IQ classification

Abstract: Are We Getting Smarter? Rising IQ in the Twenty First Century James Flynn Cambridge University Press. 2012Political scientist James Flynn's 'Flynn Effect' has potentially huge implications. Highlighted by Flynn in 1984, the Flynn Effect refers to Flynn's observation that scores on Intelligence (IQ) tests in Western countries have been rising since at least the 1930s. On assorted different tests, when a new cohort has taken the test, they have scored significantly higher than the previous cohort. Thus, for the average IQ to remain at 100 (the benchmark by which IQ comparisons are made) the scoring of the test had to be adjusted. Flynn found that Western IQs have increased by an average of 3 points per decade since the 1930s. The biggest gains are on 'shapes' (as opposed to arithmetic or verbal reasoning), usually considered the most 'culture free' part of the test. This is the part of the test believed to be closest to g, the general intelligence factor on which verbal, non-verbal and spatial intelligence are thought to develop, and which is responsible for the positive correlations among these abilities. The implications of Flynn's finding, explored in this book, are dramatic. Modern society may be making us more intelligent. But more likely, because the gains are only dramatic on shapes and IQ test measures things other than intelligence, it is making us better able to think in an abstract way. This, as Flynn argues, should be a paralleled by more complex TV programmes, more sophisticated political debate, and, simply, an ever more complex society. And, he notes, if we're not aware of then we may even end up executing mentally retarded criminals due to their IQ being established on an out-of-date test.In this new book, Flynn updates the body of research presented in his earlier book What is Intelligence? (Flynn, 2007), provides substantial appendices with all the relevant data, presents the nature of the Flynn effect in a clear manner with the educated layman in mind, and examines some sensational implications of failing to take the effect seriously. Most importantly, US states will not execute murderers if they are found to be 'retarded' (with an IQ of below 70). But if the prisoner's IQ test, whenever he took it, had not been revised for a long time then he might come out with an IQ higher than 70, even if his IQ were lower by today's norms, and so get executed. Flynn provides some surprising examples on the way in which the US justice system has executed people on the basis of higher IQ scores gained on highly questionable tests.But, in general, there are serious problems with this book. Flynn ignores evidence that contradicts his theories. He cites the Dutch scholar Jelte Wicherts when discussing 'factorial invariance' of IQ tests. Flynn explains that this term refers to the way in which the three main forms of intelligence (verbal, non-verbal, spatial) have not risen at similar rates, as spatial (shapes) has risen much faster. But this is not what Wicherts argued. Wicherts claims that specific IQ tests given to different cohorts are not 'factorially invariant' over time, meaning they are not measuring exactly the same thing, and latent variables may explain the secular gains. This has the potential to render the tests incomparable and undermine the Flynn effect (Wicherts et al., 2004). Later, Flynn claims that there is no evidence that black-white differences in intelligence are genetic, but completely ignores interracial adoption studies (e.g. Weinberg et al., 1992; Frydman & Lynn, 1989; Winnick et al., 1975) and probability analyses (Levin, 2005, pp.125-126) strongly implying that they are.For Flynn, 'intelligence' is not simply 'what IQ tests measure.' Intelligence is a general capacity to solve problems and absorb knowledge (p.7). IQ tests are very good at measuring this, but they are not perfect, meaning that other factors can contribute to a person's IQ score apart from their general intelligence. …

Abstract: The utility of the Satz-Mogel and Silverstein (two and four subtest) short forms of the WAIS-R was assessed on a group of 49 recently head-injured patients. The Satz-Mogel procedure yielded correct categorical classification of IQ in 92% of the cases and correlated highly with WAIS-R FSIQ (r = .97). Silverstein two-and four-subtest forms correlated highly with WAIS-R FSIQ (r = .82; r = .86) but resulted in a larger number of IQ classification errors. These data support the superiority of the Satz-Mogel procedure for abbreviating the WAIS-R with head-injured patients. Applications and limitations of short forms in the assessment of intelligence with head-injured patients are discussed.

Abstract: The completed WISC-Rs of 76 white urban children (6–0 to 15–9) were rescored utilizing the Satz and Mogel criteria for an abbreviated intelligence measure. Extremely high correlations for IQs (.96 – .98) and subtests (.66 – .95) were found. However, when mean differences between complete WISC-R and the shortened form were examined, significant differences between administrations were found. Furthermore, one-third of the subjects showed changes in intelligence classification levels when the abbreviated form was used. Thus, two of the three criteria previously suggested for a valid abbreviated intelligence test of: (a) a significantly high correlation between administration forms; (b) nonsignificant t-tests between the abbreviated and standard form mean IQ; and (c) low percentage of IQ classification change with the administration of the short form, were not met. It was concluded, however, that the abbreviated WISC-R may be appropriate when intelligence is a question relative to candidacy for therapy or as a noncritical, general indication of intelligence when IQ classification is not important and/or assessment time is limited.

Abstract: Intelligence study is one of keystone to distinguish individual differences in cognitive psychology. Conventional psychometric tests are limited in terms of assessment time, and existence of biasness issues. Apart from that, there is still lack in knowledge to classify IQ based on EEG signals and intelligent signal processing (ISP) technique. ISP purpose is to extract as much information as possible from signal and noise data using learning and/or other smart techniques. Therefore, as a first attempt in classifying IQ feature via scientific approach, it is important to identify a relevant technique with prominent paradigm that is suitable for this area of application. Thus, this article reviews several ISP approaches to provide consolidated source of information. This in particular focuses on prominent paradigm that suitable for pattern classification in biomedical area. The review leads to selection of ANN since it has been widely implemented for pattern classification in biomedical engineering.