Revision

Revision is a loop of three moves: test, target, teach. A closed-book test finds gaps. A gap is anything the reader struggled with on recall, got wrong, or felt unsure about. Familiar material feels like known material, and the feeling does not close any gap. The gaps stay where they were until a closed-book test finds them.

The loop works across most subjects and levels. The reader does not need a large method menu to run it.

Core takeaways

• The loop has three moves. Test means producing an answer before checking the source. Target means starting with the weakest area. Teach means reframing the material for someone who has not learned it.
• A gap is anything the reader struggled with, got wrong, or felt unsure about on a closed-book attempt. Feeling unsure counts.
• Familiar material feels known. That feeling does not close any gap.
• Writing a question is itself retrieval, and it trains the reader for question forms never practised. Premade questions save time and find gaps the writer would not think to probe.
• The weakest area goes first, even when that is uncomfortable. People prefer the material that already feels fluent, and the loop requires giving that preference up.
• Finding no gaps is a warning. It usually means the technique was done wrong or tested the wrong levels. The repair is to use more kinds of retrieval.

Test: answer before checking the source

The first move applies the knowledge. The reader produces the answer before checking the source. Questions come from two places: questions written for the session, and premade questions already in circulation.

Writing a question is itself a retrieval event. Building a question teaches even when the writer cannot yet answer it. It also trains the reader for a curveball question, meaning a question in a form the reader never practised, built from pieces the reader already knows.

Premade questions save time. They can mirror the style of the assessment, and they expose gaps the writer would not have thought to probe. Matching the assessment’s style helps, but it is not a guarantee, because the paper is not bound to last year’s form. Relying too much on premade questions is one reason a curveball question comes as a shock.

Target: start with the weakest area

The gaps found do not all get equal time. The second move starts at the weakest slice. The areas of least confidence go first, even when that is uncomfortable. Revision that starts with the comfortable areas spends the hour where it is least needed. People prefer the option that already feels fluent. Working on the weak slice is uncomfortable, and the loop requires that discomfort.

Teach: reframe it for someone who has not learned it

The third move is teaching. It is not a repeat of what was just recalled. The reader reorders and reframes the material for someone who has not learned it. The strongest form of this says why a concept matters before it uses any terminology, as if the listener were a ten-year-old.

What counts as a gap, and what to do with it

A gap is anything struggled with on recall, got wrong, or felt unsure about. An incorrect answer is one kind of gap. Feeling unsure is another, and it counts. Every gap found gets re-encoded, using the methods that taught the material the first time, connected more deeply or from another angle. Higher Order Learning is where that re-encoding should aim.

When a session finds no gaps

Finding no gaps is a warning. Unless the knowledge is genuinely free of gaps, retrieval should surface some. Finding none usually means one of two things: the technique was done wrong, or it tested the wrong levels. The levels are the ladder in Knowledge Mastery From Recognition to Usable Knowledge. A session that felt too easy and found no gaps is the same signal: the practice missed some orders of learning or some types of knowledge. The repair is to widen the kinds of retrieval used. Doing more of the same kind does not find the missed levels. The menu for that widening is the Interleaving Table.

A session that finds no gaps does not count as a run of the loop.

Where this loop fits

The same loop overlaps the advanced group-study method. That protocol is on Group Study.

This loop is the first loop. The standing schedule is Spaced Interleaved Retrieval. The fuller method menu is the Interleaving Table, and the reader can run this loop before reading that table. Gaps found in this loop feed higher-order work and the mastery ladder.

How to practice this

1. Write a few questions for the session, and take a few premade ones. Notice that writing a question is a retrieval attempt even before you can answer it.
2. Answer each question with the source closed, then check the source. Note anything you struggled with, got wrong, or felt unsure about. Each note is a gap, and feeling unsure counts.
3. Spend the next part of the session on the area you feel least confident about. Notice the pull toward material that already feels fluent. The weak area goes first anyway.
4. Explain one gap to someone who has not learned it, as if to a ten-year-old. Say why the concept matters before using any terminology. Notice that the material has to be reordered to do this.
5. Re-encode each gap with the method that first taught it, connected more deeply or from another angle. Aim the re-encoding at Higher Order Learning.
6. If the session finds no gaps or feels too easy, treat that as a warning. Widen the kinds of retrieval using the Interleaving Table. Repeating the same kind of retrieval does not fix it.

Related pages

• Spaced Interleaved Retrieval: the standing schedule this loop leads into.
• Interleaving Table: the fuller method menu, which the reader does not need before running this loop.
• Higher Order Learning: where re-encoding a gap should aim.
• Knowledge Mastery From Recognition to Usable Knowledge: the levels a wrong order of testing misses.

Open questions

• Whether least-confident-first outperforms other allocations of the hour.

Sources

• Dunlosky, J., Rawson, K. A., Marsh, E. J., Nathan, M. J., & Willingham, D. T. (2013). Improving students’ learning with effective learning techniques: Promising directions from cognitive and educational psychology. Psychological Science in the Public Interest, 14(1), 4–58. Rereading rates low utility; practice testing rates high.
• Rowland, C. A. (2014). The effect of testing versus restudy on retention: A meta-analytic review of the testing effect. Psychological Bulletin, 140(6), 1432–1463.
• Kornell, N., & Bjork, R. A. (2008). Optimising self-regulated study: The benefits, and costs, of dropping flashcards. Memory, 16(2), 125–136. Fluency and familiarity are used as cues for knowing; people prefer the fluent option.
• Slamecka, N. J., & Graf, P. (1978). The generation effect: Delineation of a phenomenon. Journal of Experimental Psychology: Human Learning and Memory, 4(6), 592–604. Building the question is itself a generation.
• Fiorella, L., & Mayer, R. E. (2013). The relative benefits of learning by teaching and teaching expectancy. Contemporary Educational Psychology, 38(4), 281–288.
• Fiorella, L., & Mayer, R. E. (2014). Role of expectations and explanations in learning by teaching. Contemporary Educational Psychology, 39(2), 75–85. Preparing to teach, and teaching, beat preparing for a test.