I Was Wrong About Speed Reading: Here are the Facts

Seven years ago, I read some books and articles on speed reading and started practicing some of the methods. I found I was able to increase my reading speed from 450 word per minute to 900 in the drills, so I published an article entitled, Double Your Reading Rate, which has since become one of the most popular on this website.

When I wrote the piece, I based the article purely on my personal experience along with the how-to books I had read. I didn’t have any solid scientific research to back my experiments.

Since that time, I’ve had some lingering doubts about speed reading. In addition to seeing some flickers of research that made me suspicious about speed reading programs, I had mostly stopped using the techniques I originally advocated. My reading diet had switched from lighter self-help, to denser and more academic writing. That meant comprehension, not speed, was the bottleneck I was trying to improve.

Now, nearly a decade later, I decided to do some in-depth research into speed reading to bring you the facts.

Is It Possible to Read 20,000+ Words Per Minute?

Some speed reading claims can be tossed aside immediately. Claims that you can read a book as fast as you can flip through a phone book are completely impossible on anatomical and neurological levels.

First we have anatomical reasons to throw out absurdly high reading rates. In order to read, the eye has to stop at a part of the text, this is called fixation. Next, it must make a quick movement to the next fixation point, this is called a saccade. Finally, after you jump a few points, the brain has to assemble all this information so you can comprehend what you’ve just seen.

Eye-movement expert Keith Raynor, argues that even going beyond 500 words per minute is improbable because the mechanical process of moving your eye, fixing it and processing the visual information can’t go much faster than that.

Speed reading experts claim that they can work around this problem by taking in more visual information in each saccade. Instead of reading a couple words in one fixation, you can process multiple lines at a time.

This is unlikely for two reasons. One, the area of the eye which can correctly resolve details, called the fovea, is quite small—only about an inch in diameter at reading distance. Processing more information per fixation is limited by the fact that our eyes are rather poor lenses. They need to move around in order to get more details. This means that eyes are physically constrained in the amount of information they achieve per fixation.

Second, working memory constraints are at least as important as anatomical ones. The brain can hold around 3-5 “chunks” of information at a time. Parsing multiple lines simultaneously, means that each of these threads of information must remain open until the line is fully read. This just isn’t possible with our limited mental RAM.

What about systems like Spritz? Spritz works by trying to avoid the problem of saccades. If each word appears in the same place on the screen, your eye can stay fixed on that point while words flip through more quickly than you could hunt them down on a page. Indeed, using the application gives a strong impression that you can read very quickly.

Their website claims to have research showing faster reading speeds, but unfortunately I was not able to find any independent, peer-reviewed work substantiating these claims.

Working memory constraints here too, enforce a limit on the upper speed you could use Spritz and still be considered to be “reading” everything. Remember reading was a three step process: fixate, saccade and process. Well that processing step slows down regular reading too. If there are no pauses in the stream of words, there isn’t enough time to process them and they fall out of working memory before they’re comprehended.

Is It Possible to Make Moderate Speed Gains Through Training?

The evidence is clear: anything above 500-600 words per minute is improbable without losing comprehension. Even my own perceived gain of 900 word per minute meant that I was probably losing considerable comprehension. This was masked because the books I was reading had enough redundancy to make following along possible with impaired comprehension.

However, according to Raynor, the average college-educated reader only reads at 200-400 words per minute. If 500-600 words forms an upper bound, that does suggest that doubling your reading rate is possible, albeit as a hard upper limit. Can we still get moderate speed reading gains?

There seems to be some mild evidence here in favor of speed reading. One study of a course had some students quadruple their speed. Another study showed some speed reading experts reading around the 600 word per minute level, roughly twice as fast as a normal reader.

However there’s a trap here. Speed reading may possibly make you a faster reader, but it’s not clear the speed reading techniques are the cause. Second, speed reading trainees tended to read faster, with less comprehension, than non-speed readers. Since measuring comprehension is more difficult than speed, I believe many new speed readers can fall into the trap I did: believing they’re making an unqualified doubling of their reading rate, when in reality, they are doing so at a significant tradeoff of comprehension.

Do Speed Reading Techniques Work?

If the evidence suggests that reading faster may be possible, albeit more modestly, it casts a much harsher light on certain speed reading dogma. The most dangerous is the idea that subvocalization should be avoided to read faster.

Subvocalization is the little inner voice you have when reading that speaks the words aloud. When you started reading you probably spoke out loud with that voice, but you learned to silence it as you got older. If you turn your attention to it, however, you can still hear yourself making the sounds of the words in your head.

Speed reading experts claim that subvocalization is the bottleneck that slows down your reading. If you can learn to just recognize words visually without saying them in your inner voice, you can read much faster.

