Book a free consultation

From the classroom

Three study habits that work, and the two your child is probably using instead

8 min read

Key takeaways

  • Testing beats reviewing. Closing the book and trying to produce the answer builds memory. Looking at the answer does not.
  • Four short sessions beat one long one. The gap between sessions is doing the work, not the total minutes.
  • Mixed problem sets beat blocked ones. In a randomized trial in Grade 7 math classrooms, mixed practice scored 61 per cent against 38 per cent.
  • Highlighting and rereading are the two techniques students use most, and both were rated low utility in the largest review of the evidence.

In 2013, five cognitive scientists sat down with several decades of research and did something nobody had quite done before. They took the ten study techniques students actually use, graded each one, and published the results.

The two techniques your child almost certainly uses came bottom.

What the researchers actually did

The review was led by John Dunlosky at Kent State University, with Katherine Rawson, Elizabeth Marsh at Duke, Mitchell Nathan at Wisconsin and Daniel Willingham at the University of Virginia. It ran in Psychological Science in the Public Interest, and it is still the reference point twelve years later.

They did not just ask whether a technique works once. They asked whether it works across different ages, different ability levels, different subjects, and different kinds of test, because a technique that only helps clever seventeen-year-olds with multiple choice is not much use to a parent. Then they graded each one high, moderate or low.

TechniqueRatingWhat it means
Practice testingHighQuizzing yourself instead of reviewing
Distributed practiceHighSpreading study out over days
Interleaved practiceModerateMixing problem types together
Elaborative interrogationModerateAsking why a fact is true
Self-explanationModerateExplaining your own reasoning aloud
SummarizationLowWriting summaries of what you read
HighlightingLowMarking up the textbook
Keyword mnemonicLowMemory tricks built on keywords
Imagery for textLowForming mental pictures while reading
RereadingLowGoing over the notes again
Utility ratings from Dunlosky, Rawson, Marsh, Nathan and Willingham, Psychological Science in the Public Interest, 2013.

Their words on the bottom of that table: “Five techniques received a low utility assessment: summarization, highlighting, the keyword mnemonic, imagery use for text learning, and rereading.”

Here are the three worth teaching, in the order I would introduce them.

1. Test, do not review

Practice testing earned one of only two high ratings. Dunlosky and his colleagues wrote that it and distributed practice “benefit learners of different ages and abilities and have been shown to boost students’ performance across many criterion tasks”.

The idea is almost annoyingly simple. Instead of looking at the material again, close it and try to produce it. The effort of pulling something out of your head is what makes it stick. Looking at it does not.

What it looks like on math homework

Cover the worked example at the top of the page, then do it from scratch. If your child can only do the question with the example visible, they cannot do the question.

Start from the end of the exercise set. Question 20 is where the interesting problems live. Questions 1 to 5 are usually a warm-up designed to be done without thinking, which is precisely the state you do not want.

And ask for the answer before the book gives it. “Roughly what should this come out to?” costs nothing and turns a reading exercise into a retrieval one.

Fair warning: this feels worse than reviewing. That is the point. Reviewing feels smooth and produces very little. Testing feels effortful and produces most of the learning.

2. Spread it out

The other high rating went to distributed practice, which is a technical way of saying do not cram.

The evidence here is enormous. Nicholas Cepeda, Harold Pashler and colleagues pulled together 839 separate assessments of spaced practice, drawn from 317 experiments across 184 published articles, and analyzed them in Psychological Bulletin. It is one of the most thoroughly established findings in the study of memory.

The part that matters for a parent is subtler than “start earlier”. What they found is that the ideal gap between study sessions grows with how long you need to remember the material. If the test is on Friday, a short gap is fine. If the material has to survive until the June exam, or until next year’s course builds on it, the gaps need to be longer and the returns keep coming.

Which explains something most families have watched happen. Cramming on Thursday genuinely works for Friday, and that is exactly why children keep doing it. It stops working in June, and by then nobody connects the two.

What it looks like on a school week

Twenty minutes on four nights beats eighty minutes on Thursday. Same total time, a different result, and it is an easier sell to a tired thirteen-year-old than an hour and a half in one sitting.

