Do losses loom larger than gains during value-directed encoding?

Abstract

People tend to encode high-value information better than low-value information. This value-directed encoding effect (VDE) is usually studied in the context of future gains, whereas the impact of future losses is rarely studied. This is a notable gap because behavioral economics suggests that participants are highly motivated to avoid losses. Here, we compared the encoding of potential gains, losses, and neutral items in three recognition experiments. We also manipulated time constraints such that participants had 1 s (Experiment 1), unlimited encoding time (Experiment 2), or 3 s (Experiment 3) to encode the valued words. VDE effects were absent when participants were given only 1 s to encode in Experiment 1. When given unlimited encoding time in Experiment 2, Win words were later recognized more often than Loss words, which were recognized more often than Neutral words. Participants also spent longer encoding Win versus Loss words. In Experiment 3, when encoding was restricted to 3 s, Win and Loss words were similarly later recognized at rates higher than Neutral words. Thus, potential future losses do not loom larger than gains during encoding. However, both yield better encoding than Neutral items if participants are given more than 1 s to encode. Additionally, in all three experiments, memory for the prior value of Win and Loss items was robust, and analyses demonstrated that VDE recognition effects were unnecessary for robust memory of value.