تفاوت سن در رد حافظه کاذب: اثر ارائه دستورالعمل هشدار دهنده و کاهش نرخ عرضه

Two experiments were conducted to examine whether children of different ages differ in their ability to reject associative false memories with the Deese–Roediger–McDermott (DRM) paradigm. Two different types of manipulations that are thought to facilitate false memory rejection in adults—slowing the presentation rate and issuing explicit warnings—were analyzed in younger and older children. The results showed that older children were more able than younger children to reject associative false memories through warnings and by slowing the presentation rate. We conclude that although older children are, in general, more prone to produce false memories with the DRM paradigm, they are also more able to reject them when certain conditions facilitate the editing process.

Although extensive research has been conducted to analyze the developmental pattern of false memories, both implanted and spontaneous, only a few studies have been concerned with exploring age differences in false memory rejection. The current study examines the extent to which children of different ages are able to use rejection strategies to avoid false memory production. Research carried out on false memories in children has flourished with studies on implanted suggestion conducted with the misinformation paradigm (e.g., Bruck and Ceci, 1999, Ceci, 1997, Ceci and Bruck, 1993, Ceci et al., 2000, Holliday et al., 2002 and Roebers and Schneider, 2000) and the imagination inflation paradigm (Ceci et al., 1994 and Pezdek and Hodge, 1999), and these studies have demonstrated that as children get older they become more immune to suggestions (Bruck and Ceci, 1999 and Ceci and Bruck, 1993). One of the explanations for such a decline in suggestibility is given by the fact that there is an increase in source monitoring ability with age, referring to the ability to correctly remember the context in which a particular memory originated. A large amount of literature has shown that young children demonstrate a poor ability to monitor the sources of their memories (for reviews, see Lindsay (2002), and Roberts and Blades (2000)), a difficulty that is particularly apparent in situations involving reality monitoring (e.g., Foley and Johnson, 1985 and Foley et al., 1983), discrimination between self-performed and other-performed actions (Foley, Ratner, & Passalacqua, 1993), or judgments about similar external sources (Lindsay et al., 1991 and Poole and Lindsay, 1995). For a long time, implanted and misinformation studies led to the conclusion that young children are more prone to any kind of false memories than older children or adults because younger children have more difficulty in monitoring or rejecting false memories. However, research on spontaneous false memory using the Deese–Roediger–McDermott (DRM) paradigm (Deese, 1959 and Roediger and McDermott, 1995) has since provided a different conclusion. This paradigm, first used with adults and then used with children, is based on a procedure in which lists of associates are presented to the participants, always omitting the converging word that is linked to all of the presented words of a list. For example, one list could be composed of several associates, such as bed, rest, awake, tired, dream, wake, snooze, blanket, etc., without the inclusion of its converging word sleep. The impressive result of this paradigm is that the converging words (critical items) may be equally or even better recalled and recognized than words presented in the lists ( Roediger & McDermott, 1995). Using this paradigm, recent studies have found an increase in false memory with age, from childhood to adulthood (e.g., Brainerd et al., 2002, Brainerd et al., 2006, Carneiro et al., 2007, Howe, 2006, Howe et al., 2008 and Odegard et al., 2008), which seems to contradict the developmental pattern usually found with implanted false memories. However, more recent studies have shown that the age increase in false memory is not a finding specific to DRM studies. When connected meaning tasks are used, implanted false memories can also increase with age, as the studies of Connoly and Price, 2006 and Fazio and Marsh, 2008, and Ceci, Papierno, and Kulkofsky (2007) demonstrated. These findings have led some authors (e.g., Brainerd, Reyna, & Ceci, 2008) to argue that there is no dichotomy between implanted and spontaneous false memories in regard to developmental reversals. According to fuzzy trace theory, connected meaning paradigms could lead to increases with age because of two factors: children’s limitations in forming semantic (gist) relations and older children’s or adults’ difficulty in using their superior true (verbatim) memories to suppress meaning-based errors in tasks requiring the processing of many meaning-sharing items (Brainerd and Reyna, 2005 and Brainerd et al., 2008). Rooted on different conceptual grounds, an associative activation theory (Howe, 2006 and Howe, 2008) has proposed that age-dependent increases in the number, strength, and automaticity of interitem associations can explain false memory developmental patterns. The role played by associative relations has been emphasized in some influential accounts of DRM effects (for an extensive presentation, see Gallo, 2006), and the approach continues to inspire the works of numerous researchers. However, it