WJEC A2 Psychology PY4 - Issues in the Measurement of Intelligence

• Binet believed intelligence is a dynamic process and not fixed, prompting the Frenchman in 1905 to develop first formal pen and paper measure of intelligence
• teamed up with Lewis Terman of Stanford Uni in the USA to create first IQ test which is still used today, known as the Stanford-Bient (SB) test in 1916, measuring subsets such as verbal, maths and visual reasoning and administered individually, standardised using white American children
• more recently there has been the Wechsler test (WAIS) in 1955 which has taken over as the most popular IQ test, measuring verbal and performance tasks, standardised on 2200 people in the USA
• both tests result in a single number known as your IQ (Intelligence Quotient)
• other tests also developed eg BAS, Raven's Matrices, STAT etc but none as widely as BS and WAIS
• however the use of these tests has raised a number of key issues

1. Use of IQ Tests

• ultimately why do we use IQ tests - what purpose do they serve? what do we do when we know our IQ, how can it be used?

(+) the original purpose was to reduce intellectual inequalities by identifying children lagging behind peers so they could be given extra help and support (Alfred Binet believed IQ was not determined or fixed from birth and could be developed). This was triggered by Binet's fears that children may be subjectively classified as intelligent by teachers, based on opinion rather than objective evidence

(-) however this original purpose changed and was used by Eugenicists to justify their treatment of those who they believed were 'feeble-minded' and risked 'contaminating' the human race, as they believed intelligence was fixed; wanted to treat those considered 'Morons' or 'Imbeciles' by IQ tests (and waste of invested money to improve intellects) by preventing immigration rights, enforced sterilisation, extermination eg

2. Internal Validity: are IQ tests measuring what they're supposed to be measuring?

Concurrent Validity (whether new tests of IQ are accurate, by comparing results of same people doing test with performance on existing valid tests)

(+) comparing WAIS and SB scores, concurrent validity is high - people tend to score similarly on both so person's results correlate with the other

(-) however concurrent validity with older traditional tests compared to modern tests (which measure more than 'g', like Sternberg's Triarchic Abilities Test: STAT) is low and has a weak correlation, raising question 'why'?, what sort of intelligence is being measured and do the tests have construct validity?

Construct Validity (a means of assessing the accuracy of a test by examining the extent to which performance on an IQ test measures an underlying theoretical belief)

(+) if one suggests intelligence is a single entity like Spearman's 'g' then such IQ tests have high construct validity

(-) BUT others questioned whether intelligence is more than 'g', such as Gardner and Sternberg, if one takes this view then tests like WAIS and SB could be argued to be measuring only one element of intelligence; some argue intelligence has been reified

Criterion/Predictive Validity (measuring the accuracy of IQ tests by comparing test scores with another current/future criterion like exam grades - if IQ is measuring something real, two measure should correlate)

(+) WAIS demonstrates a correlation of +0.5 with later school grades

(-) however, although test has some predictive validity, it cannot explain ALL exam grades (otherwise it would be +1), therefore suggests other factors must be involved

(-) Helen Bee (1994) argues it is impossible to measure potential or underlying ability, and IQ tests ignore factors such as motivation and effort

(?) however professionals (doctors/lawyers) have average IQ of 120 whereas semi-skilled occupations have 100, suggesting may predict type of job a person is likely to do later in life, but not necessarily how good they'll be at it

3. External Validity: concerns the extent to which the findings of a test/study can be generalised beyond the setting in which they were measured

Cultural Bias: argued that differences in IQ test scores between individuals within and between groups and subgroups may not reflect genuine intelligence differences but just differences in how measuring tool (IQ test) works in different places with diff people, eg today African American children in the USA score on average 10 points lower than white children on a range of IQ tests (although continuing to narrow) - is this genetic or due to cultural differences?

(?) Arthur Jensen (1980) differences are 80% genetic between black and white scores

(-) not all cultures (or sub-groups within a culture) agree that intellectual/theoretical ability is as important compared to those who constructed the test and use it most (North America + Europe); for example Sternberg (1993) found that amongst immigrants from Cambodia, Mexico, The Philippines and Anglo-Americans, all (bar Americans) rated motivation, social skills and practical school skills as just as, or more, important than cognitive skills and intelligence

(-) the early SB test was standardised using white people (children) only and the WAIS with Americans only; this is an ethnocentric way of standardising tests and perhaps reflects the norms (average IQ also) of only one cultural group

(-) the tests were designed by Western, middle class, educated individuals - many argue the questions on IQ tests contain references to vocabularies, objects, general knowledge bases and experiences far more familiar to Western, white, middle class, educated individuals