Memory Recall, Dynamics

Recency

In immediate free recall, participants are far more likely to begin recall with the final list item than with an item from any other list position (Figure 1A). This tendency persists for the first several responses, after which recalls tend to come from more distributed list positions. Participants’ tendency to begin recall at the end of the list has been strongly linked to the well-known recency effect—the increased probability of recalling items from the end of the list. The striking recency effect seen in the data (Figure 1B) is greatly reduced when participants are asked to perform an unrelated cognitive task, such as mental arithmetic, in between list presentation and the recall period (delayed recall). Although the recency effect is easily disrupted in delayed free recall, other manipulations that influence overall recall performance have little effect on the recency effect.

Figure 1 Serial position effects. (A) The probability of recalling a list item from each list position in output positions 1, 3, 5, and 7. (B) The probability of recalling list items in any output position (the serial position curve) for data from the same immediate free recall studies.

Dissociations between recency and pre-recency effects in recall have led some theorists to argue for a fundamental distinction between short-term and long-term memory. In this view, recency arises due to retrieval from a limited capacity short-term store (STS) whose contents are easily displaced by new information. In contrast, recall of pre-recency items arises from a search of associative memory, where associations between items reflect both newly formed associations between items that were together in STS [Page 493]and long-standing associative knowledge concerning the items themselves. If, however, recency depends exclusively on the operation of STS, then one would not expect to find recency in continual distracter free recall—a task where participants perform a distracting task (e.g., mental arithmetic) after every list item, including the last one. According to the STS account of recency, the final distracter should greatly attenuate the recency effect, as in delayed free recall. However, in continual-distracter free recall, one observes a strong recency effect and participants are nearly as likely to initiate recall with the final list item as in immediate free recall. This “long-term recency” has been taken to support the view that recency reflects a more general forgetting process that operates at both short and long time scales.

...