The Flynn Effect Across Four Decades and Three WISC Versions in a Single Sample of U.S. Children

• In Event: The Environment’s Role in Cognitive Development across Contexts and Developmental Stages

Abstract

The secular rise in IQ scores over time, known as the Flynn Effect, provides key evidence that environmental influences can have drastic effects on the development of cognitive ability. The Flynn Effect has particularly important implications for children, altering rates at which children meet criteria for intellectual disability and learning disorders (Kanaya & Ceci, 2012; Kanaya, Ceci, & Scullin, 2003). Although it is well-known, the Flynn Effect often is not controlled for in developmental analyses. Previous studies have documented the Flynn Effect by either 1) observing cohorts from more recent generations score systematically higher than previous cohorts on the same test (cohort effects), or 2) giving the same sample two versions of a cognitive ability test and noting lower mean scores on newer versions than older versions (test version effects). Each of these approaches has limitations. Cohort effects, typically studied in European military conscripts, may not generalize to other samples, especially children, for whom cohort effects have been rarely examined. Test version effects may be confounded by measurement artifacts (e.g., changes in content across versions). In this study, we analyzed the Flynn Effect across middle childhood and early adolescence in the Louisville Twin Study (LTS; 7-15 years; N = 1187, 52% female, 89% White), in which longitudinal cognitive ability data were collected from 1957-1999 using three versions of the WISC (WISC, WISC-R, and WISC-III). Leveraging this unique data structure, we examined cohort and test version effects simultaneously in the same sample, which had not been done previously, thereby addressing the respective limitations of each approach. In addition to WISC index scores (e.g., FSIQ), subtest scores were also available, enabling us to model Flynn Effects at both the common factor and individual subtest levels. By fitting longitudinal models, we were able to control for age effects (i.e., gains attributable to intellectual development during childhood) and ascertain more stable Flynn Effect estimates. Analyses revealed significant (p < .05) evidence of the Flynn Effect across ages 7-15 years as both 1) children’s mean cognitive ability scores increasing over the four decades of cohorts included in the study and 2) mean scores decreasing upon the introduction of newer WISC versions. Effects on latent measures of cognitive ability consistently equaled gains of approximately three IQ points per decade (e.g., cohort effect on FSIQ = 3.05 points per decade), while magnitude fluctuated across individual subtests. Subtests’ gains were not strictly proportional with their loadings on a general cognitive ability factor, offering a nuanced understanding of how the Flynn Effect operates across cognitive domains. Practical and research implications of our findings are discussed, with emphasis placed on the environment’s role in driving Flynn Effect gains across childhood.

Authors

• Evan Giangrande, University of Virginia - Charlottesville
  Presenting Author

• Eric Turkheimer, University of Virginia
  Non-Presenting Author