UC Berkeley

I examine psychological biases that influence "top-of-mind" status in task-based environments and choice under risk, and the implications of those biases for predicting and explaining consumer behavior in the field. I develop (in joint work with Jeff Holman) a model of prospective memory, defined as the capacity to recall tasks to be carried out in the future. Motivated by the economics and psychology literature on overconfidence, we introduce memory overconfidence into the model, and show that it leads to 1) inefficiently low rates of task completion, and 2) the prediction that the probability of task completion may vary inversely with the length of time allocated to completing the task. We discuss two instances where consumers face tasks that broadly fit our model - submitting rebates and cancelling negative-option subscriptions.

I then present empirical evidence of this latter instance through a field experiment with a membership-based website. I find that implementing "camouflage" pricing - that is, charging amounts such as $20.13 or $19.83 rather than "standard" amounts such as $20 or $19.95 - significantly lowers cancellation rates. Assuming that credit card statements can serve as a reminder to cancel unused subscriptions, and that the strength of this reminder is greater with standard amounts that tend to "stand out" on a bill and draw more attention than camouflage amounts, the lower cancellation rates in the camouflage-pricing groups are consistent with the prediction of our prospective memory model.

I also consider how what is top of mind may affect consumers' choices under risk. I exploit a "natural experiment" from the trading card industry to find evidence that existing models of choice under risk may understate the extent to which salient outcomes affect the valuation of risky prospects. I track auction sales of a particular card that experienced a sudden jump in value, as well as auction sales of the sealed, unopened card boxes into which this card was randomly inserted at known odds. I estimate the price jumps for both the single card and the unopened boxes, and back out the "decision weight" that would reconcile these price movements. The estimated decision weight is well beyond the range of previous experimental estimates. I discuss how consumers' attention to particularly salient or vivid outcomes (i.e., the extent to which they are top of mind) may provide an explanation for the result.