Network analysis encompasses a diverse range of new and rapidly-developing techniques, such as social network analysis, dynamic network analysis, and psychological network analysis. Network models provide a graphical representation of relationships (edges) between variables (nodes). In this presentation, we highlight the benefits of studying complex systems using psychological network analysis of cross-sectional data, which is an underrepresented dimension of CDST approaches to SLD.
A number of individual difference constructs related to SLD have been conceptualised as complex systems, such as L2 motivation (Henry, 2017), anxiety (Gregerson, 2020), and willingness to communicate (MacIntyre, 2020). To date, these constructs have mainly been explored from time-intensive perspectives, using longitudinal data to examine changes in a single variable over time. Cross-sectional data is rarely used in CDST research, but could offer an additional, relation-intensive perspective that is currently missing from our line of enquiry. Network analysis can identify the structural relationships between and across individual difference constructs by analysing which components interact to form a system, and how these components are related to other systems. As such, network analysis of cross-sectional data can provide us with a nomological net; a snapshot of a system in time. For example, psychologists are using network analysis as an exploratory tool to model psychological constructs such as intelligence (van der Maas et al., 2017), personality (Christensen et al., 2020) as complex systems.
We present two different network models of individual difference constructs, estimated from the data of 400 learners of Dutch as a second language. The first model is of L2 motivation, where we examine relationships between closely-related motivational constructs such as integrativeness and the ideal L2 self, instrumentality and the ought-to L2 self, intended effort, and attitudes towards the L2. This analysis is at item-level, to gain better insight into the instruments we are using to measure these constructs and to ascertain the extent that these constructs overlap or can be viewed as distinct sub-systems. The second model is on a more macro-level, where we use composite scores to analyse the relationships between multiple individual difference constructs. In this model, we expand our nomological network to include additional constructs such as willingness to communicate, anxiety, and self-efficacy. With these two models, we hope to highlight network analysis as a useful technique that can offer a new perspective to researching complex systems in SLD.
Christensen, A., Golino, H., & Silvia, P. (2020). A psychometric network perspective on the validity and validation of personality trait questionnaires. European Journal of Personality, 34(6), 1095-1108.
Gregersen, T. (2020). Dynamic properties of language anxiety. Studies in Second Language Learning and Teaching, 10(1), 67-87.
Henry, A. (2017). L2 motivation and multilingual identities. The Modern Language Journal, 101(3), 548-565.
MacIntyre, P. (2020). Expanding the theoretical base for the dynamics of willingness to communicate. Studies in Second Language Learning and Teaching, 10(1), 111-131.
Van Der Maas, H., Kan, K., Marsman, M., & Stevenson, C. (2017). Network models for cognitive development and intelligence. Journal of Intelligence, 5(2).
