Implicit Association Test (IAT)

Description

Measures the relative strength of automatic associations between concepts (e.g. a brand and 'trustworthy') from response-time differences in a categorisation task.

In the IAT, participants rapidly sort words or images into categories using two keys. When concepts that are strongly associated share a key, responses are faster. The D-score, computed from latency differences between compatible and incompatible blocks, indexes relative association strength. Brief IAT and single-category IAT variants shorten the procedure. The IAT has real limitations: test-retest reliability around 0.5 and modest prediction of individual behaviour, so it is best used for group-level comparisons. For UX and brand research it offers a measure less susceptible to self-presentation than stated attitudes, useful where users may not admit or know their associations.

Attributes

Loss type
Unmigrated
Effort
Days

Steps

  1. Define two target concepts (e.g. our app vs competitor) and two attribute categories (e.g. safe vs risky), with 4-6 stimuli each.
  2. Implement the standard 7-block procedure or a Brief IAT using validated software.
  3. Counterbalance block order across participants.
  4. Compute D-scores using the Greenwald, Nosek and Banaji (2003) improved algorithm, excluding trials over 10 s and participants with excessive fast responses.
  5. Compare group-level D-scores across conditions or segments, and correlate with explicit measures.
  6. Interpret cautiously: report reliability and avoid individual-level claims.

Sources

  • Greenwald, McGhee & Schwartz, Measuring individual differences in implicit cognition: The implicit association test (1998), JPSP
  • Greenwald, Nosek & Banaji, Understanding and using the IAT: I. An improved scoring algorithm (2003), JPSP

Also known as

  • IAT
  • Brief IAT
  • Implicit attitude measurement

Categories

  • User Research
  • Cognitive Psychology & Perception