The Testing Effect: Why Trying to Spell a Word Beats Looking at It

Most people study a spelling list the way they would study a painting. They look at it. They look again, a little harder, trace the tricky part with a finger, and after the third or fourth pass the word seems to settle into place, so familiar now that misspelling it feels impossible. Two days later, halfway through an email, it arrives on the screen with the wrong vowel in the middle.

That is not a failure of effort or intelligence. It is a failure of method, and a very common one, because the method that feels most like learning is one of the least effective ways to learn anything. The alternative is almost insultingly simple: hide the word, try to spell it from memory, then check. Psychologists call the benefit of that small act the testing effect, and it is among the most reliable findings in the science of learning.

What follows is what the research actually showed, and why the effort of recall is the whole point rather than an unfortunate side effect. It also covers why a wrong first attempt can help, and how to turn any list into a few minutes of practice that holds. That goes for a child's Friday test and for your own private collection of workplace nemeses alike.

What the testing effect is

The idea fits in one sentence. Pulling a piece of information out of memory strengthens that memory more than taking the same information in again.

Researchers also call this retrieval practice, and the word "test" misleads people, because it has nothing to do with grades. A test, here, is any moment when you have to produce the answer with nothing to copy: a friend reading a word aloud while you write it, a flashcard turned face down, a word spoken by a recording that you then type. The score is irrelevant. What matters is the act of fetching.

The difference between looking and fetching is the difference between recognizing a face in a crowd and drawing that face from memory. Most of us can do the first for hundreds of people. Almost none of us can do the second for anyone.

Spelling, it turns out, is a drawing task.

What Roediger and Karpicke found

The experiment that made the testing effect famous was published in 2006 by Henry Roediger and Jeffrey Karpicke at Washington University in St. Louis. College students read short passages on science topics. Some then read the passage a second time. Others put it away and wrote down everything they could remember, with no feedback on what they got right.

Tested five minutes later, the rereaders came out slightly ahead. That is the result everyone expects, and it is the result that fools us, because it is what we see in the moment: the page is fresh, the words are warm, and more study seems to produce more knowledge. Tested two days later, and again a week later, the result reversed. The students who had practiced recall remembered substantially more.

In a second experiment, some students read a passage four times, and others read it once and then tried to recall it three times. A week later, the recall group remembered about 61 percent of the ideas and the four-times readers about 40 percent. The rereaders had also predicted they would remember more. They were confident, and they were wrong.

That last detail is the most important thing in the study. Rereading produces a feeling of fluency, a sense that the material is easy and known, and that feeling is a poor guide to what you will be able to produce later. The same trap operates with spelling. After four readings, accommodate looks utterly familiar, but familiarity is not the ability to write it on a blank screen.

Later work kept pointing the same way. In 2011, Karpicke and Janell Blunt compared retrieval practice with concept mapping, an effortful technique in which students draw diagrams linking ideas, and retrieval won a week later, even on questions that asked students to draw concept maps. In 2013, a review led by John Dunlosky rated ten popular study techniques, and practice testing was one of only two to earn the highest rating. The other was spreading practice out over time. Rereading and highlighting, the habits nearly every student relies on, were rated low.

Does it work for spelling?

Most of the classic studies used prose passages, word pairs or facts, so it is fair to ask whether spelling behaves the same way. The direct evidence says it does. In a 2016 paper with the memorable title "Beyond the Rainbow," Angela Jones and colleagues compared two ways elementary school children practiced spelling words. One was rainbow writing, a classroom favorite in which the child traces a word again and again in different colors. The other was retrieval practice: study the word, cover it, try to write it, check. Retrieval produced better spelling on the later test.

There are good reasons to expect this. Spelling is a production skill. When you write, nobody shows you the word first; you hear it in your head and must supply every letter, in order, including the ones your ear cannot help with. The reading researcher Linnea Ehri has spent decades showing how skilled spellers bind a word's sounds to its letters and its meaning, a process she calls orthographic mapping. Each genuine attempt to produce a word exercises exactly that binding. You notice which part you are unsure of, whether the m doubles, which vowel hides in the unstressed middle, and the check that follows settles it.

