Tethering the elephant: Capital cases, IQ, and the Flynn effect.

Capital offenders cannot be executed if they are mentally retarded. Therefore, the IQ scores of offenders are important, and the U.S. 4th Circuit Court of Appeals has held that the Flynn effect is relevant to interpreting their IQ scores. The Flynn effect (IQ gains over time) means that different IQ tests will give different scores purely as a result of when the tests were normed. Because execution must not be a random result of what test defendants take, a formula is provided to convert IQ scores to a common metric: the norms current at the time the test was taken. The formula also includes a correction based on evidence that the Wechsler Adult Intelligence Scale-Third Edition inflates IQs because of sampling error. Given the inevitability that opposing experts will offer conflicting diagnoses, IQ scores merit special attention in capital cases. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Adjusting IQ scores for the Flynn effect: Consistent with the standard of practice?

Should psychologists adjust obtained IQ scores to accommodate the Flynn effect (J. R. Flynn, 1985)? The authors surveyed directors of doctoral training programs approved by the American Psychological Association and board-certified school psychologists and completed a systematic review of IQ test manuals, contemporary textbooks on IQ testing, federally regulated IQ testing protocols, and various sources of legal and ethical guidance. They confirmed in each instance that such adjustments to IQ scores do not comport with prevailing standards of psychological practice. Results of IQ testing may be applied to a broad range of psycholegal issues, many of which cannot be anticipated. Psychologists assist examinees, courts, and other 3rd parties most effectively by administering and interpreting IQ tests in their intended fashion. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Science rather than advocacy when reporting IQ scores.

The existence of shifts in mean IQ scores over time is well established. However, on a case-by-case basis, such shifts vary unreliably, rendering specific adjustments to a given individual's IQ score incalculable. Based upon data presented previously (Hagan, Drogin, & Guilmette, 2008) as well as a review of more recent studies that have further detailed the wide variability of mean score shifts, any proposal to “correct” IQ scores in forensic evaluations due to the “Flynn effect” (FE) is unjustifiable. To offer the court an unreliable new IQ score in place of an allegedly unreliable old one—and to do so specifically in capital murder cases as opposed to any other context—appears far more reflective of result-focused advocacy than objective scientific practice. Forensic psychologists are explicitly encouraged to address likely ranges of IQ score variability and to discuss in relevant detail the strengths and weaknesses of the specific studies—however much at odds these may be—that attempt to define and quantify mean score shifts. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Expanding variance and the case of historical changes in IQ means: A critique of Dickens and Flynn (2001).

The Flynn effect is the rise in mean IQ scores during the 20th century, amounting to about 0.33 IQ points per year. Many theoretical explanations have been proposed, though none are universally accepted. W. Dickens and J. R. Flynn's (2001) new approach explains the large IQ changes by means of recursive models of IQ growth. A salient feature of their models is that IQ phenotypes and their supportive environments are correlated; in addition, environmental effects can rebound on phenotypic IQ to increase or lower IQ. In this critique, the authors examine an empirical challenge to their models, which typically imply large changes in IQ variance. However, the historical rise in IQ mean level has not been accompanied by substantial variance changes, a finding inconsistent with the properties of the proposed model. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

The Flynn Effect and U.S. Policies: The Impact of Rising IQ Scores on American Society Via Mental Retardation Diagnoses.

Over the last century, IQ scores have been steadily rising, a phenomenon dubbed the Flynn effect. Because of the Flynn effect, IQ tests are periodically renormed, making them harder. Given that eligibility for mental retardation (MR) services relies heavily on IQ scores, renormed tests could have a significant impact on MR placements. In longitudinal IQ records from 9 sites around the country, students in the borderline and mild MR range lost an average of 5.6 points when retested on a renormed test and were more likely to be classified MR compared with peers retested on the same test. The magnitude of the effect is large and affects national policies on education, social security, the death penalty, and the military. This paper reports the perceptions of professionals as they relate to IQ score fluctuations in normal, borderline, and/or MR populations. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

L'effet Flynn et la déficience intellectuelle.

