H567/02 Section A - Gould (1982): bias in IQ testing

Gould's review and Yerkes's original research

Gould's A nation of morons was a review article, adapted from his historical critique of intelligence measurement, rather than a new experiment conducted in 1982.

The original programme was led by Robert Yerkes during US involvement in the First World War. Group testing offered a quick way to classify large numbers of army recruits. Yerkes also wanted to demonstrate that psychology could produce objective scientific measurements and believed intelligence was inherited. Gould examined whether these assumptions, the tests and their administration justified the conclusions drawn.

The historical programme tested about 1.75 million US army recruits, including White Americans, Black Americans and European immigrants. This is the population of records reviewed, not a sample that Gould recruited. The recruits were men in a military context, not a representative sample of all people or national populations. The review uses historical documents, reported test procedures, quantitative summaries and analysis of their interpretation.

Psychometric testing seeks to measure psychological characteristics using standardised tasks and scores. A score can be precise yet measure the wrong thing. Gould's central challenge was that learned language, schooling and cultural knowledge affected tests presented as measures of innate ability. Evidence of a score difference cannot by itself decide between nature and nurture.

This belongs to individual differences because intelligence tests classify differences between people. Its theme is measuring differences: the validity, fairness and social consequences of that classification. The lesson's historical detail follows OCR's guide; the original Gould article was not available for inspection in this build.

Tests and problems in their administration

Yerkes's team developed group tests intended to assign recruits to levels of ability, with an individual examination intended for those who could not complete the group tests.

The team wrote the tests in 1917. Scores produced grades from A to E, including intermediate signs; a low grade could restrict the roles considered suitable. The procedure was not simply “everyone took the same fair test”. Some illiterate or immigrant recruits were put through Alpha despite needing Beta. Growing Beta queues led administrators to change allocation rules, and camps differed in their requirements. Some Alpha failures did not receive Beta; only one fifth of Beta failures received an individual examination, according to the reviewed account.

Two problems must be distinguished. Content bias arises when performance depends on culturally learned facts rather than the intended ability. Administration inconsistency arises when comparable recruits receive different tests or follow-up opportunities. The first threatens construct validity; the second can undermine both fair comparison and repeatability. Giving a non-verbal-looking test does not automatically remove cultural demands or problems understanding instructions.

Reported results and Gould's challenge

The historical analysis converted selected scores to a common scale and reported group differences that were then interpreted as inherited differences in intelligence.