should be noted that the issue of whether there is a real dichotomy between semantic processing (as proposed by fuzzy trace theory) and associative processing (as proposed in the associative account) remains a subject of debate (e.g., Brainerd, Yang, Reyna, Howe, & Mills, 2008). At the same time, there is some evidence from DRM studies with children indicating that false memory rejection increases with age. Using a mathematical model of recognition to compare the amounts of false memory rejection in children of different age levels, Brainerd, Holliday, and Reyna (2004) and Brainerd and Reyna (2002) found an increase in false memory rejection through development. Furthermore, an event-related functional magnetic resonance imaging study conducted with the DRM paradigm suggests that the right anterior prefrontal cortex, a region activated when conditions make higher monitoring demands, is more likely to be recruited with increasing age (Paz-Alonso, Ghetti, Donohue, Goodman, & Bunge, 2008). Moreover, Rybash and Hrubi-Bopp (2000) found that when a “generate” condition previous to recall was included, where participants were instructed to think of other words that were semantically related to the presented items, young adults were able to decrease false recall, whereas children and old adults increased it. This study seemed to indicate that young adults could reject associative false recall, whereas children and old adults were unable to do so. In keeping with the same line of argument, recent research has shown that older children or adults are more efficient in the use of strategies to reject false memories (Carneiro et al., 2009, Ghetti, 2003, Ghetti, 2008, Ghetti and Alexander, 2004 and Ghetti and Castelli, 2006). Only one study suggested the opposite. Howe (2005) showed that children, unlike adults, can suppress DRM false recall after receiving directed forgetting instructions. This result led Howe to conclude that whereas adults’ false memories are more automatic and thus escape conscious control, children’s false memories are produced with greater effort and therefore are more consciously controlled, facilitating the rejection process. One way to understand the mechanisms involved in false memory rejection was provided by Gallo (2006). Gallo, assuming that false memory may result from processes that work in opposition and resting on the general source monitoring framework, proposed a distinction between two types of strategies: diagnostic monitoring and disqualifying monitoring. Diagnostic monitoring is based on the absence of an expected recollection, generating the assumption that “if that item had occurred I would have remembered it,” whereas disqualifying monitoring is based on a true recollection of another event or item, with the two memories being mutually exclusive (Gallo, 2006). So far as diagnostic monitoring is concerned, the studies by Ghetti and collaborators (Ghetti and Alexander, 2004 and Ghetti and Castelli, 2006) showed that only after 9 years of age do children begin to make use of a memorability-based strategy to reject false events of high memorability. Before that age, children were not able to apply the rule that “if that had happened I would have remembered it” to discard false events. In the case of disqualifying monitoring, Carneiro and colleagues (2009) showed that adults, unlike preschoolers and preadolescents, could spontaneously reject false memories by identifying the missing critical word and thus not recalling it later. Manipulating theme identifiability of associative lists, these authors found that false memories decreased for lists in which their critical words could be easily identified compared with lists in which critical words were hard to identify. This specific strategy—the “identify-to-reject” strategy—is used by adults to reject associative false memories (Neuschatz, Benoit, & Payne, 2003), and it seems to be a rather complex strategy not yet used spontaneously by preadolescents. According to the view that predicts increases in false memory rejection with age, one might expect smaller age differences in false memory production with the DRM paradigm whenever the situations facilitate editing processes. It is known that some factors facilitate the false memory rejection process in adults, but so far as we know none of these effects has yet been studied in children. Slowing the presentation rate and giving warning instructions to avoid false memories are two of the factors that decrease the levels of false memories in adults, and their effects can be explained by either the source monitoring account or fuzzy trace theory. So far as the time of presentation is concerned, some studies have shown that study rate effects on false recall follow a nonmonotonic function; that is, as duration is initially lengthened, false memories increase but only until a certain point (e.g., McDermott & Watson, 2001). When this inflection point is reached, the increases stop and false memories begin to decline, arguably because false memory rejection is operating. For example, some studies have shown that slowing the presentation rate reduces false recall when list words are presented from 250 ms to 5 s (McDermott & Watson, 2001) or from 500 ms to 3 s (Dehon, 2006 and Gallo and Roediger, 2002). However, the effect of slowing the presentation rate on false recognition has been less