TL;DR. Matthew Youlden, who grew up in Manchester and speaks nine languages to an advanced level, has a widely shared set of tactics. They are sensible and most of them help, but they are not all the same kind of thing. Some touch the memory mechanism directly; others simply keep you at the desk. Separating the two is the most useful step you can take when building a language plan. The measured evidence says this: in a meta-analysis of 242 studies, 1,619 effects and 169,179 participants, the strongest effects come from distributed practice at d = 0.85 and practice testing at d = 0.74, while re-reading sits at d = 0.47. Against that, interleaving runs the other way on word-based material at g = -0.39, meaning vocabulary and set phrases are better blocked than mixed. Below we label all twelve tactics against that evidence and separate the motivation tactics from the memory ones.
First, the distinction: motivation tactic or memory tactic?
Most language advice conflates two different jobs.
Motivation tactics keep you still studying six months from now. They have no measured memory effect, but without them no memory tactic survives long enough to be applied.
Memory tactics change how durably the material sticks. Their effect sizes have been measured.
You need both. But if you cannot tell which is which, you treat motivation tactics as learning methods and invest in the wrong place.
CEOtudent editorial framework: twelve tactics, labelled against the evidence
This table is a CEOtudent assessment. The figures in the evidence column are quoted from published meta-analyses; matching each tactic to that evidence is editorial interpretation. Labelling a tactic “motivation” does not mean it fails, only that it cannot be matched to a measured memory effect.
| # | Tactic | Type | Evidence status |
|---|---|---|---|
| 1 | Have a genuinely strong reason | Motivation | No direct memory effect measured; the condition that makes persistence possible |
| 2 | Commit and make yourself a promise | Motivation | Same; raises the chance the protocol is actually followed |
| 3 | Find a learning partner | Motivation and memory | Where partner conversation becomes recall under pressure, it enters the d = 0.74 band |
| 4 | Compete with your partner | Motivation | No measured memory effect |
| 5 | Learn by practising | Memory | Practice testing d = 0.74 against re-reading d = 0.47 |
| 6 | Talk to yourself | Memory | Self explanation d = 0.54; effective because it requires production from memory |
| 7 | Have fun in the language, write songs | Motivation and memory | Approaches practice testing to the extent that it involves production |
| 8 | Learn like a child, make mistakes | Memory | Errors that receive feedback; feedback is reported as an important moderator |
| 9 | Leave your comfort zone | Motivation | No measured memory effect; increases exposure |
| 10 | Listen | Memory | An input condition; does not substitute for a recall attempt on its own |
| 11 | Watch people who speak the language | Motivation and input | Passive watching sits near re-reading; limited unless combined with production |
| 12 | Do not be embarrassed by mistakes | Motivation | The companion to tactic nine |
The reading: only five of the twelve tactics connect directly to a memory mechanism. The rest are genuinely valuable but do a different job. If you have applied the list as a learning method, failed to progress and blamed yourself, the problem is not you. It is how the list is presented.
What the list is missing: a schedule
The single highest-scoring element is absent from all twelve tactics: when the review happens.
Distributed practice sits at the top of the table at d = 0.85. And the size of the gap has been studied too. A study with more than 1,350 people found that the optimal gap depends on how long you need to retain the material. For a one-week retention horizon the optimal gap is roughly 20 to 40% of that horizon; for a one-year horizon it is roughly 5 to 10%.
CEOtudent derived schedule: review gaps for language study
| How long you need to retain it | Optimal gap as a share of that horizon | Derived gap |
|---|---|---|
| 1 week (an exam, a presentation, a trip) | 20% to 40% | about 1.4 to 2.8 days |
| 1 month | not reported | not reported |
| 3 months | not reported | not reported |
| 1 year (durable command of the language) | 5% to 10% | about 18 to 37 days |
Derived by CEOtudent from the two published ratios. Arithmetic: 7 x 0.20 = 1.4 and 7 x 0.40 = 2.8; 365 x 0.05 = 18.25 and 365 x 0.10 = 36.5. Intermediate horizons are left empty because the study anchors only those two points, and interpolating would be invention rather than finding.
The practical reading: if you need a vocabulary set for an exam next week, review every one or two days. If you need it permanently, roughly monthly is enough. Note that the proportion shrinks while the actual interval grows, which is the point popular summaries most often get backwards.
Where the standard advice breaks: interleaving vocabulary
A common item in study advice is to mix topics rather than block them. Applying that literally to language learning is a mistake.
