October 2012: Schooling in Adolescence Raises IQ Scores

Brinch, C. N., & Galloway, T. A. (2011). Schooling in adolescence raises IQ scores. Proceedings of the National Academy of Sciences, 109(2), 425–430. doi:10.1073/pnas.1106077109

Summary by Dr. Michael Vaughn


Previous research has associated intelligence quotient (IQ) scores with a wide array of educational and social outcomes. IQ is a relevant yet controversial topic among educational researchers and teachers, who hold two major prevailing views of IQ. One view is that IQ is fixed and does not change. The second view is that educational enrichment can improve IQ but that the “open window” for change is early childhood. Practioners generally treat with skepticism the notion that IQ can increase during the adolescent years. The present study in Norway is therefore noteworthy because researchers found evidence that IQ scores increased during adolescence and that this increase was tied to school reforms that added 2 years of standardized compulsory schooling, resulting in a new unified middle school for grades 7–9 (affecting adolescents ages 14–16). This finding may relate to the way in which IQ is defined and measured (see recommendation 2 below for details).

Study Design

The research design of this study is a quasi-experiment. The researchers compared adolescents from prereform and postreform schooling years at age 19 and overcame the effects of reverse causation and self-selection into further education. This study also interpreted the effect of the reform on IQ relative to the calculated Flynn effect for this sample (the reform effect was greater than one-third of the observed Flynn effect). The Flynn effect is a phenomenon in which over time, IQ scores increase among new generations in industrialized nations. Researchers debate whether the Flynn effect is due to actual increases in intelligence or successive generations being exposed to more tests and subsequently getting better at taking them.

Summary of Results

This study used the IQ scores of 19-year-old males who were assessed as part of universal military service. Study participants (more than 100,000 males) represented birth cohorts in the years 1950–1958. The middle school reforms this study investigated were phased in at the local level. “Nonreform” males were from municipalities that had not yet implemented the middle school reforms and were from the same birth cohorts as the males who went to middle schools that had implemented the reforms. Results showed that middle school reforms during adolescence were associated with increases in IQ scores of 0.6 IQ points, compared to nonreform participants. Although this increase may appear to be small, it suggests that school reforms can affect IQ.


Several recommendations arise from this study.

Recommendation 1: For research studies, readers should consider two types of validity when interpreting the applicability of the findings to practice.

  • First, the “internal” validity of the study helps to determine the rigor of the findings themselves. In a quasi-experimental design like this one, participants were not randomly assigned to receive or not receive school reform, and so it is possible that the schools and students in the reform group were different from those in the nonreform group in ways that affected the outcome of interest (here, IQ). Even though the study statistically controlled for this nonrandom assignment (controlling for reverse causation, self-selection, and the Flynn effect), researchers can never remove every possible covariate, so readers should consider the possibility that any differences found could be due to something other than school reform.
  • Second, the “external” validity of the study helps to determine the generalizability of the findings. In this case, the participants and schools studied are in Norway, where classrooms, schools, and districts may differ systematically from those in the United States (e.g., in terms of socioeconomic, cultural, and linguistic diversity). Therefore, to determine whether these findings are relevant to a reader’s practice, the reader should consider whether the findings relate to characteristics of the participants that are dissimilar to those in the schools and districts of the reader’s region. 

Recommendation 2: As mentioned briefly in the overview, readers should consider whether the study reports the outcome with the strongest implications for practice.

In this case, the authors focus on IQ, which historically has been characterized as a fixed (or malleable only during early childhood) trait. As consumers of research, it is important that the theoretical basis of any study be as compelling as the analysis and results. As educators, it may be more useful to consider the malleable skills that may in fact be measured by IQ, rather than IQ itself as a static index of “smartness.” For example, many studies provide evidence that early intervention increases language and other cognitive skills. When read with that perspective in mind, the findings presented here add to such evidence that IQ (as an index of overall cognitive ability indexed by different cognitive measures) may consist of malleable cognitive skills.

Recommendation 3: Building on the recommendation above, an important contribution of this study is that it calls into question notions about the fixed nature of IQ and suggests that IQ may be more malleable, and for longer, than most of us think.

The study suggests that school reforms can make important differences in the lives of not only younger children, but also of adolescents. As such, this research suggests to educators and teachers that policies and practices that can affect adolescents in the middle school years can potentially increase cognitive ability.

Recommendation 4: When considering the effects of school interventions or policies, it is important to consider their scope.

In practical terms, interventions with small effects that can be implemented in an entire society or district actually can have large effects because of the sheer number of students who benefit. Potentially, the benefit is realized across learning in multiple domains. Examples of such interventions exist in public health (e.g., adding fluoride to municipal water supplies, leading to overall improvements in dental health).