Best Study Methods

Study advice is a crowded market, and most of it is sold on vibes. Color-coded notes look organized. Re-reading feels thorough. Highlighting feels like engagement. None of those feelings predict what you will remember in three weeks, and the research on this is awkwardly clear: the techniques students use most are among the weakest, and the strongest ones feel worse while you do them.

The best single reference here is Dunlosky et al. (2013), a review that graded ten common study techniques on the evidence behind them. Two came out with a high-utility rating: practice testing and distributed practice. A few earned a cautious middle grade. The student favorites, highlighting, re-reading, and summarizing, landed at the bottom. The top two combine into a single habit, which is the subject of our pillar guide on spaced repetition, but the full ranking is worth walking through, because knowing why the losers lose will save you hundreds of hours.

Here is the field, ranked, with what each method is, what the evidence says, and how to use it tonight.

Top tier: practice testing

What it is. Retrieving information from memory instead of looking at it: flashcards answered before flipping, past papers done closed-book, blank-page recall, problems re-solved from scratch. Known in the literature as retrieval practice or active recall.

The evidence. The deepest evidence base of any technique. Roediger and Karpicke (2006) showed students who practiced recalling a passage remembered far more a week later than students who re-read it repeatedly. Karpicke and Blunt (2011) found retrieval practice beat even elaborate concept mapping, including on inference questions. Dunlosky’s review rated it high utility across ages, materials, and test formats.

Tonight. Take today’s lecture, close the notes, and write everything you remember on a blank page, then check. That single exercise, the blurting method, is practice testing with zero setup. For repeatable testing, turn your notes into question form once and reuse them; a flashcard generator that builds cards from your own notes does the conversion in minutes.

Top tier: distributed practice

What it is. Spreading study of the same material across multiple sessions separated by days, instead of massing it into one sitting. Same hours, different placement.

The evidence. One of the oldest findings in psychology, visible in Ebbinghaus’s 1885 data and confirmed at scale by Cepeda et al. (2006) across hundreds of comparisons: spaced study reliably beats massed study for retention. The mechanism runs through the forgetting curve: each spaced review catches a memory mid-decay and resets it onto a flatter trajectory.

Tonight. Stop planning study in subjects-per-day and start planning in returns-per-topic. Whatever you learn tonight, book a ten-minute revisit tomorrow and another this weekend. Three short touches outperform one long one.

The two top-tier methods compound: testing yourself on a spaced schedule is the single highest-yield habit in the literature. That combination, plus pacing it to an exam date, is exactly the loop StudyDone automates from your own notes and slides.

Mid tier: interleaving

What it is. Mixing related topics or problem types within a session (ABCABC) instead of finishing one before starting the next (AAABBB).

The evidence. Strongest for math and physics problem sets, where the real exam skill is identifying which method a problem needs, something blocked practice never trains because the chapter heading gives the answer away. Rated moderate by Dunlosky et al., promising but less universal than the top tier.

Tonight. Doing twenty integration problems? Shuffle in some differentiation and series problems so every question forces a “what kind of problem is this?” decision. Expect to feel worse and score better later; that mismatch is Bjork’s desirable difficulty signature.

Mid tier: self-explanation

What it is. Pausing while studying to explain to yourself why a step works, how a new idea connects to what you know, or why an answer is correct, rather than just confirming that it is.

The evidence. Moderate utility in Dunlosky’s review, with good results for problem-solving domains. It builds the connected understanding that makes facts easier to retain and retrieve.

Tonight. After each worked example or dense paragraph, answer aloud: “Why does this step follow?” The disciplined, four-step version of this idea is the Feynman technique; use it on the two or three concepts in each course that everything else hangs on.

Mid tier: elaborative interrogation

What it is. Asking “why would this be true?” of factual statements and generating the answer yourself. Why do beta blockers worsen asthma? Why did inflation follow that policy?

The evidence. Moderate. Works best when you already have some background knowledge to connect the new fact to; weak when you are a complete novice in the area.

Tonight. As you review a fact list, force a one-sentence “because…” onto each item before moving on. Facts with reasons attached outlive facts without them.

