What the Research Says About Learning

One 90-Minute Workshop Cut the Performance Gap Between Top and Bottom Students by 50% — Here’s What They Taught

Here’s a number that should make you stop scrolling: a single 90-minute workshop on learning strategies, followed by three coaching sessions for the lowest-performing students, reduced the final-exam performance gap between high and low performers by roughly 50% compared to a previous cohort that received no training. That’s not a marginal gain from a fancy app or a biohacking supplement. That’s a half-day investment that nearly halved the distance between the students who were failing and the ones who were acing it. The intervention wasn’t about studying harder. It was about studying differently — and the specific strategies they taught are things you can test on yourself this week.

What the research actually shows

The “Study Smart” program, tested on 125 first-year pharmacology students and compared to a historical cohort of 120 students, targeted the gap between what cognitive science knows and what students actually do. The researchers didn’t just lecture. They ran a 90-minute interactive workshop on why common strategies like rereading and highlighting are weak, and how to replace them with distributed practice (spreading study sessions over time), interleaving (mixing topics), elaboration (explaining in your own words), and self-testing. Then, the bottom 20% of performers on the first midterm got three 30–45 minute coaching sessions to troubleshoot implementation.

The result wasn’t subtle. The final-exam performance gap between high- and low-performing students shrank by roughly 50% compared to the previous year’s cohort. That’s a massive effect size for a low-cost, low-tech intervention. But here’s the catch: the comparator was a historical control group, not a randomized trial within the same semester. That means we can’t rule out cohort differences — maybe the intervention year had smarter students, easier exams, or different grading. Still, the pattern is consistent with decades of cognitive load theory and memory research.

Speaking of which: Cognitive Load Theory (CLT), developed by John Sweller over decades, provides the theoretical backbone for why these strategies work. CLT argues that our working memory has a very limited capacity — you can hold roughly 4–7 chunks of information at once. When you reread a textbook, you’re not actually transferring information into long-term memory; you’re just keeping it in working memory for a few more seconds. Self-testing and elaboration force you to retrieve and re-organize that information, which is what actually builds durable memory. Sweller’s 2023 retrospective analysis of CLT’s evolution noted that the theory grew stronger precisely because of replication failures — when studies failed to find expected effects, researchers had to refine the model rather than abandon it. That’s a useful reminder: even the best theories have boundary conditions.

The nuance most people miss

The Study Smart study is promising, but it has limitations that matter for anyone trying to apply it. First, the “50% reduction in the performance gap” compares the intervention cohort to a previous year’s cohort, not a randomized control group. That’s a quasi-experimental design, not a true experiment. The researchers themselves note that unmeasured variables — like differences in course difficulty or student motivation across years — could account for part of the effect.

Second, the intervention wasn’t just the workshop. It included targeted coaching for the bottom 20% of students. That means we don’t know how much of the effect came from the workshop alone versus the personalized follow-up. If you’re a self-experimenter working alone, you can’t replicate the coaching component unless you hire a tutor or find an accountability partner.

Third, the strategies that work for academic learning (memorizing pharmacology facts for a multiple-choice exam) may not transfer to skill-based learning (playing an instrument, learning a language, or mastering a physical sport). The cognitive load principles apply broadly, but the specific tactics — like interleaving math problems — are optimized for declarative knowledge, not procedural skills.

Finally, the study measured self-reported study behavior, which is notoriously unreliable. Students said they changed their habits, but we don’t have objective data on how much time they actually spent on each strategy. The performance improvement is real, but the mechanism is inferred, not directly observed.

Practical implications

Test yourself, don’t re-read. The most consistent finding across cognitive science is that retrieval practice (quizzing yourself) produces far better long-term retention than re-reading or highlighting. In the Study Smart program, students were taught to replace re-reading with self-testing. For your own learning, spend at least 50% of your study time on active recall — covering up the answer and forcing yourself to retrieve it. Use flashcards, write down everything you remember from a chapter before looking at it again, or explain the concept out loud without notes.

Space your practice over days, not hours. Distributed practice — studying the same material across multiple sessions separated by at least 24 hours — consistently outperforms cramming. The Study Smart workshop emphasized this explicitly. For a personal experiment, schedule three 30-minute sessions spread across a week rather than one 90-minute session. The spacing effect is one of the most robust findings in memory research, with effect sizes that routinely exceed 0.5 standard deviations in meta-analyses.

Mix up your topics. Interleaving — studying multiple related topics in the same session rather than blocking them one at a time — feels harder and more confusing in the moment, but it produces better long-term learning. In the Study Smart program, students were taught to mix pharmacology topics rather than studying one drug class at a time. If you’re learning a new skill, alternate between related sub-skills (e.g., practice verb conjugations, then vocabulary, then sentence structure in the same session) rather than mastering one before moving to the next.

Design your own experiment

Here’s a concrete n=1 protocol to test whether these strategies work for you — because population averages don’t guarantee individual results.

What to test: Replace your current study or practice routine with a structured protocol that combines retrieval practice, distributed practice, and interleaving. Specifically: for any topic you’re learning, spend 50% of your time on active recall (no notes, just retrieval), schedule three sessions per week with at least 24 hours between sessions, and within each session, alternate between at least three related sub-topics.

How long to run it: Minimum 4 weeks. The spacing effect and retrieval practice benefits compound over time, so a shorter trial won’t capture the full effect. If you’re preparing for a specific exam or performance, run the experiment for the full preparation period.

What to measure: Pick one objective metric. If you’re studying for a test, use practice exam scores (not self-reported confidence). If you’re learning a language, use a standardized vocabulary test or a timed translation task. If you’re learning a physical skill, use a performance metric like accuracy rate or completion time. Measure your baseline before the experiment starts, then test weekly.

What confound to watch for: The biggest confound is total study time. If you unconsciously increase your study hours during the intervention week, any improvement could be due to more time, not better strategy. Keep total study time constant across the baseline and intervention periods. Also watch for the “novelty effect” — the first week of any new method often shows a boost simply because you’re paying more attention. That’s why you need at least 4 weeks: the novelty wears off by week 3.

What a positive result looks like: A consistent improvement of at least 15–20% on your chosen metric compared to your baseline, with the gap widening in weeks 3 and 4 as the spacing effect accumulates. If you see no improvement by week 4, either the strategies don’t work for your specific learning domain, or you’re not implementing them correctly (e.g., you’re “self-testing” by looking at the answer too quickly, or your interleaving is too random to be useful). In that case, try a more structured version: use a spaced repetition app like Anki for retrieval practice, and schedule your interleaving explicitly (e.g., 10 minutes on topic A, 10 on B, 10 on C, repeat).