Test-retest reliability and predictive validity of the Implicit Association Test in children

Area of Science:

• Developmental Psychology
• Cognitive Science
• Psychometric Evaluation

Background:

• The Implicit Association Test (IAT) is widely used in developmental research.
• Evidence for the reliability and predictive validity of children's IAT scores is limited and inconsistent.
• Previous studies varied in methodology, making it hard to identify factors influencing IAT reliability and validity.

Purpose of the Study:

• To investigate the test-retest reliability and predictive validity of children's IAT scores.
• To examine the influence of lag-time and domain on IAT psychometric properties.
• To provide recommendations for developmental researchers using the IAT.

Main Methods:

• Conducted five studies with 519 children aged 6–11 years.
• Manipulated lag-time between administrations and the domain of the IAT.
• Performed an internal meta-analysis using three different data analysis methods.

Main Results:

• Mean test-retest reliability effect sizes (rs = .48, .38, .34) and predictive validity effect sizes (rs = .46, .20, .10) were significantly greater than zero.
• Lag-time did not significantly moderate test-retest reliability.
• Domain differences in reliability and validity were influenced by scoring methods and sample characteristics.

Conclusions:

• Children's IAT scores demonstrate acceptable test-retest reliability and predictive validity in developmental research.
• Methodological choices, such as scoring and sample composition, impact IAT psychometric estimates.
• The IAT can be a valuable tool for developmental research when applied with careful consideration of methodological factors.