Expanding intervals

Judgement

Treat expanding intervals as one reasonable way to organise retrieval, not as the proven best pattern. The method begins with a retrieval likely to succeed, then places later retrievals progressively farther apart. Its proposed benefit is successful retrieval after increasingly difficult delays.¹

The timeline illustrates one possible expanding schedule, clarifying how retrievals can begin soon after learning and then occur after progressively longer delays.

<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 590 210" role="img" aria-labelledby="expanding-title" style="font-family: system-ui, sans-serif">
  <title id="expanding-title">One illustrative expanding retrieval schedule</title>
  <text x="24" y="30" font-size="18" fill="currentColor">One expanding schedule among many</text>
  <line x1="60" y1="120" x2="550" y2="120" stroke="currentColor" stroke-width="3" />
  <circle cx="80" cy="120" r="8" fill="currentColor" />
  <circle cx="155" cy="120" r="8" fill="currentColor" />
  <circle cx="295" cy="120" r="8" fill="currentColor" />
  <circle cx="515" cy="120" r="8" fill="currentColor" />
  <text x="58" y="94" font-size="16" fill="currentColor">Learn</text>
  <text x="130" y="154" font-size="16" fill="currentColor">Soon</text>
  <text x="270" y="94" font-size="16" fill="currentColor">Later</text>
  <text x="475" y="154" font-size="16" fill="currentColor">Later still</text>
  <text x="500" y="195" font-size="16" fill="currentColor">Time</text>
</svg>

Figure: Retrievals are spaced progressively farther apart, but this is one illustrative schedule among many. The early positive experiments were much narrower than a modern flashcard routine. Introductory psychology students learned fictitious names, with intervals measured in intervening list items and a final cued-recall test after thirty minutes. Expanding test-type practice beat comparable uniform schedules in that setting, but the study did not examine retention across days or months.¹

Broader evidence does not establish a general expanding advantage. A meta-analysis found no significant overall difference between expanding and uniform retrieval.² Another meta-analysis of second-language learning found expanding and equal spacing statistically equivalent overall.³ Replicated college experiments with GRE vocabulary pairs also found that an immediate first test helped after ten minutes, while equal spacing performed better after two days. Delaying the first retrieval improved the later result regardless of the subsequent pattern.⁴

With a deck and twenty minutes, expansion can be a starting procedure. Arrange an early successful retrieval, then widen later gaps while watching whether recall remains possible. For college students learning GRE vocabulary pairs, those experiments show that an immediate first retrieval should not be assumed best when the target is cued recall two days later. Adjust the pattern to the material and intended retention period rather than copying a named sequence.

Boundary

Spacing and retrieval practice are strong foundations. Choosing expanding rather than equal intervals is a weaker scheduling layer above them. Expansion is a reasonable procedure built on solid findings, not itself a generally superior finding. No experiment here establishes one expanding sequence across materials, learners and retention periods.