Does Speed Reading Work? What Actually Makes You Faster

No, in the form it is sold. Reading several times faster with comprehension intact has very little evidence behind it, because speed trades against accuracy. But you can genuinely get faster on a particular reading job, and that is a different question with a real answer: know why you are reading, bring knowledge to the page, and skim on purpose.

Almost every page you will find on this query stops at the first sentence. That leaves the student who has 40 pages and two hours exactly where they started, because the verdict is not the advice. This piece gives the verdict quickly and then spends its length on the second question, which is the one worth answering.

Skimming and speed reading are two different claims

Blurring these is the most common mistake in popular coverage, and it makes the evidence look more contradictory than it is. The central reference here is a 2016 review by Keith Rayner, Elizabeth Schotter, Michael Masson, Mary Potter and Rebecca Treiman, So Much to Read, So Little Time, published in Psychological Science in the Public Interest. It is a review of the reading literature written for a general audience rather than a single experiment, which matters for how much weight any one line in it can carry.

The review draws the line by definition. Skimming is where “the goal is to quickly move one’s eyes through the text to find a specific word or piece of information or to get a general idea of the text content”. It reports that “skimming rates can be as much as two to four times faster than those of typical silent reading” and that “Comprehension rates, however, are lower when skimming than when reading”. Skimming states its cost. Speed reading is the claim that you can take the gain and leave the cost, and on that the review concludes there is “little evidence for a unique behavior, such as speed reading, in which speed and comprehension are both high”.

Read that phrasing carefully, because it is doing careful work. Little evidence for something is a weaker statement than evidence against it, and the review keeps that distinction throughout. Some speed-reading claims have been tested and failed. Others have simply never been tested well.

Does speed reading work when it is actually measured?

The tested version is thinner than either side of the argument suggests, and the review says so: there is not a great deal of well-controlled research on trained speed readers. What exists points one way, with caveats worth stating.

Two early reports the review describes each recorded the eye movements of a single speed reader, and found the down-the-middle scanning pattern the courses teach. Free recall in one of them produced confused and sometimes fabricated content. A later lab test of two graduates of a speed-reading programme had each of them read a college-level textbook in under six minutes, at rates of 15,000 words per minute or higher, and both scored poorly on a multiple-choice comprehension test. The reviewers quote that investigator’s summary of the ability actually demonstrated: “a remarkable dexterity in page-turning”. These are case studies of one and two people. They are vivid, and they are not a population.

The most relevant study is bigger and more contested than the tidy version of this story allows. Calef, Pieper and Coffey measured eye movements before and after a speed-reading course in 59 students, 25 of whom took the class and 34 of whom did not, with comprehension scored on ten true-or-false questions. The trained group went from roughly 280 to roughly 400 words per minute. The Rayner review reports their comprehension falling from 81 percent to 74 percent and reads that as speed bought at the expense of comprehension. The original authors, however, described it differently: “Comprehension for the speed reading group showed an insignificant decrease.” Both readings sit on the same numbers, and the arithmetic is ours to offer rather than either party’s: seven percentage points on a ten-item test is a gap of well under one question, which is the sort of difference such a short instrument cannot resolve. So the fair summary is that the course raised reading rate by a modest amount and did not demonstrate that comprehension survived it, which is weaker than saying it demonstrably destroyed comprehension.

One more caution from the review, and it is a good one. In an older comparison of programme graduates against people waiting to take the course, the graduates read far faster and scored about the same on nonfiction. When the same comprehension test was later given to people who had never seen the passage at all, they still scored 57 percent by using general knowledge and guessing. A comprehension test that hands out 57 percent to someone who never opened the passage cannot certify that reading happened. That is worth remembering whenever an app quotes its own comprehension numbers at you.

Why your eyes are not the bottleneck

The mechanical claims underneath speed-reading training are the part where the evidence is genuinely strong, because normal reading has been measured in enormous detail. The review gathers it into one conclusion: “the ability to read is limited by our ability to attend to, identify, and understand words rather than our ability to see them”. Three specific findings sit under that.

The window you can identify words from is small. The review puts the perceptual span in English at about 14 to 15 letter spaces to the right of where you are looking, and the narrower span within which words are actually identified at about seven characters. It also reports that “readers do not access any information from the lines above or below the line currently being read”, which it notes is inconsistent with the claim that anyone takes in a whole page at once.

Backward eye movements help rather than hurt. Skilled readers move back to an earlier word roughly 10 to 15 percent of the time, often to repair a misreading. Schotter, Tran and Rayner tested this in 2014 by masking each word after the eye had passed it, so rereading became impossible. On garden-path sentences, readers who made a backward movement answered a two-choice comprehension question correctly about 75 percent of the time when they could reread, and at around chance when they could not. Courses that train you to stop going back are training away a repair mechanism.

Silencing the inner voice has been tested, and this one is evidence against rather than evidence missing. The review states that “Attempts to eliminate inner speech have been shown to result in impairments in comprehension when texts are reasonably difficult and require readers to make inferences”. Difficult and inference-heavy is a fair description of most of what a student is assigned.