A meta-analysis of 59 studies, 238 effect sizes and 158 samples found a moderate overall interleaving effect at Hedges’ g = 0.42. But the same analysis reported an advantage for blocked practice in word-based studies, at g = -0.39, with ambiguous and non-significant results for expository texts.
The translation into language learning is clean: block your vocabulary and set phrases, interleave your grammar problems and conversation scenarios. Shuffling a word list at random is not a move the evidence supports.
The second break: the outcomes measured were mostly surface ones
The meta-analysis producing those effect sizes states its own limit plainly: the majority of the studies were based on surface or factual outcomes, and caution is needed when applying the findings to deeper and more relational outcomes.
Vocabulary and conjugation tables are factual outcomes, and the numbers apply well there. Speaking fluently, catching humour and holding an argument are relational outcomes. Do not read d = 0.85 as a guarantee of fluency.
The same analysis carries a third warning: effects were much greater for lower-ability than higher-ability learners. Someone learning their first foreign language gains more from these techniques than someone learning their third.
So what should you actually do?
Six steps, each with the evidence it rests on.
- Find a reason and a partner. (Tactics 1, 2, 3, 4.) These are not memory tactics; they are the survival condition for your protocol.
- Decide your retention horizon. For an exam, or permanently? The schedule follows from this.
- Set review gaps in absolute days. Roughly monthly for a permanent goal, one to two days for a short-term one.
- Start every session with recall. Try to remember before you look at the list, then check. The gap between re-reading (d = 0.47) and practice testing (d = 0.74) comes from exactly this.
- Block the words, interleave the scenarios. Study the vocabulary set together, mix the conversational situations.
- Make feedback compulsory. Making mistakes (tactics 8, 9, 12) only becomes learning when the mistake is corrected; the presence of feedback is reported as an important moderator.
The CEO and student split applies here too. The student half asks which technique is better. The CEO half asks who is going to build the schedule, sees that the answer is nobody else, and writes the plan.
FAQ
So do the twelve tactics not work?
They do, but not all in the same way. Five connect directly to a memory mechanism; the rest keep you working. Both are necessary, but treating a motivation tactic as a memory method leads to a planning error.
How many hours a day should I study?
What the measured evidence speaks to is not duration but distribution. Spreading the same total time rather than blocking it is the source of distributed practice’s d = 0.85. One hour on three days beats three hours on one day.
Are flashcards still valid?
Yes, because the method is inherently practice testing (d = 0.74), and combined with spaced repetition software it applies the top two techniques in the table at once. Reading the card and moving on does not count; you have to actually attempt recall.
Is watching films and series enough?
Not on its own. In effect-size terms, passive watching sits close to re-reading. Unless it is combined with production, meaning speaking, writing or a recall attempt, it remains input.
Should I shuffle my vocabulary?
No. Blocked practice wins on word-based material at g = -0.39. Save the mixing for problem types and conversation scenarios.
Is learning a language as an adult genuinely harder?
The figures in this article do not answer that. What they say is that the measured benefit of these techniques is larger for learners at a lower starting level. That is a reason for an experienced learner to design the protocol more carefully, not to abandon it.
Sources and further reading
- Donoghue, G. M. and Hattie, J. A. C. (2021). A Meta-Analysis of Ten Learning Techniques. Frontiers in Education, volume 6, article 581216. Effect sizes by technique, overall mean d = 0.56, and moderator findings.
- Brunmair, M. and Richter, T. (2019). Similarity matters: A meta-analysis of interleaved learning and its moderators. Psychological Bulletin, volume 145, issue 11, pages 1029 to 1052. Overall interleaving effect and its variation by material.
- Cepeda, N. J., Vul, E., Rohrer, D., Wixted, J. T. and Pashler, H. (2008). Spacing effects in learning: a temporal ridgeline of optimal retention. Psychological Science, volume 19, issue 11, pages 1095 to 1102. Optimal gap as a proportion of the retention interval.
- Adesope, O. O., Trevisan, D. A. and Sundararajan, N. (2017). Rethinking the Use of Tests: A Meta-Analysis of Practice Testing. Review of Educational Research, volume 87, issue 3, pages 659 to 701. Practice testing against restudying.
For the wider framework behind these protocols, see our piece on fast upskilling protocols and our ranking of 12 study techniques by research strength.
This content was compiled with the support of AI following in-depth research, then written and prepared for publication by the CEOtudent editorial team.
This post is also available in:
