Dr. Wendun Wang
Period: September 2026 – August 2030.
Funded by: NWO
With the rapid advancement of digitalization and the big data revolution, multi-dimensional panel data are gaining popularity across various scientific fields. Unlike traditional panel datasets, which contain observations across two dimensions, multi-dimensional panel datasets introduce a third or even higher dimension, offering richer and more complex information.
These datasets often emerge from origin-destination flow data, hierarchical structures, or the merging of multiple standard panel datasets. Network dependence is a prominent feature of this type of dataset and prevalent in many areas of social science, where the behavior and outcomes of one unit are influenced not only by his/her own characteristics but also by the features and outcomes of others. Incorporating network dependence is both theoretically and empirically important.
A major challenge in modeling this dependence in empirical studies is that the network structure is often unobserved by researchers and decision-makers for various reasons, depending on the research context. In multi-dimensional panel data, modeling network dependence becomes even more complex as it can manifest across multiple dimensions and intertwine between them.
My research focuses on identifying and recovering the latent network dependence in multi-dimensional panels, without observing the determinants and prespecifying the functional form of the network.