Low tier: highlighting and underlining

What it is. You know what it is. You have the four dried-out highlighters to prove it.

The evidence. Low utility. Marking text requires almost no processing of its meaning, and in some studies heavy highlighting did no better than reading without a marker. Its real function is emotional: visible proof that studying occurred.

Salvage. Highlighting is acceptable as triage, a first pass marking what deserves to become a question. The learning happens later, when you test yourself on what you marked. A highlight that never becomes a retrieval attempt is decoration.

Low tier: re-reading

What it is. Reading the chapter again. The most popular study technique among students, by survey after survey.

The evidence. Low utility. The second read produces fluency, the text looks familiar, which students misread as knowing it. Roediger and Karpicke’s re-reading groups felt confident and performed worst a week out. Recognition is cheap; exams demand production.

Salvage. One careful read for understanding is necessary. It’s the third and fourth reads that should be replaced with retrieval, because by then you are skimming words your eyes already trust.

Low tier: summarizing without recall

What it is. Writing condensed versions of the material while looking at it. Copying the textbook’s sentences with fewer words.

The evidence. Low utility as typically practiced, because most students summarize with the source open, which makes it transcription. Summarizing well is a skill, and even done well it underperforms testing.

Salvage. Close the book first. A summary written from memory is retrieval practice wearing a summary’s clothes, and it instantly shows you what never made it into your head.

The rest of the field

Dunlosky’s review also graded two techniques students ask about constantly. Keyword mnemonics, inventing a vivid image linking a new term to a familiar word, earned a low rating overall: genuinely handy for stubborn arbitrary pairings like foreign vocabulary, but the images take effort to build and can fade faster than plainly learned material. Imagery while reading, picturing the content of each paragraph, fared similarly: pleasant, occasionally helpful for concrete texts, nowhere near the top tier. The pattern across both is the same one running through this whole ranking. Techniques that decorate the encoding step keep underperforming techniques that practice the retrieval step.

How to actually combine them

Rankings are tidy; semesters are not. Here is the working order for a real course, say a six-week pharmacology block:

1. First contact: understand. One careful read or lecture, using self-explanation on anything mechanism-shaped. No highlighter heroics.
2. Same day: convert. Turn the material into questions and problems. This is the last time passive review is allowed to feel like studying.
3. Every day after: retrieve on a spaced schedule. Short daily sessions, misses returning sooner, topics interleaved.
4. Final stretch: simulate. Timed past papers or generated quizzes under exam conditions.

For structuring the daily sessions themselves, time-boxing helps some students and hurts others; the honest breakdown is in our Pomodoro technique guide. But whatever container you pour the time into, the contents are what the research keeps voting for: test yourself, space it out, and treat the comfortable methods with suspicion.

FAQ

What is the single most effective study method according to research?

Practice testing, also called retrieval practice. Dunlosky et al. (2013) reviewed ten common techniques and rated practice testing and distributed practice as the only two with high utility across subjects, ages, and test formats. Combining the two, testing yourself on a spaced schedule, beats either alone.

Is highlighting actually bad for studying?

Highlighting is not harmful, just close to useless on its own. It marks text without processing it, so it earned a low-utility rating in the research. It is fine as a first-pass triage tool, provided you convert the highlights into questions you test yourself on later.

Why does re-reading feel effective if it isn't?

Re-reading creates perceptual fluency: the text looks familiar, so your brain mistakes recognition for knowledge. Exams demand production, not recognition, which is why fluency built by re-reading collapses under exam conditions while retrieval practice holds up.

How should I combine these methods in one study session?

Use understanding-focused techniques like self-explanation on first contact with material, then switch to retrieval for everything afterward. A solid session is a short read of new material, immediate self-testing on it, then spaced review of older material, with topics mixed rather than blocked.

Do these rankings apply to subjects like math and languages?

Yes, with format changes. In math, practice testing means re-solving problems from scratch and interleaving problem types. In languages, it means producing vocabulary and grammar from memory on a spaced schedule. The principles transfer even though the materials differ.