So what actually makes you faster on a specific text?

Here is the part page one leaves out. Reading a set text in less time is achievable, and the levers are unglamorous.

Decide why you are reading before you open it. The review puts this as plainly as it puts anything: “one does not need to read the same way for every reading goal”. Hunting one definition, building a first map of an unfamiliar topic, and preparing to be examined on a chapter are three different jobs, and the speed you should read at is set by which one you are doing. Most students read everything at one pace and then feel slow. Our guide to improving your reading comprehension covers the other half of this, which is what to do once the pace is right.

Bring knowledge to the page. The review attributes to earlier work the observation that effective speed readers tend to be people who already know a great deal about the topic and can therefore skim it successfully while accepting lower comprehension. Its own formulation is the useful one: “To read rapidly, you need to know enough about a topic to fit the new information immediately into what you already know and to make inferences.” The practical inference here is ours rather than the review’s, and worth flagging as such: reading the lecture summary, the syllabus entry, or a short encyclopaedia article before the assigned chapter is a speed technique, even though it looks like extra reading.

Skim deliberately, which is not the same as skimming fast. The review reports a 1982 study in which people skimming at about twice their normal rate lost accuracy equally on important information and on minor details, meaning they were not selecting anything. Naive skimming just reads everything worse. Structured skimming is different: “Effective skimmers scan a text for headings, paragraph structure, or key words to locate potentially relevant information, then read more carefully when such regions are found.” The review also notes that readers who attend more to headings write the most accurate summaries of what they read. The clearest worked example of this is a research article, which almost nobody reads front to back; our walkthrough of how to read a scientific paper sets out the order that actually works.

The strongest single result here is worth stating with its design attached. Duggan and Payne gave readers expository texts with only enough time to read half of each document. In their first experiment, skimming the whole thing beat reading half of it for memory of important ideas, and did not help for minor details or for inferences. The limit matters as much as the result, and the same paper supplies it: Experiment 2 “found no advantage of skimming over reading the first or second half of every paragraph”. Eye-tracking in that same second experiment showed skimmers spending longer on earlier paragraphs and earlier pages, which the authors read as satisficing rather than as a deliberate context-building strategy. So skimming a whole chapter beats reading half a chapter, and it does not beat reading the opening of every paragraph. Under time pressure, covering the whole text lightly while reading openings properly is the shape supported here.

Stop spending the time on rereading. The largest block of wasted minutes in most students’ reading is the second pass over material that already feels familiar, which feels productive and mostly is not. That is a separate mechanism with its own evidence, covered in our piece on the illusion of competence and why rereading feels like learning. Note the distinction against the point above: going back to repair a sentence you misparsed is useful, and rereading a chapter you already understood is the habit worth cutting.

Expect practice to work slowly. The review is candid here: “Practice with reading, and in particular reading to comprehend, does help. But it helps slowly and does not cause drastic increases in speed.” The mechanism it proposes is vocabulary and language experience rather than eye training, since words you meet often are read faster than words you rarely meet. That makes wide reading a genuine long-term speed technique and a useless short-term one. If you want a structured way to work through a chapter that builds the questioning in, the SQ3R reading method is the classic version.

How do you find out whether the fast pass worked?

Every recommendation above is measured against comprehension, and that raises an awkward practical problem: you cannot feel it. The whole trouble with reading quickly is that finishing feels identical whether the material landed or evaporated, and the page is still open in front of you while you judge. A reading speed is only a number worth having if something survived, and the only way to find that out is to close the text and try to produce the content.

That is the narrow job GeniusPal does. It takes a file you already have, a chapter, a set of lecture notes, a PDF, and writes exam-style questions whose answers are on those pages, so you can sit them with the document shut. It does not make you read faster and it is not a reading tool: there is no reader in it, no word-by-word pacing display and no timer, and nothing in it acts on the page during the pass itself. It runs afterwards, as the check on whether the pace you chose was one you could afford.

Money, since it decides what is even available to you here. Two study-set generations land in a new GeniusPal account and nothing ever adds more without payment, so the free figure is a total for the life of the login rather than one that returns each term. It runs both of them as a scored quiz and produces nothing else. The Student plan opens flashcards and written active recall together, billing 14.99 dollars per month and carrying an allowance of one hundred generations within each of those months. Paying for the Genius plan instead means a single charge of 59.99 dollars covering a year, and its unlimited wording describes a ceiling the fair-use policy has to put a number on.

The short version

The strong claim fails, and it fails for a reason you can use rather than an arbitrary one: the constraint sits in language processing, so buying a faster eye does not buy a faster reader. The review’s own closing line on the search for a special mode of reading is that “there is no such magic bullet”.

What replaces it is less exciting and considerably more useful. A student who states the purpose before opening the chapter, reads around the topic first, covers the whole text while reading paragraph openings properly, and then closes the book and tries to produce the content will get through the same material faster and know more of it than one who bought a course. None of that is speed reading. All of it is available this afternoon.