IQ is used synonymous and interchangeably with intelligence; however in this paper [pdf] Angela Duckworth et al argue that non-cognitive factors like test motivation also affect the IQ scores and have differential predictive validity.
Raven’s Progressive Matrices Example (Photo credit: Wikipedia)
- Intelligence, which is the ability to flexibly adapt to complex situations, is usually measured using IQ scores on intelligence tests. IQ scores however do not measure juts the raw intelligence; they also measure how motivated someone is to take the test and achieve a high score.
- Intelligence tests, that lead to IQ Scores, are supposed to measure the maximal intelligence ability that a person has and not the typical intelligence that he/she uses. In all intelligence testing it is assumed that the person will devote his entire attention and exert the maximum effort possible so as to achieve the highest score possible.
- While the assumption that IQ measures maximal intelligence may be true in high-stake testing situations, where the IQ results would be used for academic admissions, job placement or promotions; in normal measurement of IQ, say in a typical school setting, the stakes are quite low (there are no real/tangible repercussions of doing bad or well on the test) and hence IQ does not typically measure the maximal intelligence, but is confounded by test motivation.
- Test motivation refers to the fact that some people will be less motivated to take the test or continue with it and may display behaviors that indicate low motivation. While others may be highly motivated to take the IQ test. Thus, there would be individual differences at trait level on test motivation.
- Test motivation is also a state variable that can be manipulated by incentivizing getting high scores on the tests. When such incentives are in place, the IQ score should increase from the baseline level or when the test was given under non-incentivized conditions.
- Intelligence, as measured by IQ, has been associated with a number of good outcomes. Non cognitive factors as measured by test motivation are also theoretically linked to important life outcomes. For the purposes of this paper, two academic outcomes (years of education and academic achievement) and two non-academic outcomes (employment and criminal conviction) were measured and analyzed.
- The study 1 performed a meta-analysis of various independent samples where a comparison was made between the IQ scores received in standardized conditions vis-a-vis under incentivized conditions. For analysis the sample was divided in high IQ (those with IQ greater than 100) and low IQ (those with IQ less than 100). The main results were that incentives did result in higher IQ scores, the effect was stringer for low IQ group and there was dose-response effect with larger incentives leading to greater IQ points gains.
- Thus, for low IQ group, the lower IQ scores in standardized conditions could be due to lower intelligence or lower test motivation. If you increase the test motivation, you could bump up the IQ score of some of them. High IQ group, on the other hand had higher scores because they had both higher intelligence and higher test motivation.
- In study 2, a thin-slice video of children giving the intelligence test was behaviorally rated for signs of low test motivation. This was a longitudinal study and the IQ scores, test motivation and four types of outcomes were analyzed to find the differential impact of IQ/intelligence and test-motivation/ non-cognitive factors on life outcomes.
- The main finding was that test motivation had a significant impact, independent of IQ, on important life outcomes. This was specially pronounced for nonacademic outcomes like employment and criminal convictions. Intelligence as measured by IQ still had significant effect on all adult outcomes. They also found that test motivation predicted IQ scores, thus IQ score measures both intelligence and test motivation.
This is an important paper [pdf] that shows that IQ scores need to be interpreted with caution, and that both cognitive and non-cognitive factors are important for life outcomes.