Examining Whether Collaboration Can Improve Memory for Binding of Real-World Object Features

Abstract

Individuals can accurately store thousands of objects in their visual long-term memory. However, when objects vary on numerous features, previous research found that individuals struggle to bind the objects to their correct states (e.g., state of the studied coffee mug: full or empty). We tested whether collaboration could serve to overcome chance-level exemplar-state binding by conducting three recognition memory experiments. In Experiment 1A, participants completed 2-AFC tests, in which they had to identify either which exemplars or which exemplarstate conjunctions they had studied. Similar to previous research, we found that when participants needed to identify the exemplar-state information together, they struggled to bind this information and performed near-chance performance. In Experiment 1B, we used a withinsubject design and tested whether collaboration could enhance memory for exemplar-state binding at retrieval. To accommodate our design, we divided each task into two blocks, cutting each task in half. We found that participants who remembered individually, and those who worked collaboratively, demonstrated the ability to remember exemplars and the states of exemplars they studied. Surprisingly, they were able to successfully remember this information as a bound unit. In Experiment 2, we tested whether we could replicate this ability to successfully bind when the task becomes more challenging. Using an old/new recognition test, we found that participants who collaborated were able to discriminate above chance performance for both tasks. Thus, we found evidence that exemplar-state binding is possible by individuals who remember individually and that binding performance can be improved when individuals collaborate to remember. However, it seems apparent that the amount of information participants are required to bind impacts this ability.