The short version
Speed reading apps do not deliver reading. Nearly all of them flash words one at a time, which does speed up word identification, but it strips out preview and the ability to look back. Comprehension suffers. Normal reading with full understanding caps near 300 words per minute, app or no app.
Everyone else wrote the list. We are writing the correction.
Search for speed reading apps and you will find twenty articles ranking the ten best. They are the same article. The same names, the same screenshots, the same promise of a thousand words per minute on your commute.
Not one of them asks the obvious question. Does the method work?
The answer sits in a paper the industry never cites. In 2016 five researchers published a thirty page review in Psychological Science in the Public Interest. The title asked how we read and whether speed reading can help. The lead author was Keith Rayner, who spent forty years building the science of eye movements in reading, and who co-invented the moving window method, the instrument that measures how much text a reader takes in per fixation. He had no course to sell and no app to defend.
We sell speed reading software, so this page costs us money. Read it anyway. Then read our speed reading guide and decide for yourself.
What a speed reading app actually does
Strip off the branding and nearly all of them do one thing.
RSVP, one word at a time
Rapid serial visual presentation. The app takes your text, breaks it into single words, and flashes them in the same spot on the screen, one after another, at a speed you choose. Two hundred words per minute. Four hundred. A thousand, if you are feeling brave.
Your eyes do not move. The words come to you. The pitch is that eye movement wastes time, so removing eye movement removes the waste, and what is left is pure reading at whatever speed you dial in.
Some apps sweep a highlight along the line and drag your eye with it. Some flash chunks of two or three words. The principle does not change. Control the exposure. Shorten the exposure. Speed follows.
The fact the ten best lists never print
Mary C. Potter, of MIT, invented rapid serial visual presentation. Potter and Levy, 1969. She coined the term. Every speed reading app on your phone is running her idea.
Potter is a co-author of the 2016 paper.
Read that again. The definitive review saying RSVP does not deliver reading is co-signed by the woman who created RSVP. If anybody on earth had a reason to defend the technique, it was her. She put her name to the paper that buries it.
The other family of apps
Not all of them flash words. A second group gives you drills. Wider and wider blocks of text, timed exercises, a chart of your progress climbing week by week.
These are closer to honest, and they are the ones nobody downloads. The marketing is duller and the work is real. They still have a problem, and it is the same problem: the drill trains your eyes, and your eyes were never the slow part.
The best case anyone can make for an app
Give the other side its strongest argument. A case you have not answered is a case you have not beaten.
It runs like this. Your reading speed is held down by habit. Habits break under constraint. An app constrains you, so it breaks the habit, and once the habit is broken the speed stays up after you close the app.
The first two steps are fine. The third is where it dies.
Constraint prevents a behaviour. It does not replace it, and a prevented behaviour walks back in the moment the constraint lifts. That is what the tachistoscope classrooms found sixty years ago and it is what the app finds now. Improvement while the exercise continues. Nothing after.
There is a worse problem hiding in the argument. It assumes the habit was the fault. Look at what the app is preventing. It prevents you going back over a sentence you did not understand. That is not a habit you want to lose. That is the reader working.
To keep a gain you have to build the thing that produces it. Removing the option of doing it wrong is not the same as learning to do it right.
Be fair to the apps. They do something real.
The lazy version of this argument says the apps do nothing at all. That version is wrong, and it is why the industry brushes the criticism off.
RSVP speeds up word identification. It does. Moving your eyes costs time, and the cost is real. A saccade, the flick from one fixation to the next, takes 25 to 30 milliseconds, and you make three or four every second. Take the flicks away and the words can arrive faster than your eyes could ever have fetched them.
An app forces pace, too. You cannot dawdle. You cannot slide backwards, because the word you wanted to re-read has gone.
That is the entire case for the apps. Now watch it turn into the case against them.
Identifying words is not reading
Reading needs two things that RSVP removes.
Preview
While you fixate one word, your brain is already lifting information from the next. The word identification span runs about seven to eight characters to the right of where you are looking. Roughly the word you are on, plus the one after it.
That preview is not a bonus feature. It is a stage of the process. It is why the next word gets identified so fast when your eyes land on it.
RSVP drops one word into a box and deletes everything around it. There is nothing to preview. The app has cut out a stage of reading and charged you a subscription for it.
The ability to go back
Even skilled readers send their eyes backwards on 10 to 15% of their eye movements. The industry calls these reading regressions and sells you drills to kill them.
They are not a bad habit. They are comprehension doing its job. A clause opens, a subordinate clause cuts across it, and the sentence closes twelve words later. Your eye flicks back half a line to hold the two halves together. That flick is the repair.
RSVP has taken the words away. You cannot look back, because there is nothing to look back at. So you carry on, the sentence never closes, and you finish the paragraph having seen every word and understood none of them.
Rayner and his co-authors put it flatly. Under RSVP, comprehension invariably suffers. It gets worse the longer the text runs, because you can hold a single sentence in your head and you cannot hold a page.
The sentence that ends it
The apps optimise the one part that was never the constraint on understanding, and they break the two parts that were.
Your reading rate is set by how fast your brain pulls meaning out of a word. Not by how fast your eyes reach it. A fixation lasts about 250 milliseconds. A saccade lasts 25 to 30. The eye movements are roughly a tenth of your reading time. The rest is thinking, and no app has ever touched the thinking.
