Retrieval practice schedule for a teaching unit: four flexible weeks

Plan a retrieval practice schedule for a teaching unit by choosing which objectives to revisit, the new question for each check and the feedback that follows. The aim is not to fill four Fridays with identical tests. A good sequence introduces a concept, asks for it after a short gap, mixes it with other ideas and checks whether students can use it in a new situation. The interval and difficulty should fit the unit, so this four-week model is a template to adapt rather than a universal formula.

The What Works Clearinghouse practice guide recommends spacing learning and using quizzes for active retrieval. The Education Endowment Foundation’s metacognition guidance emphasises planning, monitoring and evaluating learning. Together, they suggest a cycle in which answers change teaching and students learn to judge what they can explain without looking.

Choose three objectives that can return in new forms

Imagine a science unit on ecosystems. Students should explain a food chain, predict a change when one population falls and read a simple data table. These are related but not interchangeable. A quiz that only asks for definitions can produce a high score while students still struggle to make a prediction. Write a small question bank with one recall item, one interpretation and one explanation per objective.

List common misconceptions in advance: arrows in a food chain may be read in the wrong direction; a population decline may be treated as a single direct effect; a table may be read without checking units. Design each distractor or explanation prompt to reveal one of these errors. The guide on diagnostic quizzes for prior knowledge helps you identify prerequisites before the unit begins.

This map prevents a schedule from becoming a calendar of repeated trivia. Keep the source material or answer rationale beside each question so the key can be checked before students see it.

Week 1: teach, then use a small exit check

After teaching the core idea, ask two or three questions without notes. One may be a familiar diagram and another a short explanation: “What does the arrow show?” Show feedback or discuss the answer promptly. If many students reverse the arrow, model the relationship with a new example before the next lesson. Do not assign a high-stakes grade to a first attempt that is intended to reveal the starting point.

At the end of the week, ask students what they can now explain and which step still feels uncertain. Their self-report is useful for planning, but compare it with an actual response. Some students feel confident because a diagram looks familiar; others understand but lack confidence. The guide to exit ticket questions that change tomorrow’s lesson shows how to turn those differences into an instructional choice.

Week 2: return after a gap and change the example

Do not begin by showing the original answer key. Present a new habitat and ask students to reconstruct a food chain from memory. Mix in one question about interpreting a table. The shift of context makes the learner retrieve the idea rather than recognise a picture they have already seen. If an answer is wrong, provide a brief explanation and a fresh attempt, not just the correct letter.

Keep the quiz short enough to review. A five-minute check with useful feedback is preferable to a long form whose results are never read. If the class splits between two answers, inspect whether the distractors represent a real misconception or an ambiguous question. Fix poor wording before treating the score as evidence about learning.

At this point students can keep a simple log: “I confused arrow direction; next time I will name who eats whom before drawing.” The note links the error to an action. It is more useful than recording a percentage without a plan.

Week 3: mix topics and require explanation

Interleave a question about food chains with a population-change case and a table. Ask students to explain one answer in a sentence. The written reason can show whether a correct choice came from a principle or a guess. It also reveals when an incorrect choice follows a reasonable interpretation of an unclear diagram. Read a sample of explanations and discuss common patterns with the class.

This week should not simply be harder for its own sake. Keep the learning objective recognisable while changing the surface example. If the new case adds unfamiliar vocabulary or an unreadable graph, the task may measure reading load rather than ecosystem understanding. Adjust the support and check access to the material.

Week 4: test transfer and choose the next revisit

Give a new, modestly complex ecosystem scenario. Students first answer a short retrieval question, then explain a prediction using evidence from a table. The longer response is essential if the unit goal is reasoning. A multiple-choice score alone cannot show whether a student can construct the explanation independently. Compare this task with the week-one objectives, not just with earlier percentages that may have different difficulty.

After review, separate “secure,” “needs another example” and “needs reteaching.” Schedule a later question for the concepts in the last two groups. A four-week plan need not end the retrieval cycle; it establishes a way to revisit material as later units begin. Tell students why an older idea has returned so they can see the pattern of learning over time.

Do not interpret every low score as forgetting. A changed question may contain a hidden prerequisite, or the original explanation may never have been clear. Ask students how they reasoned and check the item before deciding whether to reteach. The goal is better understanding, not a prettier score trend.

Four worked checkpoints for the same ecosystem objective

Put actual questions and keys in the timetable. This proposed sequence assumes that energy-transfer arrows, reading the population table and alternative prey have been taught before the relevant check. The numbers are invented. The elapsed days illustrate planning choices, not optimal memory intervals.

The gaps are three days from teaching to the first check, then seven days between checks. A timetable interruption can change them. If most learners cannot explain the first arrow even after feedback, do not wait a week merely to obey the grid: give a worked example and a supported check in the next lesson, then plan a later independent attempt.

Keep a record of support, not just a percentage

Use a small external log with objective, question version, response and support received. “Correct after the chain was displayed” differs from “reconstructed without notes”; both can be useful, but they answer different questions. The log does not need named data if you only need to adjust whole-class teaching.

Day 25 introduces an additional relationship and conditional reasoning. A lower percentage than day 4 therefore does not measure a simple amount of forgetting. Check the arrow objective separately with a comparable new chain before drawing that inference. Likewise, a higher score after a hint is not evidence that the interval caused improvement.

After day 25, choose a later revisit from the remaining difficulty: arrow direction, table reading or conditional prediction. Record why that objective returns and what would count as a sound explanation. This turns four dates into a sequence that can adapt to the answers.

Make the schedule practical in Skolina

Create one short quiz per checkpoint, using a different example each time. In Skolina, preview the answer keys and share the quiz. Advanced per-question reports require Starter, Pro or eligible institution access. Read open explanations yourself. Keep a small external note of which objective each item targets and what you changed after feedback. Check the Skolina plans if you expect to use several quizzes across classes.

To try the first cycle, create a Skolina teacher account. Start with three objectives and one five-minute check. Add the later checks to your teaching plan so retrieval happens after a real interval, with enough time to respond to the results.