Sequential Integration in Risky Choices from Experience Results in Mean-Variance Preferences

We show that sequential sampling models apply to risky choices from experience. Our paradigm also allows us to test whether samples are weighted equally, or whether they depend on the scale of values or the time at which they are sampled during the decision process. We find evidence of a small primacy and large recency bias in sequential sampling. We also prove a link between the decision boundaries of the Drift-Diffusion-Model and a Modified Probit model which demonstrates a relationship between individuals' preferences for variance in information structures and their willingness to trade of speed for accuracy. Exploratory analysis is still in progress.