Retrieval practice as a study protocol: close the source, produce the answer, then check

Goal

Replace rereading and highlighting, which feel productive, with attempts to reproduce the material from memory, which expose what is not yet known. The 2007 IES practice guide recommends using quizzing with active retrieval "at all phases of the learning process"; the 2021 review by Agarwal, Nunes and Blunt, which systematically coded 50 classroom experiments, reports benefits of retrieval practice across education levels, content areas and test formats.

Prerequisites

A source split into units small enough to recall in a few minutes (a section, a function's contract, a command's options), a blank page or file, and a place to record gaps with a date for the next attempt.

Steps

1. Read or work through one unit once, attentively, without taking notes that copy the text.
2. Close the source. Write, from memory, what the unit said: definitions, steps, the reason each step exists, one example. Write "?" where a name or value does not come.
3. Reopen the source and compare line by line. Mark three categories: reproduced correctly, reproduced wrongly, missing.
4. Turn every wrong or missing item into a prompt with its answer. Wrong items get a note on what was confused with what.
5. Record the prompts with a date for the next attempt, at least a day later; a spaced-repetition schedule can hold them.
6. On the next attempt, answer the prompts first, then recall the whole unit again. Retire prompts only after they have been reproduced correctly on separate days.
7. Once a week, recall the structure of the whole source (chapter list, main claims) from memory and compare with the table of contents.

Expected result

A short list of prompts per unit that concentrates on exactly what did not stick, and a record of recall attempts that shows which units needed how many passes. The subjective sense of "I know this" is replaced by a count of successful reproductions.

Limits and test basis

Retrieval practice improves the retention of what is retrieved; it does not check whether the source was right or whether the understanding is correct in cases the prompts do not cover. Feedback matters: comparing against the source immediately after the attempt is part of the protocol, not optional. The cited review concerns school and university classrooms; transfer to self-directed adult study of technical material is plausible but not shown by it. No measurement is claimed here.