The complication, and we will give it to you straight
An honest page shows you the evidence that cuts against it.
In the same year, a group at Sapienza University of Rome published a study on the speed limits of reading. Martelli and colleagues, in PLoS ONE. They agree with almost none of Rayner’s emphasis.
Their finding: normal text reading is capped around 300 words per minute, and the expensive step is executing the eye movements. They put readers on RSVP and got them to 1,000 and 1,200 words per minute. The eyes were the brake, and RSVP took the brake off.
That sounds like a win for the apps. Read the next line of their paper.
The authors state it themselves. The study does not address the full experience of reading, nor does it speak about text comprehension. They measured word identification. They said so.
So we have two teams. Rayner says the brain is the bottleneck. Martelli says the eye movements are the most expensive single step. They disagree about which constraint bites first.
They agree the ceiling is real, and they agree it is low. Around 300 words per minute for reading you understand. Everything above that is a different job with a different name.
The tachistoscope, for the historians
None of this is new, and the apps are not the first to try it.
A tachistoscope projects an image for a controlled length of time. The United States Air Force used them to teach pilots to recognise aircraft from a split second flash. Shrink the picture, shorten the flash, and the pilots got better at it. The method worked, for that job.
Somebody pointed the machine at words. There was a whole industry of it in the 1950s and 1960s. Machines in classrooms. Reading laboratories. A teacher turning up the dial each week. The numbers on the machine went up, and everyone believed.
The numbers on the machine were the problem. Readers got better at the machine.
It is a good story and we like telling it. It is not the argument. The argument is the seven characters and the regressions, and it would stand if the Air Force had never bought a tachistoscope.
What the apps cannot touch
Three things set your reading speed. An app reaches none of them.
The meaning. Long fixations are not a visual fault. When you stall on a word, you stall because the word was hard to identify or the sense was hard to grasp. The eye is waiting for the brain. Flashing the word faster does not make the brain quicker. It just means the word has gone by the time you get there.
Your vocabulary and your knowledge. A reader who knows the subject reads it faster, because the words come out of memory quicker. There is no app for knowing things.
Your judgement about pace. A good reader has gears. They push through a familiar report and slow to a crawl on a contract. An app sets one speed and drives you through everything at it, whether the text is a lease or a lullaby.
Attention belongs on the list too. A reader who knows they will be tested on a page reads it differently, and both speed and comprehension move. Nothing an app does puts anything at stake. You can look away and it will keep flashing words at the wall.
What about the ones that promise a thousand words a minute?
Some will show you a demonstration. You sit there, words firing past, and you feel that you are keeping up.
You are. You are keeping up with word identification. That is the one thing RSVP is good at, and Martelli’s group measured it at 1,200 words per minute. You are not reading. Ask yourself what the passage argued, and the whole thing collapses.
Ask any app maker for the number that would settle this. A comprehension score, at the claimed speed, on text the reader has never seen. It never appears. What appears is words per minute on its own, which is the one number that means nothing on its own.
We go into the ceiling in detail on read 1,000 words per minute. The short answer is that the number is real and the reading is not.
What actually works
Now the part that will annoy the industry more than any of the above.
Speed reading courses do sometimes work. Students finish them reading faster than they started. That is not a fraud and we are not going to call it one.
The gain is real. The explanation is not the one on the sales page. A course runs for 8 to 12 weeks and it makes the student read a great deal of text. That is the active ingredient. The drills were along for the ride.
Reading a lot is the shortcut. There is no other shortcut.
The second thing that works is learning to skim on purpose. Rayner called skimming a reasonable way to cope with the overwhelming amount of text we have to read, and he was right. Skimming is a skill. It has a cost, you accept the cost, and you choose when to pay it. That is a decision, not a technique you buy. Our page on skimming vs speed reading draws the line.
Reading is not one speed and it never was. Ronald Carver measured what people do at different jobs. Memorising a text runs near 140 words per minute. Learning it runs near 200. Reading it with full understanding runs near 300. Skimming runs near 450 and scanning for a single target word runs near 600, and neither of those last two is reading with comprehension. They are different jobs done at different speeds.
An app that drives you at one fixed rate has removed the only real skill in the list, which is knowing which gear you are in.
Start by measuring. Take the speed reading test and find out what you actually read at, because almost everybody guesses high. The average adult sits near 250 words per minute. The fastest competent adults reach about 550, and by then they are trading something away.
Then work through real speed reading exercises, and give it the weeks it needs. Our page on how long it takes to learn speed reading has the honest timeline.
Use an app if you enjoy it. It will pace you and it will stop you sliding backwards, and for a few weeks that will feel like progress. Just do not confuse a crutch with a leg, and do not be surprised, when you put it down, to find yourself standing where you started.
Sources
Rayner, K., Schotter, E. R., Masson, M. E. J., Potter, M. C., & Treiman, R. (2016). So Much to Read, So Little Time: How Do We Read, and Can Speed Reading Help? Psychological Science in the Public Interest, 17(1), 4–34.
Martelli, M., et al. (2016). Perceptual and Cognitive Factors Imposing Speed Limits on Reading Rate: A Study with Rapid Serial Visual Presentation. PLoS ONE, 11(4): e0153786.
Potter, M. C., & Levy, E. I. (1969). Recognition memory for a rapid sequence of pictures. Journal of Experimental Psychology, 81(1), 10–15.
