Paired-associate learning with imagery and generation manipulations

Description

Participants learn cue-target pairs under different encoding instructions, then recall targets from cues, to measure which encoding strategy produces better retention.

Participants study word pairs (boat-tree) and are later shown the first word and asked for the second. The encoding instruction is the manipulated variable. Bower and Winzenz (1970) presented 15 pairs for 5 seconds each; those told to form an interacting mental image recalled more than twice as many as those told to repeat the pair silently. Slamecka and Graf (1978) compared reading full pairs (king-crown) with generating the second word from a stem (king-cr___) and found 28 per cent more pairs remembered after generation, the generation effect. The design isolates encoding effects because the test is identical across conditions. For UX it offers a controlled way to test whether onboarding, training or labelling designs that make users generate or visualise associations improve retention of shortcuts, codes or terminology.

Attributes

Loss type
Unmigrated
Effort
Days

Steps

  1. Build a list of cue-target pairs representative of the material (for example 15 command-shortcut or term-meaning pairs).
  2. Assign participants randomly to encoding conditions (rote repetition, interactive imagery, generation from a stem).
  3. Present each pair for a fixed time, such as 5 seconds, counterbalancing pairs across conditions.
  4. Check instruction compliance with a brief strategy question after study.
  5. After a delay, show each cue and ask for the target; score exact or lenient matches.
  6. Compare proportion recalled across conditions and report effect sizes.

Sources

  • Cognitive Psychology (Goldstein, 2019)

Also known as

  • Paired-associate learning
  • Generation effect paradigm
  • Interactive imagery instruction

Categories

  • User Research
  • Cognitive Psychology & Perception