Experimental Psychology Seminar - Bastien Blain

Abstract: 

Well-being is a complex and dynamic state that fluctuates over time, often without being linked to specific, identifiable events, unlike emotions.What drives these changes in well-being? While some research suggests that well-being reflects recent experiences, particularly in gambling or learning tasks, the mechanisms underlying these fluctuations remain poorly understood. In particular, it is unclear how an individual's properties like preference and capacities to understand the structure of the world influence well-being in response to events. Here, we ask: Does well-being reflect an individual’s preferences? And, how does abstract knowledge of the world influence these fluctuations?

In the first study, we investigate the impact of altruism on well-being. While it has long been assumed that altruism promotes well-being, recent studies have produced mixed results. One possibility is that well-being is closely tied to individual social preferences. In our study, participants engaged in both altruistic and selfish actions, with only some of the outcomes realized, creating prediction errors. While generosity itself did not directly correlate with well-being, the hedonic impact of altruistic actions was greater among individuals with more generous tendencies. Through computational modelling, we found that well-being was more strongly predicted by altruistic prediction errors than by selfish ones in generous individuals. Overall, these results highlight that well-being is influenced by social preferences.

One obvious preference is for positive over negative feedback. However, what if our abstract knowledge of the world dissociates what we see (positive or negative feedback) from what we actually receive, based on our understanding of the context? The second study examines how different behavioural controllers (model-based vs. model-free) affect well-being. We designed a novel 2-step task with a strict 1-arm bandit structure, incorporating an insurance mechanism following rare transitions. This design allows us to test how behavioural controllers influence well-being. The insurance mechanism ensures that participants always receive the outcome they chose, even if the feedback seems contradictory. Our findings show that well-being is influenced by the type of behavioural controller in use in a given individual, with model-based reasoning leading to different well-being responses compared to model-free strategies.

Together, these studies provide new insights into the determinants of well-being fluctuations, demonstrating that well-being is shaped by individual properties like preferences and the ability to apply abstract knowledge in decision-making processes.

Zoom Link: https://ucl.zoom.us/j/93138064693?pwd=rgMhBIP7Xx3dpMJLYtpS6mUwOkaBoe.1
Meeting ID: 931 3806 4693
Passcode: 833033