Economists often assume decision makers are hyper-rational agents with few limits to their cognitive capabilities. Sometimes labeled "Homo Economicus", these decision makers learn quickly, perform complex math, have endless information processing capacity, and are exceptionally rational (Thaler 2000). While these assumptions make it easier to model decision making behavior, they are demonstrably false. Decision makers calculate probabilities inaccurately (Allais 1953), dislike ambiguity (Ellsberg 1961), change behavior to avoid negative emotion (Luce, Bettman and Payne 2001), and are affected by the mere presence of alternatives that should be irrelevant to their decision (Huber, Payne and Puto 1982).
These violations have led to the development of other models that are better at predicting what consumers actually do. Rank dependent utility theory (Quiggin 1981) took into account imperfect probabilistic calculations. A more recent version of this theory (Schmeidler 1989) extended the model to ambiguous decisions. Prospect theory's weighting function (Kahneman and Tversky 1979) also addressed people's imperfect probabilistic calculations while the theory's editing function accounted for some of the simplification strategies that decision makers use to overcome their cognitive limitations.
Still, while better at describing certain behaviors, all of these theories share a common limitation. Like the model of Homo Economicus, these theories model behavior "as-if" humans were performing the functions prescribed by the theory. Rank dependent utility theory predicts probabilistic behavior "as-if" people rank ordered probabilities. Prospect theory's weighting function predicts probabilistic behavior "as-if" people overestimate small probabilities and underestimate large probabilities. Prospect theory's editing function describes some editing processes "as-if" people perform them to simplify decisions, but it does not describe how they actually come to choose a specific editing function (Thaler 2000). Ultimately, all of these theories generate a single equation that predicts consumer choice "as-if" consumers calculated these values and chose the option with maximal value. These models so far have either focused on what consumers should do or have focused on predicting what consumers actually do. Decision making models frequently do not attempt to describe the cognitive processes that are actually used to make a decision.
The research described in this dissertation investigates this rarely studied area in human decision making. The research does not focus on what people do. Instead, it focuses on the decision making process itself. Recent advancements in brain imaging techniques such as electroencephalography (EEG), magnetoencephalography (MEG), functional magnetic resonance imaging (fMRI), and positron emission tomography (PET) have allowed decision making researchers to examine cognitive processes that were previously thought impossible to observe. This research uses behavior and fMRI to study the decision making process as it actually occurs in human decision makers.
The dissertation adopts a theoretical framework developed in economics (Camerer et al. 2005) and psychology (Liberman 2007) to understand neuroscientific studies of behavior. The framework is a 2x2 combination of dual process theories that draws a distinction between cognitive processes that are either automatic or controlled and that are either internally (related to internal body states) or externally (related to sensory states) focused. Research hypotheses are developed and tested based on this framework.
This dissertation contributes to the study of decision making in three ways. Theoretically, it tests an alternative model of decision making that emphasizes the cognitive process underlying decision making. Methodologically, the research demonstrates the usefulness of neuroscientific techniques as a complement to more traditional methods used in marketing research. Practically, the research contributes to a better understanding of decision making processes which could ultimately benefit society by helping consumers overcome decision biases that lead to societal problems such as drug use, obesity, and race bias.