Retrieval Practice Grid

A simple grid with columns for this week, last week, and three weeks ago. Each week, learners attempt to fill each column from memory alone — no notebooks, no phones, no prompts. The deliberately uneven spacing is the point: recall gets harder for older material, which is exactly where retrieval produces the largest durable gains.

Procedure

1. Distribute a blank grid: three columns labelled This week, Last week, Three weeks ago. Rows may be left blank or seeded with thematic headings (environment, work, emotions).
2. Silent, individual, 5 minutes. Learners try to write as many target words or phrases as they can recall from each period, with a short example sentence or collocation if they can manage it.
3. At the 5-minute mark, learners fold the paper in half and swap with a partner.
4. Pairs compare and add any items they both agree belong but one has missed. New additions go in a different colour.
5. Whole-class consolidation (3 min): teacher elicits items per column, building a board list; learners update their grids.
6. Collect or photograph completed grids. They become the baseline for the following week.

Why it works

Roediger and Karpicke's (2006) meta-analytic finding that retrieval practice produces 50–80% long-term retention gains over re-study rests specifically on the kind of effortful recall this grid forces. Re-reading a vocabulary list feels easier during study but produces weaker traces than struggling to retrieve without cues. The distributed spacing across weeks leverages the spacing effect, which Kim & Webb's (2022) meta-analysis confirms is robust for L2 vocabulary. The middle column — last week's material — is the hardest: recent enough that re-study feels redundant, old enough that retrieval requires genuine work. That is the desirable-difficulty sweet spot.

Variations

• No-feedback version: Learners keep their grid without any correction. Evidence shows even unsupported retrieval improves retention (Karpicke & Roediger 2008).
• Image-cued grid: Each row is a small image; learners write the associated word or phrase. Adds dual-coding to the retrieval.
• Collaborative grid: Teams build a single grid together. Different retrievers catch different items; the pooling rewards distributed attention in the original lessons.

Tips

• Use the grid at the start of every third or fourth lesson, not every lesson. Overuse dulls the spacing effect.
• Resist the temptation to allow phones or notebooks. The effortful retrieval is the learning event; external aids remove what makes the activity worth running.
• Patterns of forgetting in the class grid are diagnostic. Words that disappear after one week need a different encoding strategy — consider feeding them into Frayer or keyword work.