Retention intervals, causal connections, and the 2016 election

Prior work has shown that memory typically gets worse over time, and narrative memory is better the more causally connected an event is within the story. Here, we assessed memory for a real-world narrative event, the 2016 US presidential election, which played out over many months. People were asked to recall election-related news stories and write brief headlines. We assessed memory in terms of the amount of time that had passed since the event, as well as the degree to which an event was causally integrated into the larger story. The data from 944 participants involved 1115 distinct news events that were coded in terms of temporal delay and number of causal connections to other events. We used negative binomial regression to examine these influences. Surprisingly, although the amount of time had little impact on memory overall, the number of causal connections was significantly related to memory. Furthermore, memory was driven more by the number of causal consequences that followed from an event rather than those that preceded it, although these had some influence. This indicates that for complex, long-term narratives, even those that play out over many months, the role that events play in a narrative’s causal structure are better predictors of memory than temporal distance.