In 1984, Flynn pointed out a 3 to 5 point increase per decade in major IQ tests in the United States. To respect the golden standard of the 100 point average and 15 point standard deviation, it became necessary to regularly restandardize the American IQ tests. In the first part of this paper we tackle three subjects: The observation of the "Flynn effect," its probable causes and its possible limit. In the second part we show that the "Flynn effect" might present a misdiagnosis risk for mental retardation when professionals use an obsolete version of an IQ test. This caveat especially applies given that the greatest risk of misdiagnosis is for borderline cases (with scores from 70 to 85). We also raise the possibility that a more complex environment due to the evolution of technologies might necessitate adjustment of the functional adaptability criteria for mental retardation. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Looking to science rather than convention in adjusting IQ scores when death is at issue.

The progressive obsolescence of IQ test norms and associated score inflation (i.e., the Flynn effect) may have literal life and death significance in capital mental retardation determinations (i.e., Atkins hearings). Hagan, Drogin, and Guilmette (2008) asserted that IQ score corrections for the Flynn effect were inconsistent with a “standard of practice” they deduced from custom, convention, and authority. More accurately, this reflected a proposed practice guideline or recommendation for practice, rather than a standard of practice. Whether a proposed guideline or recommendation for practice, these are better informed by an analysis of the available science than accepted convention. The authors reviewed research findings regarding the occurrence of the Flynn effect in the “zone of ambiguity” (IQ = 71–80), and proposed a best practice recommendation for discussing and reporting Flynn effect correction of IQ scores in capital mental retardation determinations. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

"Massive IQ gains in 14 nations: What IQ tests really measure": Correction to Flynn.

Reports an error in "Massive IQ gains in 14 nations: What IQ tests really measure" by James R. Flynn (Psychological Bulletin, 1987[Mar], Vol 101[2], 171-191). In this article, it was wrongly stated that that U. Schallberger put forward a hypothesis or hypotheses concerning the magnitude of Swiss IQ gains; in fact, he did not. Flynn wishes to acknowledge this error and stress that his critique of these hypotheses does not apply to Schallberger. Schallberger had earlier come to the same conclusion as Flynn, namely, that the Swiss have probably made sizable IQ gains, but all estimates of magnitude would have only speculative status. (The following abstract of the original article appeared in record 1987-17534-001.) Data from 14 nations reveal IQ gains ranging from 5 to 25 points in a single generation. Some of the largest gains occur on culturally reduced tests and tests of fluid intelligence. The Norwegian data show that a nation can make significant gains on a culturally reduced test while suffering losses on other tests. The Dutch data proved the existence of unknown environmental factors so potent that they account for 15 of the 20 points gained. The hypothesis that best fits the results is that IQ tests do not measure intelligence but rather a correlate with a weak causal link to intelligence. This hypothesis can also explain difficult trends on various mental tests, such as the combination of IQ gains and Scholastic Aptitude Test losses in the United States. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

The IQ paradox is still resolved: Reply to Loehlin (2002) and Rowe and Rodgers (2002).

In their original article (W. T. Dickens & J. R. Flynn, 2001), the authors formalized the consensus model of reciprocal effects between IQ and environment and showed that it can resolve the apparent paradox between high heritability and large environmental effects. Commentators suggested that the model has undesirable properties that call its usefulness into question. J. L. Loehlin (2002) argued that incorporating persistence of IQ into the model causes problematic behavior. D. C. Rowe and J. L. Rodgers (2002) argued that an increasing correlation of IQ and environment should have caused growing variance of IQ. Empirical evidence suggests that IQ is not sufficiently persistent to cause the problems Loehlin found and that the correlation of IQ and environment has not grown much over time so that the reciprocal effects model need not imply increasing variance. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Comparison of the WAIS and the WAIS—R in the upper ranges of IQ.

35 medical students were administered 1 of 2 combination forms of the Wechsler Adult Intelligence Scale (WAIS) and the WAIS—Revised (WAIS—R) in a single session, with the WAIS and WAIS—R components presented in a counterbalanced order using a procedure that avoided the repetition of identical items. The WAIS IQ scores were higher than were the WAIS—R IQ scores. The higher the WAIS Full Scale IQ, the smaller the Full Scale IQ difference between the 2 tests. Unexpectedly, the Verbal minus Performance discrepancy was greater on the WAIS—R than on the WAIS. (7 ref) (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Reinforcement and Binet Performance in Disadvantaged Children.

