Does Speed Reading Work? What the Research Says
Courses and apps promise that you can read several times faster and still understand everything. We read the main scientific review on the subject and the studies behind it to see which of those promises hold up.
What speed reading promises
The idea has been popular since at least 1959, when Evelyn Wood launched her Reading Dynamics program. Smartphone apps later joined the courses. In 2016 five cognitive psychologists (Keith Rayner, Elizabeth Schotter, Michael Masson, Mary Potter and Rebecca Treiman) published a large review in Psychological Science in the Public Interest. It covers what science knows about reading and what that means for speed reading courses and apps. This article is built on that review.
The authors examine three main promises: that you can learn to take in more text at a glance, that you can suppress your inner speech, and that showing words one at a time on a screen can do away with eye movements.
Speed versus comprehension
The review’s main conclusion is that reading, like any other behavior, involves a trade-off between speed and accuracy. The faster words pass before your eyes, the worse you understand and remember the text. The authors consider it unlikely that you could double or triple your speed (say, from 250 to 500–750 words per minute) and understand a text as well as you do at a normal pace.
For reference, the review puts most readers’ normal pace at 200–400 words per minute. A 2019 meta-analysis of 190 studies by Marc Brysbaert found that adults average 238 words per minute on non-fiction and 260 on fiction (for English). There’s more on these numbers in “Average Reading Speed”.
Why your eyes aren’t the limit
Courses often teach you to read with your “peripheral vision”, taking in whole lines or even pages. But we see sharply only with a small central area of the retina, and its field of view is about the width of your thumb at arm’s length. That’s why our eyes jump from word to word.
More importantly, what limits speed is how quickly we recognize and understand words, and vision has little to do with it. In a study that compared many cognitive skills, the ability to identify words was the one most strongly linked to reading speed. And exercises with words flashing in the periphery did not help people read faster with good comprehension.
Another common tip is never to look back. Yet 10–15% of a skilled reader’s eye movements go backward, and many of them repair a misunderstanding. In an experiment by Schotter, Tran and Rayner (2014), the letters of each word were replaced with “x” as soon as the reader’s gaze moved on, so rereading was impossible. In normal reading, when people went back to a word, they answered a question about the sentence’s meaning correctly 75% of the time. When there was nothing to go back to, they got it right about half the time, which is chance level.
Your inner voice isn’t the problem
Many speed reading books tell you to get rid of inner speech (subvocalization), as if it were a habit left over from learning to read aloud that slows you down. Research shows the opposite: the sound of words plays a part in recognizing and understanding them, even in silent reading. For example, in quick checks of meaning, people more often wrongly accepted a sentence in which a word had been swapped for one that sounds the same (like “hair” and “hare”) than a sentence with a word that looks similar but sounds different.
Researchers have tried to suppress inner speech in two ways. When people used a feedback signal to learn to quiet the activity of their speech muscles, comprehension of easy texts didn’t suffer, but comprehension of difficult ones got worse. When people had to repeat unrelated words aloud while reading, comprehension also dropped, especially where they had to connect ideas and draw a conclusion. The sound of words was represented even in people reading sentences at 720 words per minute, shown one word at a time. The authors conclude that with a difficult text, inner speech helps rather than hinders.
What tests of course graduates found
There are direct tests too. In one study (Calef and colleagues, 1999), adults read about 280 words per minute before a speed reading course and about 400 after it, but the share of correct answers to questions about the text fell from 81% to 74%.
The most detailed study (Just, Masson and Carpenter, 1980) compared course graduates reading at 600–700 words per minute, ordinary readers at about 250 words per minute, and people asked to skim the text at the same speed as the graduates. The graduates grasped the general meaning better than the skimmers but worse than the ordinary readers, and on questions about details they were no different from the skimmers. On technical texts, where you can fill in little from general knowledge, they lost their advantage in general meaning too.
The review concludes that there is no evidence training lets you speed up dramatically and keep excellent comprehension. In essence, the courses teach skimming. It works best for people who already know a lot about the subject and are willing to give up some comprehension.