Reading exercises something else. It asks only that you recognize a word already in front of you, and recognition works on partial information: the first and last letters, the overall shape, the sentence around it. That is why so many avid readers are indifferent spellers. They have seen definitely thousands of times and built it from nothing almost never.

Why the struggle is the point

Retrieval practice has a cost, and the cost is emotional. Trying to spell a word you half know is uncomfortable. You hover over the middle, unsure whether it ends in able or ible, and every instinct says to peek, get it right, feel competent again.

The psychologist Robert Bjork has a name for this: a desirable difficulty. His argument, developed with Elizabeth Bjork across many studies, is that conditions which make learning feel slower and harder in the moment often make it last much longer, while conditions that make learning feel smooth often produce knowledge that evaporates. The strain of recall is not a sign that the practice is failing. The strain is the practice.

This changes how you should read your own sessions. A word you can spell instantly needs no more work today; spelling it again is pleasant and nearly useless. A word you have to fight for is exactly the word that benefits.

It also explains timing. Recall helps most when it is effortful but still possible, so the ideal moment to test a word is shortly before you would have forgotten it. Test too soon and the fetch is trivial. Wait too long and there is nothing left to fetch, so you are learning from scratch. That window, and the way it widens each time you succeed, is the logic behind spaced repetition, which our guide to the science of spaced repetition explains in full.

Getting it wrong first

Many careful learners, and many careful parents, try to prevent mistakes altogether, on the theory that a wrong spelling, once written, will stick. The research gives a more interesting answer.

In 2009, Nate Kornell, Matthew Hays and Robert Bjork asked people to guess an answer before they were shown it. Most guesses were wrong; the items were chosen so that people would usually fail. Yet on a later test, those people remembered the correct answers better than people who had spent the same time simply studying them. An unsuccessful attempt followed by the right answer beat no attempt at all.

One plausible explanation is that the attempt primes the search. You reach for the answer, stir up everything nearby, and when the correct answer arrives it has somewhere to land. Curiosity may help too. An answer you just missed is an answer you want.

There is a boundary, though, and for spelling it matters a great deal. The benefit depends on quick correction. A wrong spelling that is never checked is simply practice at the wrong spelling, and there is evidence that exposure to misspellings does real harm. In a 1990 study with the splendid title "Reading student essays may be hazardous to your spelling," Larry Jacoby and Ann Hollingshead found that reading misspelled words made people somewhat more likely to misspell them afterward. So the lesson is not to avoid errors. It is never to leave one standing. Attempt, check at once, and try again later. The idea is developed further in why you remember a word better after getting it wrong.

Look, cover, write, check, done properly

Generations of schoolchildren have been taught a routine called look, cover, write, check. It is plain, a little old-fashioned, and remarkably well aligned with the research, provided the cover step is taken seriously. Everything depends on it. Once the word is hidden, writing it becomes retrieval, and the check supplies the immediate correction that makes a miss useful. Skip the cover, copy the word while looking at it, and the routine collapses back into tracing.

A version that makes the most of the testing effect:

1. Say the word aloud and listen to its parts.
2. Look at it for a few seconds and name the part most likely to trip you.
3. Cover it completely, and wait a few seconds before writing.
4. Write or type the whole word from memory.
5. Uncover it and compare letter by letter.
6. If it was wrong, write it correctly once, then put it back in the pile for later.

The short wait in step three deserves a defense. Without it, you can hold the letters in short-term memory, the way you hold a phone number between reading it and dialing, and the exercise becomes copying with extra steps. A few seconds of distraction makes the retrieval real.

Better still is to replace "look" with "listen." If the word is spoken and you have to produce it without seeing it in that session, there is nothing to copy at all. That is how Spelling.School is built: a recorded voice says the word, you type it, and the app marks exactly which letters went wrong before bringing the word back after a longer gap. A patient friend, or a voice memo of yourself reading your list, does the same job.

How to study a list this way

Begin with a test, not a study session. It sounds backward, but it tells you where to spend your time. Have the list read to you and spell every word cold. The list splits into words you already have and words you do not, and only the second group needs work.

