Prep
What interleaving means
Interleaving means mixing different topics or problem types within one study session, instead of finishing all practice on one topic before moving to the next. If you are learning three kinds of problems, A, B and C, blocked practice looks like AAA BBB CCC. Interleaved practice looks like ABC BCA CAB.
The idea is simple: on the real exam, questions arrive in a random order and nothing tells you which topic each one belongs to. Blocked practice lets you skip the step of working out what kind of problem you are facing, because the chapter heading already told you. Interleaving forces you to make that decision every time.
The evidence is real but more mixed than for spaced repetition or practice testing. Interleaving helps a lot for some material and not at all, or even hurts, for others. This page explains where the line falls.
| Material | Evidence | Our recommendation |
|---|---|---|
| Similar, easily confused categories (e.g. access control models, AWS storage services) | Interleaving helps most here; paintings g = 0.67 | Interleave once you know the basics of each |
| Mathematical problem types | Small benefit, g = 0.34; two large classroom studies found no effect | Mix problem types in practice, but do not expect a large gain |
| Expository text | Ambiguous | Learn in blocks, then mix practice questions |
| Word lists, vocabulary, acronyms | Blocking was better, g = -0.39 | Use flashcards with spacing, not interleaving |
| Very different categories | Blocking helps learners see what each category has in common | Block on first learning |
What the research shows
A moderate effect overall
The largest meta-analysis, by Brunmair & Richter, 2019, pooled 59 studies and 238 effect sizes. It found a moderate overall interleaving effect, g = 0.42. (Hedges' g is read like Cohen's d: about 0.2 is small, 0.5 moderate, 0.8 large. So 0.42 is a small-to-moderate benefit.)
The average hides big differences between types of material:
- Paintings (learning to tell artists' styles apart): g = 0.67, a moderate-to-large benefit.
- Mathematical tasks: g = 0.34, a small benefit.
- Expository texts: ambiguous results.
- Words (such as vocabulary lists): blocking was better, g = -0.39.
Similarity is the key moderator
Brunmair and Richter concluded that interleaving helps most when the categories are similar to each other and easy to confuse. Carvalho & Goldstone, 2014 explain why: interleaving puts similar items side by side and highlights the differences between categories, while blocking highlights what items in the same category have in common. So interleaving pays off when your main problem is telling similar things apart, and blocking pays off when categories are already very different and you still need to grasp what each one is.
Where it has failed
Honesty matters here. Rowlandson et al., 2025 ran two large classroom experiments on learning defined mathematical categories and repeatedly failed to detect any difference between interleaving and blocking. Interleaving is not a universal rule.
The ten-technique reviews reflect this middle position. Dunlosky et al., 2013 rated interleaved practice "moderate utility", below practice testing and distributed practice. In the meta-analysis by Donoghue & Hattie, 2021, interleaved practice had an overall effect of d = 0.47, below distributed practice (0.85) and practice testing (0.74). Their breakdown by level is striking: d = 0.65 across 79 effects for university students, but 0.00 for secondary and -0.20 for primary students. Adult learners preparing for professional exams are the group where it looks most useful.
Why interleaving feels worse while you do it
Interleaved practice usually feels slower and produces more errors during practice. That is part of why learners avoid it. Interleaving is a classic desirable difficulty: it feels effortful, so learners read the effort as a sign of poor learning, when it often predicts better long-term performance (de Bruin et al., 2023).
A practical consequence: do not judge interleaving by how your practice session went. Judge it by your score on a mixed set a week later. Students who block their practice see high scores during practice and are then surprised when the exam mixes everything up.
Interleaving in certification exams
Certification exams are full of confusable categories, which is exactly where the meta-analysis found interleaving helps most. Examples:
- Security+ and CISSP: access control models (discretionary, mandatory, role-based, attribute-based), attack types that look alike (phishing variants, injection types), control categories and types.
- Network+ and CCNA: routing protocols, wireless standards, similar troubleshooting symptoms with different causes.
- AWS and AZ-900: storage and database services that solve overlapping problems, pricing models, support plans.
- PMP and CAPM: the right response under predictive versus agile versus hybrid approaches.
For these, the hard part of the exam question is often not recalling a fact but recognising which concept the scenario is about. Interleaved practice trains exactly that.
For material that is mostly isolated facts, such as port numbers or acronyms, the evidence for words suggests interleaving adds little. Use active recall and spacing for those.
How to interleave without getting lost
Our recommendation, combining the similarity findings from Carvalho and Goldstone with how certification courses are structured:
- Learn new topics in a block first. On first contact you need to understand what each category is. That is where blocking helps.
- Interleave as soon as you have two or more similar topics. Mix practice questions across them, especially topics you tend to confuse.
- Make the comparison explicit. After each mixed set, write one line per confused pair: "X is used when..., Y is used when...". A comparison table works well.
- Interleave old with new. Mixing in questions from earlier topics also spaces them out, so you get spacing for free.
- End with fully mixed sets. In the final weeks, every practice set should be random across all exam domains, like the real exam.
What an interleaved session looks like
Here is a 45-minute session for someone who has already studied four related Security+ topics: symmetric encryption, asymmetric encryption, hashing and digital signatures. These are classic confusables, because all four appear in questions about protecting data, and the scenario usually describes a goal rather than naming the tool.
- Minutes 0 to 25: 20 mixed questions. Five per topic, shuffled. Before answering each one, name the category in a word: "hashing", "signature". Then answer.
- Minutes 25 to 35: review. For every miss, note whether you got the category wrong or the detail wrong. Category errors are the ones interleaving fixes.
- Minutes 35 to 45: comparison table. Four rows, three columns: what it guarantees, what key it uses, a typical exam clue. Write it from memory, then check.
A blocked version of the same session, 20 questions on hashing in a row, would feel smoother and score higher on the day. It would also never make you choose between hashing and a digital signature, which is the decision the exam actually asks for.
The same template works for any set of confusable topics: AWS storage classes, cloud service models, risk responses in the PMP, or routing protocols for the CCNA.
Limits of the evidence
- Most studies use short lab tasks: paintings, bird species, simple maths problems. Few test complex professional content over weeks.
- Classroom results are mixed, as Rowlandson et al. show.
- Interleaving dissimilar topics is unlikely to help and may slow initial learning.
- Interleaving is often confounded with spacing in real study, because mixing topics automatically spreads each one out. Some of its real-world benefit may come from spacing.
- We know of no trial of interleaving with certification candidates specifically, so the protocol below is our application of the research, not a tested programme.
How to use this for a certification exam
Example: CompTIA Security+, eight weeks out, five study days per week.
- Day 1. Take our free Security+ practice test. Note which questions you missed because you confused two similar concepts. Those pairs are your interleaving list.
- Weeks 1 to 3: block, then mix. Learn each new domain in a block. From week 2, end each session with 10 mixed questions: 5 on today's topic and 5 on similar topics from earlier.
- Weeks 4 to 6: confusable pairs. Twice a week, do a 20-question set built only from your interleaving list. After each set, update a one-page comparison table of the confused pairs.
- Weeks 7 to 8: full mixing. Every practice set is 30 to 40 questions drawn at random from all domains. Retake our free practice test at the start of week 7 and week 8 and compare with day 1.
- Check. If your mixed-set score is more than 10 points below your score on single-topic sets, you are relying on context. Add more interleaved sets.
For networking exams, our Network+ practice test works the same way. For how interleaving fits with the other techniques, see study techniques ranked by evidence.