Two serious problems plague research in decision sciences. One concerns aggregation of individual choice data. Much research in individual decision making routinely aggregates data across decision makers or across repeated choices made by a given individual. Yet, as the theory of voting methods has famously shown, such aggregated choices may not match the choices of any single decision maker at any given time. This problem has been known under the heading of "voting paradoxes." A second problem in the decision sciences concerns a conceptual, mathematical, and statistical disconnect between major theories on the one hand, and empirical data on the other hand. Most theories are static, whereas behavior is highly variable. In this research we address the first of these problems through a quantitative framework that dissociates individual decision making from group or societal choice. The solution is to make variability of preferences an inherent part of the theory. The second problem is addressed by leveraging recent advances in mathematical modeling and in statistical inference. These advances allow for the conduct of quantitative contests among major decision theories using laboratory and survey data.
