Advertisement
Not a member of Pastebin yet?
Sign Up,
it unlocks many cool features!
- IQ checklist:
- - range restriction
- - selection effects (Berkson's paradox or unrepresentative groups like Mensa)
- - differing levels of measurement error:
- - comparing different constructs
- - tests of different lengths
- - sum scores vs factor scores
- - subtests vs full-scale tests
- - claims of IQ changes due to environment or intervention:
- - fadeout over time
- - failure of test invariance: post-test factor analysis shows gains are subtest-specific or on factors other than claimed
- - demand + motivational effects (non-blinded/passive control groups eg in dual n-back)
- - controlling for intermediate variables/'Everest regression': adding moderators/covariates which are correlated with IQ without acknowledging the (often dubious) causal model this corresponds to
- - group comparisons not adjusted for Flynn effect or Wilson effect (spurious gains/losses or differences in heritabilities)
- - lack of genetically-sensitive design while investigating causal questions
Advertisement
Add Comment
Please, Sign In to add comment
Advertisement