Apps that show one word at a time
Apps like Spritz show text one word at a time in a single spot on the screen, at a set speed. The method is called rapid serial visual presentation (RSVP), and the promise is to save the time spent on eye movements. Media reports citing Spritz claimed that 80% of reading time goes to eye movements. The review estimates that the eye jumps themselves take about 10% of the time, and text processing doesn’t stop during them.
The cost is that with word-by-word display you can’t go back, and you can’t catch the next word out of the corner of your eye, while in normal reading the brain starts processing that word before you look at it. At a normal pace, comprehension with RSVP was comparable to ordinary reading in a number of studies. But in the work the authors reviewed, speeding up the display beyond a normal pace always reduced comprehension and memory, and without pauses between sentences comprehension suffered even at normal speed.
Benedetto and colleagues (2015) tested Spritz directly. People read an excerpt from Orwell’s novel 1984 either normally or one word at a time, with the option to pause the display. Once pauses were counted, the speeds came out almost the same: 200 and 209 words per minute. Word-by-word display tired the eyes more and worsened literal comprehension of the text. On questions that required drawing a conclusion, participants did equally well in both versions.
Skimming: a fair trade
Skimming means moving your eyes quickly through a text to find something or get the general idea. It can be 2–4 times faster than normal reading, but comprehension is lower. That’s the price, and it isn’t a flaw: if you’re looking for a specific fact or going through material on a familiar subject, the trade is worth it. If you need to understand a text fully, you’ll have to read at a normal pace and sometimes reread.
You can skim well or badly. In one study, people who skimmed at twice their normal speed got equally worse at recognizing both the important ideas and the small details, which means they weren’t picking out what mattered. What helps is looking at headings, paragraph structure and keywords, then reading carefully where you find what you need. Readers who paid more attention to headings wrote the most accurate summaries of the text.
What actually makes you read faster
Reading practice, especially reading for understanding, does help, but slowly and without sudden jumps. What works is experience with language, not eye training:
- Vocabulary. Common words are recognized faster than rare ones, and the more often you meet a word, the faster you read it. Reading more, and reading different things, does more good than rereading the same passage.
- Knowledge of the subject. The more you know about a topic, the easier it is to anticipate what comes next and fit new information into what you already know. That’s why it’s worth reading about many different things.
- Experience with written language. Books contain words and phrasing that rarely come up in conversation, and the only way to get used to them is to read.
The authors of the review remind us that the usual 200–400 words per minute is faster than we normally take in information by listening, and that for most tasks it’s a good speed.
How to tell if you’re reading faster
If you want to read faster, the honest path is more modest than the ads: read more, read different things and go at your own pace. To see whether anything is changing, measure your speed now and then, on similar books and in similar conditions, and make sure you understand what you read. Compare yourself with yourself, not with champions.
Before you choose any training, find your starting point: the online reading speed test takes a couple of minutes.
Test any method on yourself, with numbers instead of promises
Measure your speed before and after a course: read a paper book with the SpeedReading timer and photograph the page. The app counts the words and shows on a chart whether anything has really changed.
Sources
- Rayner K., Schotter E. R., Masson M. E. J., Potter M. C., Treiman R. So Much to Read, So Little Time: How Do We Read, and Can Speed Reading Help? Psychological Science in the Public Interest, 2016, vol. 17, no. 1, pp. 4–34. doi:10.1177/1529100615623267, PubMed. The studies by Calef and colleagues (1999) and Just, Masson and Carpenter (1980), and the experiments on suppressing inner speech and on skimming, are cited from this review.
- Schotter E. R., Tran R., Rayner K. Don’t Believe What You Read (Only Once): Comprehension Is Supported by Regressions During Reading. Psychological Science, 2014, vol. 25, no. 6, pp. 1218–1226. doi:10.1177/0956797614531148, PubMed.
- Benedetto S., Carbone A., Pedrotti M., Le Fevre K., Bey L. A. Y., Baccino T. Rapid serial visual presentation in reading: The case of Spritz. Computers in Human Behavior, 2015, vol. 45, pp. 352–358. doi:10.1016/j.chb.2014.12.043.
- Brysbaert M. How many words do we read per minute? A review and meta-analysis of reading rate. Journal of Memory and Language, 2019, vol. 109, 104047. doi:10.1016/j.jml.2019.104047.