster than outsiders do.
Ideas have to be put together
The same study found longer pauses at points where information must be integrated, such as the ends of sentences, where the reader assembles what the sentence said and connects it to what came before. Technical prose asks for a great deal of this. A sentence may depend on a definition from two pages back, a condition stated in the previous clause and a figure on the facing page.
Eye movements change with difficulty
Rayner (1998), reviewing two decades of eye-movement research, describes what happens as text gets harder: fixations lengthen, the eyes move forward in shorter steps, and they go back to earlier words more often. All three reduce words per minute. None is a fault. They are how a skilled reader responds to demanding material.
Not everything on the page is prose
Technical documents contain equations, tables, code, diagrams and citations. Words-per-minute does not describe studying a chart or working through a formula at all. A page with 200 words and one dense table can take longer than a page with 500 words of plain explanation. For this kind of content, a word-based estimate is a starting point at best. The guide on how text difficulty affects reading speed goes into the features that slow a text down.
Purpose changes speed as much as genre does
The type of text is only half the picture. What you intend to do with it matters at least as much.
Carver (1992) describes reading as a set of distinct processes, or gears, each with its own typical rate. For the college students he studied, his estimates are about 600 WPM for scanning, 450 for skimming, 300 for ordinary reading, 200 for reading to learn and 138 for memorising. These are estimates for his population and should not be read as fixed values for everyone. The point that carries over is that the same person, on the same text, moves at very different speeds depending on the goal.
That is why genre labels mislead. A novel read for a literature exam is study reading. A technical changelog scanned for one version number is scanning. Fiction is usually read in an ordinary-reading gear and textbooks in a learning gear, so some of what looks like a genre effect is really a purpose effect.
The practical consequence: keep more than one pace
A single personal WPM will be wrong for much of what you read. A better approach is to hold two or three figures and pick one per text:
- An easy pace for narrative and familiar, lightly written material.
- A standard pace for general non-fiction such as articles and reports.
- A careful pace for dense, unfamiliar or study material.
If you have not measured your own, the presets in the reading time calculator give reasonable planning values: 300 WPM for brisk reading of easy text, 238 as the research average for non-fiction, and 180 for careful reading. Only 238 is a published average; the other two are round planning values near the ends of the typical range.
The choice changes the estimate considerably. A 4,000-word text takes 13 min 20 sec at 300 WPM, 16 min 48 sec at 238 and 22 min 13 sec at 180. For technical documents with tables, formulas or code, treat even the careful figure as a minimum.
How to test yourself on different kinds of passage
You can find your own paces with the reading speed test. It offers five original passages of roughly 300–450 words. You choose one, click Start Reading to begin the timer, read as you normally would, click Finished Reading and answer four comprehension questions. The result shows raw WPM, your time and your comprehension score, with a comparison to 260 WPM for the narrative passage or 238 WPM for the other passages. These are the adult English fiction and non-fiction averages reported by Brysbaert (2019), not targets for every reader.
To compare text types:
- Take the test on a passage that feels easy and familiar to you, and note WPM and comprehension.
- On another occasion, take it on the passage you expect to find hardest, and note the same two numbers.
- Repeat with the remaining passages over several days, so tiredness does not distort the comparison.
- As a practical rule, keep results with at least three of four questions correct. A fast time with two correct needs another attempt before you use it for planning. The quiz samples a few details, so this cut-off is not a validated measure of understanding.
Suppose your results are 270 WPM on the easiest passage and 205 on the hardest, both with all four quiz questions correct. Treat these as provisional planning values, not established personal paces. A 5,000-word paper would take about 18 min 31 sec at the first and 24 min 23 sec at the second. A demanding paper could take longer than either estimate. Our guide to how long a research paper takes to read builds a fuller time budget, from a quick relevance check to a careful review.
Two limits apply. The passages are short and written for a general reader, so none of them stands in for a specialist document in your field; for that, time yourself on a real sample of the material. And one attempt per passage is a small sample. The guide to building a personal reading speed baseline explains how to turn several attempts into figures you can plan with, and average reading speed gives the wider context for the 238 and 260 figures.