Here the evidence is clear: subvocalization is necessary to read well. Even expert speed readers do it, they just do it a bit faster than untrained people do. We can check this because that inner voice sends faint communication signals to the vocal cords, as a residue of your internal monolog, and those signals can be measured objectively.

It’s simply not possible to comprehend what you’re reading and avoid using that inner voice. So reading faster means being able to use this inner voice faster, not eliminating it. To further that, expert speed readers who were studied also subvocalized, they just did it faster.

The other main recommendation I made in my speed reading article was using a pointer. This means moving your finger or a pen to underline the text as you read it. This technique is supposed to help you make eye fixations and reduce the random wandering of the eye which can waste time. One study suggests that this apparent function isn’t realized, and that the pointer functions as a pacing device, while actual eye fixations are uncorrelated with pointer or hand movements.

If You Shouldn’t Speed Read, How Should You Read Better and Faster?

In my research for this article, I did find a couple factors that were associated with better reading speed, without sacrificing comprehension. None of these are magic fixes for your reading woes, but a mild treatment that works is better than a fantastic one which doesn’t.

Reading Tip #1: Skim Before You Read

Many speed reading courses are actually teaching skimming techniques, even if they package it as “reading” faster. Skimming is covering the text too fast to read everything fully. Instead, you’re selectively picking up parts of the information.

Skimming, isn’t actually a bad method, provided it’s used wisely. One study found that skimming a text before going on to reading it, improved comprehension in the majority of cases.

Reading Tip #2: Improve Your Fluency to Improve Your Speed

Fluent recognition of words was one of the major slowing points for readers. Subvocalization, that mythical nemesis of speed readers, is slower on unfamiliar words. If you want to speed up reading, learning to recognize words faster seems to improve your reading speed.

Fluency isn’t just an issue for reading in your non-native language. It also matters for technical documents or prose which uses unfamiliar vocabulary. If I’m reading a text that uses jargon like mRNA, or obscure words like synecdoche, I’m going to pause longer. That will slow my reading speed down.

The best way to improve fluency is to read more. If you read more of a certain type of text, you’ll learn those words faster and read better. If you’re a non-native or fluency significantly impacts your reading speed, then even a tool like Anki may be useful for learning hard words.

Reading Tip #3: Know What You Want, Before You Read It

Part of the reason skimming first might appear to help is that it allows you to map out a document. Knowing how an article or book is structured, then, allows you to pay more attention to the things you think are important.

Another tip offered in a lot of speed reading courses, which is likely good advice, is to know what you’re trying to get out of a text before you read it. Thinking about this before you start reading allows you to prime yourself to pay attention when you see words and sentences that are related. Even if you’re reading at a speed which has some comprehension loss, you’ll be more likely to slow down at the right moments.

This isn’t always possible. I read a lot of books unsure about what I want to discover in them. Fiction and reading for pleasure can’t be reduced to a mission objective. However a lot of bland, necessary reading in our lives fits this type. Speeding it up might be worthwhile if it leaves us more time for reading with curiosity.

Reading Tip #4: Deeper Processing Tasks to Improve Retention

Sometimes you don’t want speed at all—you want near full comprehension. When I was in school, I needed to read most textbooks in a way that I could retain nearly every fact and idea I encountered later. It’s not just full comprehension you want, but long-term memory of the information.

Here cognitive science offers some suggestions. A principle of memory is that we remember what we think about. So if you want to remember the ideas of a book, highlighting bolded passages isn’t the best idea. Highlighting causes you to think about bolded words, not what they means.

Some alternatives are taking paraphrased, sparse notes or rewriting factual information you want to remember as questions to self-quiz later.

Conclusion

I was wrong. Subvocalization shouldn’t be avoided. Doubling your reading rate may be possible from a lower range (250 to 500 words per minute, for example), but it’s probably impossible to go beyond 500-600 words and still get full retention. Speed reading may have some redemption as an alternative to skimming text, but even here the benefits come from how speed readers conceptually organize the text, and not on the mechanics of eye movements.

In terms of accuracy, my original article hasn’t aged too well. In my more recent courses, I still teach speed-reading, but I had already shifted mostly to the speed-reading-as-intelligent-skimming paradigm which is a bit more defensible. Still, I’ll be sure to include this research in any new courses I develop.

I apologize to any readers who may have gotten outsized hopes about what speed reading could accomplish. My goal, as always, isn’t to present a fixed dogma of what it takes to learn better, but to research and experiment with new ideas. Unfortunately, sometimes that’s a path that dead-ends or winds back on itself. In any case, I’ll always do my best to share whatever I find with you.