Then keep going back. When a unit finishes, it should still turn up once a fortnight. The natural instinct is to consider a unit done when the test on it is done, and that instinct is the whole problem.

3. Mix the problem types together

This is the one almost nobody does, and in math it may be the most valuable of the three.

A normal math worksheet is blocked. Twenty questions, all quadratics. Interleaved practice mixes them, so no two consecutive problems need the same method.

Doug Rohrer and colleagues at the University of South Florida tested this properly. They ran a randomized controlled trial across 54 Grade 7 math classes in five Florida schools, 787 students and 15 teachers. Every class did the same eight practice worksheets over roughly 103 days. The only difference was the order of the questions. Then, a month after the last worksheet, an unannounced test.

Test scores one month after practice, by how the practice was ordered
Mixed problem types61%
Blocked by topic, the usual way38%

Rohrer, Dedrick, Hartwig and Cheung, Journal of Educational Psychology, 2020. 787 Grade 7 students across 54 classes. Effect size d = 0.83.

Sixty-one per cent against thirty-eight. Same worksheets, same teachers, same amount of practice. Only the order changed.

The reason is worth understanding, because once you see it you cannot unsee it. A blocked worksheet tells your child which method to use before they read the first question. The heading says “Factoring”, so every answer is factoring. They practice executing a method they were handed.

A test does not hand it to them. A test gives them a problem and expects them to work out which method it needs, which is a completely different skill and one that blocked practice never touches. That is why a child can do every homework question and still open the test and go blank. They did not fail to learn the method. They never practiced choosing it.

One honest note. Dunlosky’s team rated interleaving only moderate in 2013, because at that point the evidence was thin. The strongest evidence arrived seven years later, and it came out of real math classrooms.

What it looks like at your kitchen table

When a unit ends, do not file it away. Build a short set of six questions pulled from three different units and shuffle them. Six mixed questions are worth more than twenty from the same section.

You do not need to write the questions. Take them from old homework, cover the headings, and change the order.

The two to stop, and why they feel like they work

Highlighting and rereading are the two techniques students reach for most, and both were rated low utility.

They are not useless because students are lazy. They are useless because they are convincing. Reading something for the third time makes it feel familiar, and familiarity feels almost exactly like knowing. Your child finishes an hour of rereading genuinely believing they have got it.

Then the test asks them to produce it from nothing, and recognition turns out not to be the same thing as recall at all.

If your child’s textbook is more yellow than white, that is not studying. It is decorating.

You do not have to ban either one. Just make them the first ten minutes rather than the whole hour, and spend the rest of the time on the three above.

What a week actually looks like

Putting all three together, for a test two weeks out:

  1. Four short sessions a week rather than one long one. Twenty minutes is plenty.
  2. Every session starts closed book. Try the problem first, check afterwards.
  3. Every session includes at least two questions from an earlier unit.
  4. The night before is for sleep, not for the first real attempt at the material.

None of this takes more time than what your child is doing now. It is the same hours, arranged differently.

Where we come in

Most of what we do is these three things, done consistently by somebody who is not their parent. That last part matters more than it should. A child will argue with you about closing the book and will not argue with a tutor, and no amount of being right about the research changes that.

A related piece, if this was useful: how to tell whether your child understands the math, or has just memorized it.

Sources

William Thieu, founder of InnEd Academy

About the author

William Thieu

Founder, InnEd Academy

William founded InnEd Academy in Whitchurch-Stouffville in 2021. He grew up without a tutor or a mentor, turned his own grades around in Grade 12, and came out of it with offers from Waterloo Engineering and Ivey Business. He now runs a small-group math program from Kindergarten to Grade 12, four children to one tutor, and has worked with more than 150 local families. He writes here about the Ontario curriculum and what he sees at the whiteboard.

Connect on LinkedIn

Questions about your own child?

Thirty complimentary minutes with William. No pressure and no sales script, and we will tell you honestly whether we are the right fit.

Book a free consultation

Back to all articles

Book a free consultation Call