The relationship between intelligence and academic performance is a classic topic in cognitive science, often revisited in light of new psychometric methods and specific educational contexts. The longitudinal study conducted by Gygi et al. (2017) offers a relevant contribution by exploring the predictive validity of four intelligence tests widely used in German-speaking countries: IDS, RIAS, SON-R 6-40 and WISC-IV. The research sought to assess the extent to which these tools can predict, three years in advance, academic performance in mathematics and language in school-age children.
The relevance of the study lies in the fact that, although intelligence is considered the most robust single predictor of school success (Deary et al., 2007), there is still a lack of data on the predictive effectiveness of specific tests when the validation criterion is school grades – a measure that, unlike standardized tests, integrates aspects such as motivation, self-regulation and socio-educational context.
The sample involved 103 children (M = 9.17 years), of whom 54 were followed longitudinally to collect school grades after three years. The results showed that all tests presented significant predictive validity for the overall grade point average. However, only the IDS and RIAS tests were significant predictors for both mathematics and language. The SON-R 6-40 demonstrated prediction only for mathematics, and the WISC-IV did not present significant prediction for either subject alone.
These differences deserve critical attention. The IDS, for example, includes subtests that assess aspects of working memory (phonological and visuospatial), components that are already well-established as correlates of performance in mathematics and language. Similarly, the RIAS integrates verbal and nonverbal reasoning tasks, capturing aspects of both fluid and crystallized intelligence. The SON-R 6-40, on the other hand, due to its nonverbal nature, is more sensitive to predicting mathematical abilities, which are often associated with visuospatial reasoning. The WISC-IV, in turn, although internationally recognized, was limited in the specific context of the sample analyzed, perhaps due to the lower relative weight of verbal and working memory components in the estimates of general intelligence used in the analysis (FSIQ).
It is important to highlight that the predictive effect found was of small to moderate magnitude, lower than the estimates reported in previous meta-analyses (Roth et al., 2015). The authors point out that this finding can be explained, at least in part, by variance restrictions in school grades (concentrated in the passing range), by the small sample size and by possible atypical characteristics of the group (with IQ scores slightly above the population mean).
The choice of school grades as a performance criterion also raises methodological considerations. Because they reflect continuous assessment, grades are influenced by multiple non-cognitive variables — which can dilute the pure effect of intelligence. On the other hand, this same characteristic gives ecological validity to grades as an indicator of functional performance in real learning contexts.
The study also raises important practical implications. School and clinical psychologists should carefully consider the components assessed by each intelligence test, especially when the goal is to estimate future performance in specific subjects. Tests that integrate working memory and verbal reasoning tasks, such as the IDS and RIAS, have proven to be particularly useful for this type of prediction. Furthermore, the exclusive use of total IQ (FSIQ), without attention to the particularities of specific indices, may mask information relevant to understanding the student’s cognitive profile.
Finally, the study highlights the need for further research with larger and more diverse samples, including children with special educational needs or with different cognitive profiles, in order to consolidate the predictive validity of the tests evaluated. Despite the limitations, the findings of Gygi et al. reinforce the importance of intelligence as one of the pillars of academic performance, while inviting a more sophisticated and multidimensional understanding of this construct in real educational contexts.
Reference:
GYGI, Jasmin T.; HAGMANN-VON ARX, Priska; SCHWEIZER, Florine; GROB, Alexander. The Predictive Validity of Four Intelligence Tests for School Grades: A Small Sample Longitudinal Study. Frontiers in Psychology, vol. 8, p. 375, 13 March. 2017. Available at: https://doi.org/10.3389/fpsyg.2017.00375. Accessed on: 17 June. 2025.
