The power of 1000 samples
- Autor(en)
- Jakob Pietschnig, Martin Voracek, Anton Formann
- Abstrakt
The accumulated empirical evidence for long-term gains on standardized intelligence test scores in the general population, known as the Flynn effect (Flynn, 1984, 2009; Wicherts et al., 2004; Colom et al., 2005), for the most part is based on: (1) large and representative (population-based) samples; (2) healthy male subjects (e.g., from conscript mass-testing); (3) different measures of fluid intelligence; (4) data from the Anglo-American sphere; and (5) study periods prior to the 1970s. In contrast, here we provide first-time meta-analytic evidence for the Flynn effect based on: (1) small and non-representative samples; (2) samples of normals and patients of both sexes; (3) a single measure of crystallized intelligence sensu Cattell (the German-language Multiple-Choice Vocabulary Test, MWT; Lehrl et al., 1995); (4) data from the German-speaking countries of Central Europe; and (5) a more recent study period (subsequent to the 1970s). As for the central finding, sample size-weighted mean MWT IQ figures from about 1000 retrievable samples (collected between 1973 and 2009; total N = 39,000+) correlated significantly positive (r = +.37, p < .001) with study year. This effect generalized across sample type (normals vs. patients), publication status (published vs. unpublished data), and publication language (English vs. German). The estimated IQ gain per decade (Jensen’s Delta IQ) amounted to 2.2 IQ points. This novel finding contains several points of interest: (1) in contrast to the majority of findings from Anglo-American countries, in German-speaking countries the Flynn effect is also observable for the crystallized component of intelligence (Voracek, 2006); (2) in contrast to recent evidence from Scandinavia for an end or even reversal of the Flynn effect
(Flynn, 2009), IQ gains apparently are still ongoing in German-speaking countries; and (3), intriguingly so, the Flynn effect is even ascertainable through meta-analytic aggregation of a great many samples, all of them small and non-representative, thereby demonstrating the pervasiveness and robustness of this still poorly understood phenomenon.
- Organisation(en)
- Publikationsdatum
- 09-2009
- Peer-reviewed
- Ja
- ÖFOS 2012
- 501004 Differentielle Psychologie
- Link zum Portal
- https://ucrisportal.univie.ac.at/de/publications/55fa4872-5743-4d5e-a904-468ca0cc372b