Test-taking motivation has been viewed as an important determinant of the attenuated IQ scores of disadvantaged children. This study compared the effect of a material reward (money) and the symbolic reward generally used (praise) in the test-taking situation on Binet IQ scores of severely disadvantaged fourth-grade children. Forty-six boys and 46 girls were administered the Stanford-Binet Intelligence Scale, Form L-M, with either material reward (money) or symbolic reward (praise) for correct responses. The IQ scores of the four groups, constituted on the basis of reward condition and sex, were compared. Results indicated no IQ difference due to reward and no interaction between reward and sex. Boys made significantly higher IQ scores than did girls. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Flynn effects not genetic and unrelated to race differences.

Comments on the article by J. R. Flynn (see record 1999-00167-001) which argued that massive IQ gains over time test the IQ-intelligence equation, reveal groups who achieve far beyond their mean IQ's and falsify prominent arguments for a genetic racial IQ gap. The author contends that Flynn fails to report in sufficient detail a summary of what the gains over time do and do not tell us about the nature of the Black–White difference. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Aging and vocabulary score: A meta-analysis.

Vocabulary scores were examined in a total of 210 articles, containing 324 independent pairings of younger and older adults, from the 1986-2001 issues of Psychology and Aging. The average effect size, favoring the old, was 0.80 SD. Production tests yielded smaller effects (0.68 SD) than multiple-choice tests (0.93 SD). Both age and education were found to be partially independent determinants of performance in production tests; age effects disappeared in multiple-choice tests as soon as education was taken into account. In addition, the Wechsler Adult Intelligence Scale-Revised Vocabulary subtest (D. Wechsler, 1981) was also found to be sensitive to the Flynn effect Q. R. Flynn, 1987; i.e., increasing test scores with advancing birth year). The results question the approach of using age-group equality in vocabulary scores as a check on sample equivalence. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Comparison of the WAIS and WAIS—R: Different results for different IQ groups.

Using a procedure that eliminated repetition of identical items, thus avoiding order effects, we administered the Wechsler Adult Intelligence Scale (WAIS) and the WAIS—Revised to 108 subjects. All correlations between the two tests were significant and similar to those reported in the WAIS—R manual. For the group as a whole, verbal, performance, and full scale IQ scores on the WAIS—R were significantly lower than their respective WAIS scores; however, this difference was not consistent across IQ levels. Subjects of both average and borderline intelligence had WAIS IQ scores significantly above their WAIS—R scores. For the mildly retarded subjects, the performance IQs were equal for the WAIS and WAIS—R, whereas the WAIS—R verbal and full scale IQ scores were higher than the corresponding WAIS IQ scores. However, these score differences were small (1 point) and of little practical value. The differences of moderately retarded subjects, on the other hand, were large and in the reverse direction: the WAIS—R IQ scores were significantly higher than the WAIS IQ scores. Clinical implications of these findings are discussed. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Attention problems, inhibitory control, and intelligence index overlapping genetic factors: A study in 9-, 12-, and 18-year-old twins.

It is assumed that attention problems (AP) are related to impaired executive functioning. We investigated the association between AP and inhibitory control and tested to what extent the association was due to genetic factors shared with IQ. Data were available from 3 independent samples of 9-, 12-, and 18-year-old twins and their siblings (1,209 participants). AP were assessed with checklists completed by multiple informants. Inhibitory control was measured with the Stroop Color Word Task (Stroop, 1935), and IQ with the Wechsler Intelligence Scale for Children (Wechsler et al., 2002) or Wechsler Adult Intelligence Scale (Wechsler, 1997). AP and inhibitory control were only correlated in the 12-year-old cohort (r = .18), but appeared non-significant after controlling for IQ. Significant correlations existed between AP and IQ in 9- and 12-year olds (r = ?.26/–.34). Inhibitory control and IQ were correlated in all cohorts (r = ?.16, ?.24 and ?.35, respectively). Genetic factors that influenced IQ also influenced inhibitory control. We conclude that the association between AP and inhibitory control as reported in the literature may largely derive from genetic factors that are shared with IQ. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Genetic influences on the overlap between low IQ and antisocial behavior in young children.