 


Read This Next
#1000
  • Ross

    How can you claim to read that fast without a photographic memory, obviously a story is easy to read as you fill in the blanks but what about a book such as the spectacle of society, without previously reading so your brain has no previous understanding. Or you would be a great philosopher fully understanding everything first try

  • As long as you stay focused only on words, you are going to continue to be a slow reader (<200 wpm). Here's the ONE THING every speed reader eventually has to figure out. You aren't reading words. You're reading IDEAS. And this is the opposite of skimming, because this is actually thinking about the real meaning of what you read.

    If a person with an average reading speed, which is 200 wpm, were to even double that to 400, this would put them in the top 1% of adult readers. And it would also change their life.

    No, you may not learn to read thousands of words per minute, but I don't know why the writer of this blog would try to convince slow readers not to even try to improve this important skills.

  • I don’t do speed reading, I vocalize. As I am Brazilian, I only read English books to practice my second language.
    I noticed that I improved my listening and speaking a lot, just reading. The improvement was bigger than usual when I read The Lord of the Rings, that has a rich vocabulary. I think this is because I sub-vocalize when I am reading. Speed reading also don’t give me pleasure.

  • Graeme

    Great article, thank you!

    I teach the LSAT, for which reading speed is very relevant. I originally taught some speed reading. But after reading that subvocalization can be objectively measured, I realized subvocalizations claims must be unfounded.

    I ended up shifting to the skimming as an auxiliary tool approach you describe. Spefically for the LSAT, I teach people how to skim faster for information *after* they’ve read the passage. This seems to be trainable. (Though in the spirit of this article, I should probably search for literature on this, rather than just trust my anecdotal observations).

    I’ve found history of reading + vocabulary acquisition to be a very important factor. Invariably, slower readers just don’t know as many words. They also tend to have the most difficulty with the complex arguments you find on the LSAT.

    One thing you might find useful for retention is spaced repetition. I’ve used this when learning language via Pimsleur, and I think it has some merit when reading a text. That would be reading once, then re-reading/re-skimming later.

    Once you’ve already read, it doesn’t take nearly as long to read/skim again, and your skimming retention will be higher. Due to spaced repetition, this will greatly increased the retention.

    And type of text definitely matters! According to online tests I read around 500-600 when reading their simple, factual texts. I think I’m using my background in similar texts to fill in some gaps. Whereas when I read a more complex text, I read much slower. On a different calculator, I was 305 for huckleberry finn, and 461 for Alice and Wonderland.

    And while I had full comprehension for both books (I think), I was also reading them a bit uncomfortably fast. If reading for pleasure only I would read much slower. Whereas for the factual information I would probably read quite quickly if I just need information, but slower if it was a complex argument or a topic I cared particularly about. I think there must be some intuitive tradeoff in comprehension vs. speed in very fast readers, and it relates to objective.

  • Tony a

    I’ve been speed reading since my early teens, now 45 🙂 ‘tech news’ items and fiction stories etc are fine, no worries 900 words/min is my avg speed with no subvocalization. (using a ipad app that displays words in the middle of the screen – so you don’t need to move your eyes – however when you blink, you know you have missed a few words lol – app is called QUICKREAD in the app store ) but for real technical text ( like a programming language) I cannot speed read at any real pace. never have. If i need to memorise papers of tech it’s slow going.
    ABOUT ME, my life is a total mess, my IQ 188 and I’ve been stuck in China for last five years ( and I don’t even speak the darn language, and ordering food by pictures is a total gamble! and when the gf cooks, you just smile and eat it. and don’t ask questions, you won’t like the answers lol). I work with a couple of startups ( like the biotech startup. it’s new biotech that kills cancer cells in 45seconds using a biotech plant based agent and lasers, and it actually works ) and I also work with a fuel cell & flow battery startup also here in China.. oh yeah also a VR/AR startup that’s the fun job.

  • Peter McArthur

    Thank you, thank you, thank you.

    Congratulations on being the man who may well have put the speed reading myth to bed. Even better, you have replaced the myth with practical tips on how to read quickly. I especially like your advice to skim before you read; like all the best advice it’s simple, obvious in hindsight, and usually ignored.

  • Magally Vivas

    Thank you. I believe that the most important words in a reading is the content, what you want to express, to convey ideas. Read more always take us to learn the meaning of it (words), to think, to express ideas.

  • I respect you so much for writing an article like this, thank you

  • greentea9

    Thank you for writing this post. For the longest time I thought I was a slow reader especially because of my tendency to subvocalize. I now feel like my subvocalizing tendencies have been validated if that makes any sense lol!

AS SEEN IN