consistent (e.g., McCabe and Smith, 2002, Seamon et al., 1998, Seamon et al., 2000 and Seamon et al., 2002). From a monitoring perspective, a manipulation of the presentation rate can facilitate false memory rejection in two different ways. On the one hand, when the presentation rate increases, participants have more time to engage in item-specific processing of list items, resulting in more distinctive recollections of each item and, thus, facilitating diagnostic monitoring (Gallo, 2006). On the other hand, slowing the presentation rate could give more time to the participants to “figure out” the converging word, to realize that it was not presented, and to hold it in their minds so as not to recall it in the future, thereby facilitating disqualifying monitoring (Gallo, 2006). Also, fuzzy trace theory could explain presentation rate effects by assuming that a longer time exposure to the word lists increases reliance on specific traces of the list items (i.e., verbatim traces), resulting in greater discriminability between verbatim traces and the gist traces and, thus, facilitating the rejection of false memories (Brainerd, Reyna, Wright, & Mojardin, 2003). So far as warnings are concerned, several studies have shown that specific instructions given before study can reduce false memories, although they are not able to eliminate the DRM effect completely (Gallo et al., 1997, McDermott and Roediger, 1998, Starns et al., 2007, Watson et al., 2004 and Westerberg and Marsolek, 2006). However, if these warning instructions are provided after study, they are not efficient in reducing false memories (Gallo et al., 2001 and Neuschatz et al., 2001; but see also McCabe & Smith, 2002). It has been thought that warning instructions before study may encourage disqualifying monitoring because participants can strategically identify the related lure at study and thereby avoid false memories at test (Neuschatz et al., 2003). Also, for fuzzy trace theory, warning instructions might reduce false memories by inducing participants into verbatim-based responding. As Reyna and Kiernan (1994) demonstrated, the type of instructions that participants are given—emphasizing either gist-responding or verbatim-responding—can influence the size of the false memory effect. Two experiments were conducted to analyze the effects of giving warning instructions and slowing the presentation rate on children’s false memories. The first experiment analyzed the effect of giving warnings, and the second one analyzed the effect of giving warnings and slowing the presentation rate. The second experiment also allowed us to examine whether the contributions of each factor summate to produce a more pronounced effect on false memory rejection. We predicted that the monitoring limitations of young children, or their less developed verbatim traces, would prevent them from engaging in memory editing, and therefore we also predicted that young children would not be able to decrease false memories following warning instructions or by slowing the presentation rate. If this is the case, it would be possible to argue that although, in general, younger children produce fewer associative false memories than older children and adults, younger children are also less able to reject them. Experiment 1 The first experiment analyzed children’s ability to use disqualifying monitoring via warning instructions. Although, as mentioned previously, fuzzy trace theory could also provide an explanation for age differences in false memory rejection, we focus mainly on the monitoring account because warning and presentation rate manipulations are likely to involve different strategies of monitoring that are important for explaining age differences. With this purpose in mind, younger and older children were warned about the DRM phenomenon before the presentation of lists, and they were instructed to identify the critical word so as not to recall or recognize it later. In the case of younger children, it was expected that the warning instructions would not decrease false memories. Some arguments could justify this prediction. As mentioned previously, younger children have general difficulty in using monitoring or verbatim processing to discard false memories, but they also might be unable to engage in the process of gist extraction necessary to identify the critical words. According to this last alternative, the inability to extract the gist or the difficulty in using gist to avoid recalling or recognizing the critical words would result in a failure to apply this particular rejection strategy. In the case of preadolescents, the forecast is not so clear-cut. Although the study by Carneiro and colleagues (2009) showed that preadolescents are not able to spontaneously use the identify-to-reject strategy, it is not known whether preadolescents are able to apply it if they are instructed to do so. This experiment was concerned with this issue—analyzing the ability of preadolescents to reject false memories through disqualifying monitoring when the strategy is experimenter given. Because the warning effect had already been observed in adults’ recall and recognition tasks, this experiment included both tasks. The recognition task was administered with and without a previous recall task controlling for eventual recall contaminations in the recognition results (Roediger & McDermott, 1995).