In this workshop, I introduce a novel latent space modeling approach to item response data analysis. In this approach, item response data are viewed as the bipartite network between respondents and items where a respondent-item tie is made when a correct response is given to the item. This approach's most practical benefit is that it supplies an interaction map, a two-dimensional Euclidean space, to represent the relationships (or interactions) between respondents and items, between items, and between respondents.
I will give some background and basics of latent space modeling for social network data and explain how to view and understand item response data as a special type of network data. I will then present the mathematical formulation and Bayesian estimation of the proposed latent space item response model. With empirical examples, I will demonstrate how to fit the proposed model and how to process and interpret the parameter estimates and the estimated interaction map. All demonstrations will be given in R and R Studio using the package developed for the proposed latent space approach to item response data.