The well-documented relation between the phenotypes of low IQ and childhood antisocial behavior could be explained by either common genetic influences or environmental influences. These competing explanations were examined through use of the Environmental Risk Longitudinal Twin Study 1994-1995 cohort (Moffitt & the E-Risk Study Team, 2002) of 1,116 twin pairs and their families. Children's IQ was assessed via individual testing at age 5 years. Mothers and teachers reported on children's antisocial behavior at ages 5 and 7 years. Low IQ was related to antisocial behavior at age 5 years and predicted relatively higher antisocial behavior scores at age 7 years when antisocial behavior at age 5 years was controlled. This association was significantly stronger among boys than among girls. Genetic influences common to both phenotypes explained 100% of the low IQ-antisocial behavior relation in boys. Findings suggest that specific candidate genes and neurobiological processes should be tested in relation to both phenotypes. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Comparison of age-extended norms for the Wechsler Adult Intelligence Scale—Revised in patients with Alzheimer's disease.

Recent studies by R. J. Ivnik et al (see record 1993-04116-001), J. F. Malec et al (see record 1993-04120-001), and J. J. Ryan et al (see record 1991-08835-001) have provided age-extended norms for the Wechsler Adult Intelligence Scale—Revised (WAIS—R). The current study compared IQ scores based on these newer age-extended norms in 216 elderly Alzheimer's disease (AD) patients. Results showed that when the norms from Ryan et al were used, IQ scores were consistently the same as or higher than when WAIS—R manual norms were used. When the norms provided by Ivnik et al and Malec et al were used, IQ scores tended to be lower than WAIS—R manual norms for younger patients with more intellectual impairment. Results illustrate the importance of reporting the normative sample upon which IQ test scores for older adults are based and provide guidelines for selecting which set of age-extended WAIS—R norms to use with cognitively impaired elderly Ss. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Efficacy of the Satz-Mogel short form WAIS—R for tumor patients with lateralized lesions.

The ability of the P. Satz and S. Mogel (1962) short form Wechsler Adult Intelligence Scale—Revised (WAIS—R) to provide equivalent information about IQ scores and age-corrected scale scores was examined as a function of side of lesion for 34 left hemisphere damaged (LHD) and 29 right hemisphere damaged (RHD) patients having primary brain tumors. Correlations between the 2 forms were significant for all WAIS—R scores with the exception of the Object Assembly subtest for RHD patients. The short form significantly overestimated Verbal IQ, Performance IQ, Full Scale IQ, Similarities, and Picture Arrangement subtest scores. It significantly underestimated Object Assembly subtest scores. A marked percentage of patients showed 1 or more category changes for most WAIS—R scores and 2 or more category changes for some of these scores. The distribution of discrepancies between scores and the number of category changes did not differ significantly for LHD and RHD patients. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Composite reliability and standard errors of measurement for a seven-subtest short form of the Wechsler Adult Intelligence Scale—Revised.

The composite reliability and standard errors of measurement were computed for prorated Verbal, Performance, and Full Scale IQ scores derived from a 7-subtest short form of the Wechsler Adult Intelligence Scale—Revised (WAIS—R) proposed by L. C. Ward (see record 1991-00137-001). The results indicate that this short form provides IQ scores that are as reliable as the complete WAIS—R. Together with previous findings that this short form has better concurrent validity than other short forms, the present findings justify substituting it for the complete WAIS—R in most clinical and research applications. The advantages of and objections to using short forms of the WAIS—R are discussed. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

The IQ paradox: Resolved? Still an open question.

A generalized Dickens-Flynn (2001) model is presented and various simulations undertaken with it to give readers a better sense of the properties of such models. In particular, the inclusion of moderate degrees of persistence of intelligence and intelligence-relevant environment did not have much impact on the overall behavior of the models, although more extreme degrees of persistence did. Even moderate degrees, however, affected the internal relationships in the models. The importance of specifying the time scale and of addressing developmental aspects of the models is emphasized. It is noted that the translation of individual changes to population changes is not a simple matter in resolving the "IQ paradox" of large population gains over time in intelligence test scores. (PsycINFO Database Record (c) 2010 APA, all rights reserved)