There has been considerable debate within the literature between those who view recognition as the result of two processes (recollection and familiarity) and those who consider that familiarity is all that is required. Within those models that propose a role for recollection, the dominant view is that recollection involves recalling qualitative content from the study episode and thus is thought to be responsible for tasks that require episodic memory such as list discrimination tasks.
Familiarity is not generally thought to be helpful in identifying source although it may be correlated with it. Eluding the debate between a single and dual process interpretation, I instead propose that a role for content is likely for some tasks, but not necessarily for all episodic tasks.
In this thesis, the necessity of recalling qualitative information in order to adequately perform an episodic recognition task was examined across seven experiments. In order to properly address this question it was necessary that the methodology inhibited the use of content in distinguishing between episodes. All experiments contained the same basic design where participants receive multiple study and test lists constructed of the same items. In this list-specific design, the same words are tested on each list, however the targets and distracters are randomized on every list. In consequence all items become more familiar and recollected content becomes less useful with each successive list. Despite the difficulty of the task and the reduction in the availability of discriminating content, across all experiments it was found that most participants were still able to perform the task.
The Remember-Know (RK) procedure has often been applied in order to ascertain the involvement of recollection and familiarity in recognition tasks, however there has been a fair amount of controversy regarding the validity of the RK procedure. Despite the controversy the procedure is still frequently used. Including RK judgments allowed some assessment of whether content could still be useful in a list specific task. In addition, I was able to evaluate how the addition of the RK procedure might affect recognition performance.
All three experiments incorporated in the second chapter demonstrated that participants could adequately perform the list specific task. Some large differences in recognition performance were observed when RK judgments were included for the first time on the final list compared to conditions where participants were well practiced at the task or were yet to make RK decisions. However it was difficult to form strong conclusions from the data due to limitations of the design. It was further difficult to make strong conclusions about whether or not content was still useful to participants, because despite some very poor performance with the inclusion of the RK task on the final list, participants were still more likely to attribute R judgments to hits than false alarms.
In the third chapter the methodology of the initial experiments was refined. From the results of the experiments in this chapter I was able to more categorically conclude that the inclusion of the RK paradigm can alter recognition and sometimes R judgments. I was also able to more accurately measure the contribution of recollection to the task. In the last experiment in this chapter I included a justification condition where participants were asked to write down the content they had recollected. This manipulation greatly reduced the likelihood that participants would produce a R judgment and also adversely affected recognition performance. In addition, while R judgments were still frequently accurate (more often made to hits than false alarms), a chi square analysis which tests for an association between R judgments and hits, failed to indicate that content could have been useful for the majority of participants. However, there was some indication that participants were more capable of associating hits and R judgments when they were more practiced at the RK procedure.
The final experiment, included in chapter 4, supported previous findings and also provided further clarification as to how RK judgments might be affecting recognition. It was possible that knowledge about a role for familiarity and recollection in recognition memory might be enough to alter performance on the recognition task. It was also possible that actively making RK judgments affected the task. The results of the final experiment indicated more strongly that actively making RK judgments as participants perform the recognition task, could adversely affect recognition performance. Implications for the RK paradigm are discussed, as are implications for the role of content recollection in recognition memory and episodic memory.