Savings (relearning) method

Description

Measure the time or trials needed to relearn material after a delay and express the reduction against original learning as percent savings, revealing memory that recall misses.

Ebbinghaus, acting as his own participant, learned lists of 13 nonsense syllables at a constant rate, recorded the time to first learn each list, waited a set delay, then recorded the time to relearn it. Savings equals original learning time minus relearning time (for example 1,000 s minus 400 s gives 600 s), usually expressed as a percentage of original time. Plotting percent savings against delays from 19 minutes to 31 days gives the savings curve: memory falls quickly over the first two days and then levels off. Relearning can be counted in trials rather than seconds. Its value is sensitivity: savings detect residual learning even when free recall is zero. For UX it suits measuring how much of a learned interface survives a break, for example in infrequently used enterprise or safety-critical tools.

Attributes

Loss type
Unmigrated
Effort
Weeks

Steps

  1. Define a learning criterion, for example two error-free completions of a task or list.
  2. Measure time or trials to reach the criterion in the first session.
  3. Wait a planned retention interval; use separate groups for several intervals (for example 1 day, 1 week, 1 month).
  4. Measure time or trials to reach the same criterion again.
  5. Compute percent savings as (original minus relearning) divided by original, times 100.
  6. Plot savings against delay and compare curves across designs or user groups.

Sources

  • Cognitive Psychology (Goldstein, 2019)
  • Cognitive Psychology (Sternberg & Sternberg, 2012)

Also known as

  • Savings method
  • Ebbinghaus relearning
  • Relearning (savings)
  • Savings curve

Categories

  • User Research
  • Cognitive Psychology & Perception