Intelligence Testing: Uses & Limits
Intelligence Testing: Uses & Limits
Origins & Development
- Attempts to measure human intelligence in Western tradition date back to the mid-1800s.
- Originally used for military purposes and by eugenicists.
- Alfred Binet developed the first ‘IQ’ tests designed to identify school children struggling academically.
- IQ tests quantify intellectual ability by comparing scores against a cohort's expected standards (e.g., children of the same age).
- The original IQ score was a standardized score indicating performance relative to cohort expectations.
- A higher score indicates exceeding expected standards while a lower score indicates underperformance.
Initial Applications
- Primarily used in education.
- Identifying students with low ability for remedial help.
- Identifying students with high ability for advanced learning opportunities.
- Modern neuropsychological tests can target specific cognitive impairments.
What is Measured
- Intelligence tests for adults do not depend on developmental benchmarks but rely on comparisons to cohorts.
- IQ tests measure individual differences rather than an absolute amount of intelligence.
Standardisation & Ranking
- Intelligence tests are viewed as tools to rank performance on intellectually demanding tasks.
- Rankings are often expressed in percentile scores, indicating the percentage of individuals scoring higher or lower.
- Example: University admissions scores (e.g., ATAR).
Standardised Scores
- Standardized intelligence tests convert ranking into scores with a fixed reference point.
- A common IQ scale sets average at 100, with a standard deviation of 15.
- A Z-score sets the average at 0 and standard deviation at 1.
Standardised Distribution
- Standardized intelligence scores typically follow a Gaussian (normal) distribution, also known as the Bell Curve.
- Most values cluster around the average, with extreme scores being less common.
Application of Measurement
- Intelligence is a useful but imperfect predictor of life outcomes, explaining about 60% of variations like academic performance.
- Intelligence tests come with a degree of measurement error.
Limitations, Assumptions & Harms
- Intelligence testing has several historical problems categorized into three areas:
- Validity issues with tests.
- Interpretation problems of test results.
- Social and policy implications due to misinterpretation of differences.
Reliability & Validity
- Quality of measurement tools is assessed via Reliability & Validity:
- Reliability: Consistency in results across multiple tests.
- Validity: Whether the test measures what it purports to measure.
- An intelligence test is valid only when the design aligns with the characteristics of the test-takers.
Reliability & Validity Illustrated
- A Reliable intelligence test yields similar scores for the same person across attempts.
- A Valid intelligence test should measure intelligence, not cultural familiarity or socioeconomic status.
Contextual Assumptions
- Assumptions in intelligence testing can undermine validity if mismatched with test-takers’ characteristics.
- Many tests may reinforce existing social hierarchies based on race, gender, and culture, comparing averages from groups for whom the test was designed.
Common Assumptions
- Intelligence tests usually assume:
- Familiarity with the test language.
- Knowledge of specific cultural references.
- Experience with written problem-solving formats.
- Motivation to score well and a suitable mental state during testing.
- Knowledge of expected answer formats.
Assumptions Beyond Language
- Even non-verbal tests carry assumptions about familiarity with test formats and cognitive patterns in processing questions.
Validity of Comparisons
- Invalid IQ test comparisons reinforce social hierarchies, marginalizing groups that perform poorly on tests designed for mainstream cultures.
- Meaningful IQ comparisons require inclusive standardization for the entire group under comparison, which is infrequently achieved.
Conclusion
- IQ tests primarily rank individuals against an average standard.
- Standardized scores are adjusted based on reference groups.
- The assumptions underlying tests can lead to invalid scores or comparisons.