Negative Reinforcement in the Description-Experience Gap

When faced with decisions in everyday life, sometimes an individual knows the probability of an event occurring and bases their decision on that knowledge (decision from description). In other situations, the individual does not know that probability and bases their decision on information learned through experience (decision from experience). In decisions from description, people are more risk averse in decisions involving gains (positive reinforcement) and more risk seeking in decisions involving losses (negative punishment). The reverse is evident in decisions from experience – the description-experience gap.

Across eight experiments I investigated decisions about loss prevention (negative reinforcement). These types of decisions are analogous to many everyday decisions, particularly within medical contexts but have not been examined previously. Participants completed a computer task where they could prevent the loss of points (negative reinforcement condition), as well as gain (positive reinforcement condition) and lose (negative punishment condition) points by clicking on images of doors. On each trial, participants chose between a safe (guaranteed chance of a smaller outcome) and a risky (non-guaranteed chance of a larger outcome) alternative. Across experiments, I varied the outcome probability of the risky alternative’s larger outcome.

In Study 1 I investigated whether I would observe a description-experience gap in decisions involving loss prevention. I also examined whether I would replicate a description-experience gap in decisions about gains and decisions about losses. Participants completed description, experience, and hybrid (i.e. a combination of descriptive information and experience) conditions. The outcome probability of the risky choice alternative’s larger outcome was either 50% or 20% across experiments. In decisions about loss prevention, information obtained from description produced lower risk preference than information obtained from a) experience when the outcome probability of the larger outcome was 50%, and b) hybrid information when the outcome probability of the larger outcome was 20%. Regarding decisions about gains and decisions about losses, the data of the current thesis lend only partial support to the phenomenon of a description-experience gap.

Because in many everyday situations an individual has both descriptive information and experience available to them, in Study 2 I investigated decisions from hybrid information more closely. I investigated whether the hybrid’s description or its experience component drives risk preference in situations where those two components conflict. Participants completed experience, hybrid/conflict, and hybrid/congruent conditions. The outcome probability of the risky choice alternative’s larger outcome was either 50%, 20%, or 10% across experiments. Results suggested that conflicting descriptions affected participants’ choice behaviour where the likelihood of the risky alternative’s larger outcome occurring was 10% and 50%, but not when it was 20%.

Across experiments, choice behaviour in decisions about loss prevention and in decisions from hybrid information did not consistently follow the predictions of previous literature regarding decisions about gains/losses, and decisions from description/experience, respectively. Decisions about loss prevention and decisions from hybrid information should be added to existing frameworks of decisions making to enable researchers to arrive at predictions of choice behaviour in those decisions. I proposed first steps towards a framework of decisions from hybrid information.