Interleaved Practice for Math: Why Mixing Problem Types Works
What interleaved practice is, what a 54-class randomized trial found (61% vs 38% a month later), why mixing problem types works, and how to use it for the GMAT.
By the Relvora team · Updated · 5 min read
Most math practice is blocked: a page of percent problems, then a page of rate problems. It feels efficient, and your accuracy climbs fast. Research suggests it's also one of the least effective ways to prepare for a test where problem types arrive mixed.
In short: Interleaved practice mixes different problem types in one session, so you must choose the method for each problem yourself. In a randomized trial with 54 seventh-grade math classes, interleaved practice led to 61% on a surprise test a month later, against 38% for blocked practice. For GMAT Quant, that means mixed sets, naming the method before solving, and judging progress by delayed, mixed checks rather than in-session streaks.
What is interleaved practice?
Blocked practice groups problems by type: ten percent problems, then ten work-rate problems. Once you've solved the first one, you know the method for the next nine.
Interleaved practice mixes the types: a percent problem, then a rate problem, then a remainder problem, then another percent problem. Each time, you have to work out what kind of problem it is before you can solve it.
Interleaving isn't just "doing more variety." It works best when it mixes problems that look similar but need different methods, so you're forced to tell them apart.
What does the research show?
The strongest evidence comes from a cluster-randomized trial by Doug Rohrer, Robert Dedrick, Marissa Hartwig and Chi-Ngai Cheung, published in the Journal of Educational Psychology in 2020.
- Who: 54 seventh-grade math classes, 787 students in five schools in one school district.
- What: over about four months, classes were randomly assigned to do their regular practice assignments either interleaved or blocked. Both groups saw the same practice problems; only the order changed.
- Result: on an unannounced test one month later, the interleaved group scored 61% and the blocked group 38%. The effect size was d = 0.83, which is large for an education study.
The U.S. Institute of Education Sciences' What Works Clearinghouse reviewed the study and rated it as meeting its standards without reservations, its highest rating for study design.
The percentages above are from the published paper. The Clearinghouse's own summary reports the two groups' means slightly differently, but it shows the same large, statistically significant gap.
Why does mixing help?
The researchers' explanation is about choosing a strategy. In a blocked assignment, the heading tells you the method, so you practice executing it but never selecting it. On a real exam, selection is half the job. Mixed practice trains both.
Interleaving also spaces practice out. When problems of one type are spread across many sessions, you return to each type after a gap. A large meta-analysis by Cepeda and colleagues (2006, Psychological Bulletin) found that spaced practice reliably beats massed practice for retention. Many studies have also found that retrieving something from memory strengthens it more than reviewing it does, the "testing effect" shown by Roediger and Karpicke (2006).
What are the caveats?
Be honest about what the trial does and doesn't show:
- The students were 12- and 13-year-olds, not adults preparing for a graduate admissions test. The mechanism should carry over, but this trial didn't test GMAT or GRE preparation.
- Interleaved practice can take longer. Each problem starts with working out the method, so a mixed set usually takes more time per problem than a blocked one.
- It feels worse. In-session accuracy is usually lower with mixed practice. That makes blocked practice look more effective than it is.
What does interleaved practice look like for GMAT Quant?
Here is a mini mixed set of four original questions. Before solving each, name the topic and your first move.
A price drops from $80 to $68. What is the percent decrease? First move: change ÷ original. 12 ÷ 80 = 15%.
One worker paints a fence in 4 hours, and another in 12 hours. How long do they take working together? First move: add rates, not times. 1/4 + 1/12 = 3/12 + 1/12 = 4/12 = 1/3 of the fence per hour, so 3 hours.
What is the units digit of 3²³? First move: find the cycle. Units digits of powers of 3 repeat 3, 9, 7, 1. Since 23 = 4 × 5 + 3, 3²³ has the third digit in the cycle: 7.
The average of 5 numbers is 12. One number is removed, and the average of the remaining 4 is 10. What number was removed? First move: switch to totals. 5 × 12 = 60 and 4 × 10 = 40, so the removed number is 60 − 40 = 20.
In a blocked set, question 2 would sit among nine other work-rate problems and you'd add rates automatically. Mixed in like this, you have to notice that "together" means adding rates. Adding times is a classic trap (see 10 common GMAT Quant traps).
How do you apply interleaving in your own study?
- Learn a new method with a few examples in a row. When a method is brand new, a short block of worked examples helps you follow it.
- Switch to mixed sets quickly. Once you can follow the method, put it into sets with other topics.
- Name the method before solving. Say or write "work rate, add rates" before you touch the numbers. This is the skill that blocked practice skips.
- Keep old topics in rotation. Each session should include topics from earlier weeks, not just this week's.
- Don't judge by in-session accuracy. A lower score on a mixed set doesn't mean you're learning less.
- Check later. The real test of a skill is a fresh, mixed question a few days later. See spaced practice for the GMAT.
How does Relvora use interleaving?
Relvora Prep builds interleaving into its mastery rule. A skill counts as mastered only after mixed practice where you get at least 4 of your last 5 answers right within 1.5 times the target time, across different questions and patterns. Mastery is then confirmed by a correct answer 2 or more days later and followed by spaced reviews. A streak inside one blocked session doesn't count.
The app also has a no-pen mode for practice away from your desk. Questions that need working out become 20-second speed cards where you pick the fastest first move. That trains method selection directly. Speed cards are tracked separately and never count toward mastery.
Relvora covers GMAT Quant, GMAT Data Insights, GRE Quant and SAT Math, and has a free plan. To see what a mixed set feels like, try your free 3-minute set with a free account. To find out which skills to mix in first, take the free diagnostic.
GMAT™ is a trademark of GMAC. GRE® is a registered trademark of ETS. Relvora is not affiliated with or endorsed by either.
Sources
- Rohrer, Dedrick, Hartwig & Cheung (2020), A randomized controlled trial of interleaved mathematics practice, Journal of Educational Psychology 112(1), 40–52 (DOI)
- What Works Clearinghouse review of the Rohrer et al. study (IES)
- Efficacy study of interleaved mathematics practice, project page (IES)
- Cepeda, Pashler, Vul, Wixted & Rohrer (2006), Distributed practice in verbal recall tasks, Psychological Bulletin 132(3) (PubMed)
- Roediger & Karpicke (2006), Test-enhanced learning, Psychological Science 17(3) (PDF)
Frequently asked questions
What is interleaved practice in math?
Interleaved practice means mixing different types of problems in one session instead of doing a block of the same type. It makes you decide which method each problem needs, which is what a real exam asks you to do.
Is interleaved practice better than blocked practice?
In a randomized trial of 54 seventh-grade math classes, students who did interleaved practice scored 61% on a surprise test a month later, against 38% for blocked practice. Blocked practice often feels better during the session, but the mixed group remembered more.
Why does interleaved practice feel harder?
Each problem can be any type, so you must work out the method before you solve it. That extra step lowers accuracy during practice. The same step is what helps you on test day, where questions also arrive mixed.
Should I ever practice one topic at a time?
A few examples in a row are reasonable when you first learn a new method. Once you can follow the method, switch to mixed sets so you also practice recognizing when to use it.