Modeling Preference Data

We provide a gentle overview of modeling choice data, with an emphasis on statistical models that allow treating both observed and unobserved effects due to the decision makers and choice options. We first consider the situation when decision makers express their preferences in the form of liking judgments or purchase intentions (as in conjoint studies).Then, we consider applications that involve partial and/or incomplete ranking data -including paired comparisons and first choices. In this case, we assume that choice outcomes are a result of a maximization process, i.e., decision makers are assumed to select or choose options that have the highest utility among the considered options.