Both recent experience and past experience affect cue effectiveness. When learning lists of words, recall is higher when:

• Cues have stronger forward connections to their targets
• The cues are connected to fewer words in memory (to reduce competition)
• When studied words have preexisting connections to cues
• When the test cue and target word share associates because there is an indirect connection

A model that might be able to explain part of this is the PIER 2 model. In the PIER 2 model, shared associates indirectly link the test cue to the target which "adds" to the overall strength of this relationship. The authors stated that the model predicts that shared associates will ALWAYS facilitate recall… except for one way of escaping this prediction which is the assumption that encoding a familiar event actually produces more than an implicit representation (consisting of the target and its associates as activations in long-term working memory), but also an explicit representation that is generated that links the target word to its context. So, the model then assumes that each representation contributes independently to recovery during cued retrieval. Implicit representation may facilitate recall and explicit representation may disrupt recall.

The lists of targets consisted of:

1/3 with small sets of associates with strong primary associates

1/3 with small sets with weaker primaries

1/3 with large sets of associates with weaker primaries

Hypotheses:

-Targets with weaker primaries will be better recalled if they involve shared associate connections with the test cue.

-Targets with strong primaries will have lower recall with shared associates (because of interference effects).

-Actual study of the primary associates will not have effects on the magnitude of the interference effect.

Procedure

Participants were shown words on a screen at a 3-second rate. They were asked to read each word aloud and to remember as many as possible. At test they were told to read each cue word aloud and then to recall a list word that was meaningfully related to the cue.

Results for Experiment 1

Recall was more likely with shared associates in the ‘weak’ conditions, and was less likely in the ‘strong’ condition. (Table 1, pg. 656). Examination of the interaction concluded that recall was significantly facilitated when a weak primary was a shared associate produced by both the target and the test cue but when a strong primary was a shared associate there were significant reductions in cued recall.

Experiment 2

Procedures and materials were the same except that the large set targets were no longer used allowing focus on the interaction of primary strength with shared associate status. The manipulation was that half of the participants studied the targets and their primaries and the other half studied only the targets.

Hypotheses:

Studying both the targets and the primaries would not have any influence on the magnitude of the interference for targets with strong primaries (but would probably lower recall). Why? Because strong primary associates are likely to become false memories because of their strong activation by the target during study that they may actually reach conscious awareness.

Studying the weak primary items may reduce the facilitation effect. Why? Because a weak primary is likely to become a competitor.

Procedure was the same except the two groups listened to either lists with 16 targets or lists with 16 targets and 16 associates.

Results

Those studying only the targets showed the same effect as in experiment 1. When the primaries were studied with the targets, this pattern was not the same but the facilitation effect was reduced for weak primaries (Table 2, pg. 658).

When both the target and the associate were studied, the probability of an intrusion was higher than would be expected.

Discussion

The retrieval of information that has recently been encoded is a function of the interaction between new and old information. The same type of information can have opposite effects on performance. These opposing effects (according to the authors) appear to be the result of the automatic activation of related information in memory.

General Implication: When a recent experience strongly revives a past experience, interference is likely when the retrieval cue reactivates memories for both experiences. But when an experience weakly revives a past experience, retrieval will be greater when both experiences are reactivated because the experience is more likely to be retrieved.

PIER 2 is capable of explaining both the positive and negative effects produced by shared associates. (by adding the assumption that strong associates are so intensely activated that they create their own explicit representations that become functionally connected to the context of the experiment.)