The Se-Kang Lab for Psychometrics at Fordham University studies the psychological measurement of individuals.
We investigate methods to detect a small number of central response patterns in a population – these are referred to as "core profile patterns." We can then summarize individuals as combinations of these core response patterns. This methodology is referred to as profile analysis – a type of dimensionality reduction techniques for response patterns. Applications of profile analysis include psychological and educational assessment. We study methods such as Profile Analysis via Multidimensional Scaling (PAMS) and Profile Analysis via Principal Component Analysis (PAPCA).
In addition to profile analysis, our lab studies Correspondence Analysis (CA), a technique for visualizing the associations between the levels (categories) of a two-way contingency table. CA displays the row and column categories in a low-dimensional space so that their relative locations are indicative of their associations. Typically, these associations are quantified by the chi-squared distance; however, chi-squared distances are not easily interpretable. We have developed methods to provide meaningful, interpretable estimates of these category associations.