:The ability to categorize, compare and segregate the roll dynamical behavior of various vehicles from one another is a subject of considerable research interest. A number of comparison paradigms have been developed (static stability index, roll couple methods, etc.), but all suffer from lack of robustness: results developed on the basis of a particular comparison metric are often not able to be generalized across vehicle lines and types, etc., or they simply do not segregate vehicles at all. In addition, most models do not describe vehicle dynamics in sufficient detail, and some contain no dynamics at all (e.g., static stability index = t/2h).In the present work, static stability index, a two-degree-of-freedom roll model and a three-degree-of-freedom roll and handling model were used to locate eigenvalues for a sample of 43 vehicles consisting of (1) passenger cars, (2) light trucks, (3) sport/utility vehicles and (4) minivans. The objective of the work was to distinguish whether static stability index (in the simplest case) and pole location, damping ratio (where applicable), resonant frequency and time constants (in the case of dynamic models) differ from one another within and across such classes of vehicles. Parametric studies were also conducted to evaluate the effects of changes in various vehicle properties such as wheelbase, t/2h ratio, c.g. height and forward velocity on rollover. Results showed that the (1) neither the widely-used static stability index t/2h nor any of the dynamical models used below are capable of unambiguously discriminating rollover behavior between vehicle classes, and (2) that there are significant differences in damping ratio, natural frequency, and time constants when vehicle categories are compared. The models can, however, distinguish the rollover performance of cars from non-cars. In addition, the results help to quantify the effects of incremental changes in vehicle properties on these categories for two exemplar vehicles.