The Human Intelligence Podcast IQ Test Bias Explained: Myths, Mistakes, and Evidence with Craig Frisby
Sep 30, 2025
Dr. Craig Frisby, Emeritus faculty at the University of Missouri and author of Essentials of Evaluating Bias in Intelligence Testing, dives deep into the complexities of IQ test bias. He clarifies that bias stems from measurement errors, not mere score gaps. The discussion covers landmark cases like Larry P. that reshaped these debates, critiques of race-norming methods, and the necessity for standardized testing practices. Frisby emphasizes the importance of data-driven evaluations to ensure fairness in intelligence assessment across diverse demographics.
AI Snips
Chapters
Books
Transcript
Episode notes
Let Data, Not Belief, Guide Evaluation
- Use statistical tools rather than preconceived beliefs to evaluate tests.
- Let data determine whether items function equally across groups before judging bias.
Item Functioning Defines Bias
- Biased items often behave differently in difficulty or discrimination across groups.
- Proper items should show increasing pass probability with higher ability in every group.
Match Item Difficulty Order
- Check item rank-order of difficulty across groups and remove inconsistent items.
- Keep items that show the same easy-to-hard ordering for all groups.
