Introducing a science interest network model to reveal country differences.

In this article, a science interest network model (SINM) is introduced and a first empirical test of the model is presented.The SINM models interest as a dynamic relational construct, in which different interest components, that is, affective, behavioral, and cognitive components and related motivational components mutually reinforce one another within the development of interest.The dynamical relational perspective hypothesizes that the mutual interactions between interest components underlie the development of interest.Applying the SINM to the PISA 2015 data of 2 countries, that is, the Netherlands and Colombia, we were able to not only illuminate the structure of interactions between the different variables (i.e., indicators) in the networks of different groups of adolescents but we could also make predictions about which variables are of structural importance within the interest construct and as such worthwhile to test as potential candidates for intervention.Additionally, we were able to replicate earlier findings of the literature, namely that (a) enjoyment is central within the interest network and that (b) important structural differences exist in the interest network across countries, which, for instance, point to differences in domain specificity of interest between countries.While the network approach is sensitive to structural differences in science interest across countries, the network structure is stable across subgroups within countries.Future studies are proposed to test theoretically important assumptions of a dynamical perspective on interest, such as the causal role of different interest components in the development of interest. Educational Impact and Implications StatementStudents with a genuine interest in science are generally more inclined to engage with sciences -for example, by keeping up with news about science or going to a science center.They will learn as a result of this engagement, and are more likely to take up a career related to sciences.In this paper, we introduce the science interest network model, which reveals how mutual interactions of specific behaviors, enjoyments, knowledge components, values, and motivational components such as selfefficacy, constitute the science interest construct.We compare the science interest networks of 15-year-olds from two different countries (the Netherlands and Colombia), using the data of a large-scale assessment, PISA 2015.Important structural differences exist in the science interest network across countries.We found that in the Netherlands science interest is domain-specific (e.g., interest in climate change, but not motion and forces), whereas in the Colombia science interest is more domain-general.Moreover, different indicators are central in the national science interest constructs, suggesting to focus interventions in different countries on different aspects: Enjoyment is central in science interest for adolescents in the Netherlands, whereas having opportunities for learning appears to be more central in Colombia.