Encoding Strategies: How to Get Information Into Memory

Encoding strategies are ways of processing new information so it gets stored in long-term memory in a form you can find again. The rule behind almost all of them is simple: you remember what you think about. If you study a page by noticing its words, you store words. If you study it by asking what it means, why it's true, and how it connects to what you know, you store meaning, and meaning lasts.

This is why two students can spend the same hour on the same chapter and walk away with very different memories. This guide explains how memory encoding works, the levels of processing research behind it, its main limit, and eight practical encoding strategies, with a 30-minute routine to try on your next chapter.

What is memory encoding?

Memory has three stages. Encoding is getting information in, storage is keeping it over time, and retrieval is getting it back out. Most study advice targets retrieval, but a memory that was encoded weakly has little to retrieve.

Encoding needs attention. In a 1996 set of experiments, Fergus Craik and colleagues had people learn word lists while doing a second task. Splitting attention while learning caused large drops in memory. Splitting attention while recalling caused small or no drops. In plain terms: the moment you first meet new material is the moment distractions cost you most.

Psychology classes usually name three types of encoding:

Semantic encoding produces the strongest memories, and the classic experiments below show by how much.

Levels of processing: Craik and Lockhart's idea

In 1972, Fergus Craik and Robert Lockhart proposed the levels of processing framework. Instead of treating memory as a set of boxes (short-term, long-term), they argued that how well you remember something depends on how deeply you processed it. Shallow processing looks at surface features. Deep processing deals with meaning and connections. Deeper processing leaves a more durable memory.

The framework was deliberately loose, and critics pointed out that "depth" was hard to measure on its own. But it changed the question researchers asked. Instead of "how long did you study?", the question became "what were you doing with the information?" Ask the same question about your own sessions: two hours of rereading is two hours of shallow processing, and the clock won't show it.

The Craik and Tulving experiments

In 1975, Craik and Endel Tulving tested the idea directly. People saw a series of words, one at a time, and answered a quick question about each. The study was described as a perception and reaction-time task, so they didn't know a memory test was coming. The question type controlled the depth of processing:

• Case: "Is the word in capital letters?"
• Rhyme: "Does the word rhyme with ___?"
• Category: "Is the word a type of ___?"
• Sentence: "Would the word fit in this sentence: ___?"

Then came a surprise recognition test.

The meaning questions took only about 150 to 200 milliseconds longer to answer, yet recognition jumped from below 20% to over 90%. Experiment 2 used a harder test (the 60 studied words mixed in with 120 new ones) and found the same pattern: for words answered "yes," recognition rose from 15% after case questions to 81% after sentence questions.

The most important result for students came in Experiment 5. The researchers made the shallow task slow: people had to decide whether a word matched a pattern of consonants and vowels, which took about 1.7 seconds, twice as long as the sentence question. More time didn't help. Words from the quick sentence task were still recognized better: 82% versus 57% for the slow shallow task (for "yes" answers, Table 4). The kind of thinking mattered, not the time spent.

Also notice that "yes" answers were remembered better than "no" answers. A likely reason: a word that fits the sentence links up with its meaning, while a word that doesn't fit has less to attach to. That's an early hint of why connecting new material to what you know is so powerful.

What to do with this: for each new term, don't stop at reading the definition. Use it in a sentence of your own and say which category it belongs to. Each takes a few seconds, and they are the study versions of Craik and Tulving's two deepest questions.

Is deeper processing always better?

Not quite, and this matters for exam prep. In 1977, Morris, Bransford, and Franks had people study words by meaning or by rhyme, then tested them in one of two ways. On a standard recognition test, meaning-based study won, as expected. On a test that asked which new words rhymed with studied words, rhyme-based study won.

This is called transfer-appropriate processing: memory is best when the processing you do while studying matches the processing the test requires. For you, it means:

• If your exam asks you to apply concepts to new cases, practice applying concepts to new cases.
• If you'll need to pronounce vocabulary in an oral exam, say it aloud while studying.
• If you'll solve problems under time pressure, solve problems, don't just reread solutions.

Meaning-based encoding is the best default, because most exams test meaning. But the final weeks before an exam should look like the exam. The same matching idea applies, more weakly, to your surroundings: see whether studying in the same place helps.

8 encoding strategies that work

Each strategy below forces semantic or elaborative encoding, meaning you connect the new idea to things you already know. Two of them, elaborative interrogation and self-explanation, were rated "moderate utility" in a 2013 review of study techniques by John Dunlosky and colleagues.

A few notes on the evidence:

• Self-reference. A 1997 meta-analysis by Cynthia Symons and Blair Johnson confirmed that relating material to yourself improves memory, probably because the self is a well-organized body of knowledge that prompts both elaboration and organization.
• Saying it aloud. Colin MacLeod's team found that reading words aloud improves memory compared with reading silently, but mainly when you do it for some items and not others. Saying a word aloud makes it stand out. Read everything aloud and nothing stands out.
• Generating and explaining. Our guides to the generation effect, elaborative interrogation, and dual coding cover the research and limits of each.

What weak encoding looks like

You can usually spot shallow encoding by how it feels: easy, smooth, and forgettable.

• Rereading a chapter you've already read. The words feel familiar, but you're processing text, not ideas. See active recall vs passive reading.
• Highlighting without deciding why a line matters. Our post on whether highlighting is effective explains why it rarely helps on its own.
• Copying slides word for word in lectures. Paraphrasing forces you to process meaning.
• Studying with notifications on. Divided attention hits encoding hardest, as Craik's 1996 experiments showed.

If you want the bigger picture of why surface study fails, deep vs surface learning compares the two approaches across a whole course.

A 30-minute encoding routine for one textbook section

Use this on one section of about 5 to 10 pages.

1. Preview (3 minutes). Turn the headings into questions. Try to answer two of them before you read, even if you're guessing. Wrong guesses still prime you to notice the right answer.
2. Read one chunk at a time (12 minutes). After each subsection, stop and answer: what's the main claim, why is it true, and how does it connect to the previous subsection?
3. Make it yours (5 minutes). Write one example from your own life or another course for the hardest idea. Sketch one diagram.
4. Organize (4 minutes). Build a short outline with no more than four or five top-level points, each with its supporting details.
5. Close the book and recall (6 minutes). Write everything you remember on a blank page, then check and mark gaps in a different color.

Step 5 is where encoding turns into a durable memory. Retrieval doesn't just measure learning, it strengthens it, so active recall is the natural next step after any encoding strategy. Come back to the same section after a day and again after a few days.