Work on the misses with the cover-and-check routine, then test them again later the same day or the next morning. Words you get right move to a longer gap; words you miss stay close. Shuffle the order every time, because a list recited top to bottom teaches the sequence as much as the words.

Keep sessions short. Ten minutes a day across a week beats an hour on Sunday, for reasons that have as much to do with spacing and sleep as with recall, and the two effects are strongest together. A word spelled correctly on three separate days is far more secure than one spelled correctly three times in a row.

For a child's weekly list, that means a short test on Monday, practice on the misses through Wednesday, and a light check on Thursday, instead of one tearful push the night before. For an adult, it means a personal list of the dozen or so words that genuinely cause trouble, drawn from your own writing, and a few minutes of recall on most days.

How often to test yourself

A reasonable default is to test a new word on the day you meet it, again the next day, then after three or four days, a week, and a few weeks. Each success earns a longer gap. Each failure sends the word back to the start.

You can run this with three envelopes. The first holds words you missed today and is tested daily. The second holds words you have spelled correctly once and is tested every few days. The third holds words you have spelled correctly on three different days and is tested weekly. Any miss, from any envelope, sends a word back to the first.

Research by Nicholas Cepeda, Harold Pashler and their colleagues suggests the best gap depends on how long you need the memory to last: short gaps for a test on Friday, stretching gaps for a word you want for life. For the fuller story of why memory fades on the schedule it does, read about the forgetting curve.

What to do next

Choose five words you reliably get wrong. Write them on a card and turn it face down. Tonight, have someone read them to you, or record yourself saying them and play the recording back, and spell each one from memory. Check against the card and circle the misses.

Tomorrow, test only the circled words; two days later, test all five. Keep that up for a week, a few minutes at a time, and you will feel, quite physically, the difference between knowing what a word looks like and being able to produce it.

If you would rather have the words spoken and scheduled for you, the five-word demo on the home page needs no sign-up. Either way, the rule is the same. Stop looking. Start fetching.

Questions people ask

What is the testing effect?

The testing effect is the finding that trying to recall information from memory strengthens it more than studying the same information again. It is also called retrieval practice. For spelling, it means hiding a word, trying to write it from memory, and then checking, instead of reading the word over and over. The effect has been shown in hundreds of experiments and is one of the most reliable findings about learning.

Is it better to test yourself or reread?

Testing yourself is better for anything you need to remember beyond the next few minutes. In a well-known 2006 study by Henry Roediger and Jeffrey Karpicke, students who practiced recall remembered far more a week later than students who reread the same material, even though the rereaders felt more confident. Rereading builds familiarity, which feels like knowledge but does not give you the ability to produce a word on a blank page.

Does getting a spelling wrong first help you learn it?

It can, provided you see the correct spelling straight away. Research by Kornell, Hays and Bjork in 2009 found that attempting an answer before seeing it improved later memory, even when the attempt was wrong. The danger is a wrong spelling that is never corrected, which becomes practice at the error. Always check immediately after an attempt, and try the word again later.

How often should I test myself on spelling?

Test a new word on the day you learn it and again the next day. If you get it right, wait three or four days, then a week, then a few weeks before the next try. Any word you miss goes back to daily practice until you have spelled it correctly on several separate days. Short sessions of around ten minutes, most days, work better than one long session a week.

Why does look, cover, write, check work?

Look, cover, write, check works because the cover step forces you to produce the word from memory, and that act of recall strengthens memory far more than copying does. The check step then gives immediate feedback, which turns a wrong attempt into learning. The method stops working if the cover step is skipped or rushed, because writing a word you can still see, or still hold in short-term memory, is only copying.

Why can I read a word easily but not spell it?

Reading only requires you to recognize a word that is already in front of you, and recognition can work from partial information such as the first and last letters and the overall shape. Spelling requires you to produce every letter in the right order with nothing to go on. That is why keen readers can misspell words they have seen thousands of times. Practice that makes you produce the word, rather than look at it, closes the gap.

Fix the words you actually get wrong

A few minutes a day, spaced repetition, built only for spelling. Try five words without signing up.