This article presents a novel exploratory multigroup approach that quantifies relative group differences within an item response theory framework using tools from functional data analysis. Specifically, examinee groups are formed using different clustering methodologies based on background and attitudinal variable profiles. Item parameters for the groups are then calibrated separately and linked onto a common scale for comparison. Using non-parametric splines and tools from functional data analysis, groups are plotted in a two-dimensional principal components space and their relative difference is interpreted with respect to the different group profiles; each step in the methodology is illustrated with data from the 1999 TIMSS study.