The primary question addressed by this experiment is whether performance changes when people complete the DRM task a second time after either a week or a month. Consequently, following the completion of the DRM task, participants were debriefed about the nature of the paradigm and its ability to induce false memories, and were later retested. Even under these conditions, the findings revealed that prior experience and knowledge about the nature of the paradigm is not enough to make participants immune to these types of memory distortions. The main findings of the experiment are subsequently discussed.
Contrary to what was expected, participants still falsely recognised the non-presented critical lure when retested after a week or a month. Although participants did identify a slightly lower proportion of non-presented critical lures, this minor improvement was not significant. With regards to studied items, both groups of participants recognised a somewhat smaller proportion of studied items when they were retested. Still, these changes in performance were insignificant. Despite having experienced the task on a previous occasion and being forewarned about the memory illusion, participants remained highly confident in their memories of critical lures. When retested they still believed that these critical lures had been presented on the studied word lists. A significant decrease was observed in confidence ratings of studied items, regardless of whether participants were retested after a week or a month. Participants were more confident in their recognition of studied items during the initial phase of the experiment than when later retested.
One relatively surprising finding is the fact that when participants were retested after either a week or a month, they remained highly confident in their memories for critical lures, but not for those of studied items. If warning would have been effective, the opposite pattern would have been observed. The expected pattern can be explained in the context of activation/source-monitoring framework and fuzzy-trace theory. As the former explains, memory records for studied items should hold more perceptual, contextual, and emotional details than those for critical lures (Neuschatz et al., 2003; Neuschatz et al., 2001). Therefore, warning participants about the nature of study lists should have encouraged them to adopt more stringent criterion, reducing false memories and confidently discriminating between sources of memory (Neuschatz et al., 2003; Neuschatz et al., 2001). Similarly, fuzzy-trace theory suggests that warning should induce participants to rely on verbatim-based representation rather than gist, to reduce false memories (Neuschatz et al., 2003; Neuschatz et al., 2001). If participants would have been able to discriminate between true and false memories accurately, confidence for critical lures should have been lower in comparison to those of studied items (Anastasi et al., 2000; Neuschatz et al., 2001). However, it seems that warning might have been ineffective as it appears that at least from participants standpoint, true and false memories are experienced similarly (Anastasi et al., 2000; Toglia et al., 1999). This explanation could account for the confidence ratings for critical lures and studied items observed. It is difficult to determine the exact nature of these findings. It is possible that these results are merely due to chance and that if the study were to be replicated different results would be obtained. Therefore, future replication of the present study is necessary to determine the stability of these findings.
Furthermore, it appears that warning and prior experience to the DRM were ineffective in encouraging participants to reduce false memories. Consistent with past research, we expected that participants would be able to edit their memories when retested as they previously experienced and learned more about the task (Neuschatz et al., 2003; Gallo et al., 1997). Our experiment was similar to that of Gallo et al. (1997) in that participants were given detailed information about the word lists. They were told that the words were semantically related to a non-presented word, the critical lure, and that the word lists were designed to make people falsely recognise this non-presented word. Therefore, participants should have been able to use this information to identify the critical lure when studying word lists and encode it as ‘not present’ (Gallo, 2010; Gallo et al., 1997). When asked to complete the recognition task, recollecting this information should have allowed participants to disqualify the critical lure as having been previously presented (Gallo, 2010; Gallo et al., 1997). However, in the present experiment participants received the forewarning instruction either one week or one month prior to the presentation of word lists on the retest. Although we did expect that warning would be effective and allow participants to reduce the proportion of critical lures reported, there are some plausible explanations as to why these findings might not have been replicated.
One possible explanation as suggested by Neuschatz et al. (2003) is that the failure to attenuate the false memory effect may result from the nature of the warning instruction. Some warning instructions assert either explicitly or implicitly to avoid false memories by identifying the critical lure during the presentation of word lists and encoding it as missing (Neuschatz et al., 2003). However, this raises the issue that if participants are not able to identify the critical lure, the warning is highly unlikely to be of any use (Neuschatz et al., 2003). Given the nature of the DRM task and the ease with which participants report the critical lure during the recognition test, it is unlikely that they will be able to avoid reporting a critical lure which they are not able to consciously think of (Neuschatz et al., 2003). Moreover, it has been demonstrated that the critical lure is more identifiable in some of the lists presented in the DRM task (Neuschatz et al., 2003). Hence, it is conceivable that in the present experiment the critical lure was only highly identifiable in some of the word lists and was less identifiable in others. This may have weakened the effect of warning. Nevertheless, further research is required as academics have not specifically determined what makes some critical lures more identifiable than others (Neuschatz et al., 2003).
In addition, another possibility is that participants might not have been able to reduce false memories because warning instructions may have faded or been forgotten with the passage of time (Schacter, 1999). It has been argued that interference processes can account for a great deal of everyday forgetting (Hardt, Nader, & Nadel, 2013). After initial learning takes place, mental activity which is either related or unrelated to the task at hand can impair memory (Hardt et al., 2013). As a result, the consolidation process for the memory of the learning task can be disrupted (Hardt et al., 2013). Therefore, warning instructions might have been ineffective in allowing participants to reduce false memories in the retest because these instructions might not have been consolidated in the first place. Nonetheless, it is hard to tell whether warning instructions might have faded from memory or whether interference processes might have disrupted the consolidation of memory by the time participants completed the retest, even for those who were retested after a week. To ascertain whether it is likely that warning instructions might have faded or been forgotten over the delay period, future research might investigate the effectiveness of forewarning after shorter delay periods, for example 24 hours.
Furthermore, given that the study was conducted online, it is plausible that participants did not pay sufficient attention when warning instructions were encountered once the initial DRM task was completed. This would render warning instructions ineffective as they would not have been encoded in the first place. It is hard to tell whether encoding may or may not have taken place. Thus, future replication in the lab is highly encouraged to increase the control over the methodological procedure. Replication would further allow us to conclude whether these findings are stable and reliable.
Nevertheless, assuming that the results obtained in this experiment are robust and replicable, this study has practical value to those researchers who are attempting to replicate studies and may be required to use participants who are not naïve to the DRM paradigm. Our findings reveal that even when participants have experienced the paradigm quite recently – a week ago – they are still highly prone to making the same memory illusions. Even in studies in which the false memory effect has been reduced, it has not been eliminated. This suggests that knowledge might not be sufficient for people to control their susceptibility to false memories. Therefore, it might be acceptable to recruit participants who are aware of how the paradigm works because the present experiment along with past research has demonstrated that underlying mechanisms, which do not seem to be under conscious control, account for the fact that people are not able to prevent the critical lure from being falsely reported.
To conclude, research reported here fits with the general findings in demonstrating the persistence and robustness of false memories within the paradigm. It has been demonstrated that even after a week of having completed the task, people can be highly prone to these types of memory distortions at very similar levels. The DRM paradigm has enhanced awareness that memory is indeed fallible. Although we might believe that our memory will not be deluded, this research, similar to previous studies, has proven it most likely will.