نیاز به شناخت و حافظه کاذب در الگوی دییز -روئیدگر-مک درمات

Two experiments investigated whether false recognition in the Deese–Roediger–McDermott (DRM) paradigm is mediated by individual differences in need for cognition. In Experiment 1, participants were presented with word lists composed of associates which converge on a non-presented critical word. On a subsequent recognition test, high need for cognition participants falsely recognized a greater proportion of critical words as having been previously studied than did low need for cognition participants. Experiment 2 replicated Experiment 1, and also tested a manipulation of list strength. Word lists used were either strong or weak in terms of eliciting the critical item. These experiments show that individual differences in approach to information processing tasks can affect the rate of false memory elicited in the DRM task.

A popular technique for studying memory illusions in the laboratory is the Deese–Roediger–McDermott (DRM) paradigm (Deese, 1959 and Roediger and McDermott, 1995). The DRM task involves presentation of word lists consisting of semantic associates to a critical, non-presented word. For example, the words “snow, ice, chilly, weather, air, and frost” are all strongly associated with the critical semantic associate “cold.” The DRM paradigm typically produces high levels of false memory for the critical words, and false recall can be as high as accurate recall of studied items (Roediger & McDermott, 1995). Several theories have been put forth to explain false memory in the DRM task. According to fuzzy trace theory (Brainerd and Reyna, 2002 and Brainerd et al., 2001), at the time of encoding, both a verbatim representation of the surface form and a gist representation of the meaning of a word list are formed. The theory posits that false recollection on the DRM task occurs via the gist representation. At the memory test, the gist trace results in a sense of familiarity for the critical word, and the participant incorrectly ascribes this to having studied the word previously. According to the activation-monitoring approach (Gallo and Roediger, 2002 and Roediger et al., 2001), during list study, semantic associates to list words may be consciously or unconsciously activated through either elaborative processing of the list or spreading activation through a semantic memory network. The critical item is the highest semantic associate to the list words, and false memory for critical items results from their activation during encoding. Both the fuzzy-trace and activation-monitoring accounts hold that information processing during list encoding is an important mechanism of false memory. Further, there is a great deal of research showing that encoding conditions at study affect levels of false memory, such as presentation duration (McDermott & Watson, 2001), presentation modality (Gallo, McDermott, & Percer, 2001), and levels of processing (Thapar and McDermott, 2001 and Toglia et al., 1999). The effect of levels of processing is particularly relevant to the current research. Studies show that task instructions promoting deeper, semantic processing of DRM lists increase false memory for critical items (Chan et al., 2005, Thapar and McDermott, 2001 and Toglia et al., 1999). Although processing during encoding is a mechanism of false memory in the DRM task, very few studies have examined how individual differences in processing affect performance on the task. Yet individual differences in cognition and personality, such as working memory capacity (Engle, 2002), reflectivity/impulsivity (Swanson & Schumacher, 1986), and need for cognition (Cacioppo & Petty, 1982) have moderating effects on cognitive activity. A few studies do indicate that individual differences can affect DRM task performance. For example, higher vividness of imagery is associated with greater false recognition (Winograd, Peluso, & Glover, 1998). Expertise can also increase false recall in the expert’s domain (Baird, 2001). And recently, Watson, Bunting, Poole, and Conway (2005) found that, when warned about false memory in the DRM task, those with higher working memory spans produced lower levels of false recall, compared to those with lower spans. These studies illustrate that DRM task performance can be mediated by individual difference variables. The primary goal of the present research was to examine how an individual difference variable that is specific to encoding and information processing affects DRM task performance. 2. Need for cognition While some individuals actively seek opportunities to engage in effortful thought, others prefer to avoid extensive processing of information. This difference in information processing was conceptualized by Cacioppo and Petty (1982) as need for cognition (NFC). Those who are high in NFC engage in more elaborate information processing, which may be characterized by a deeper, semantic orientation ( Kardash & Noel, 2000). On the other hand, those low in NFC are less motivated processors and utilizers of information. NFC predicts performance on a variety of tasks (see Cacioppo, Petty, Feinstein, & Jarvis, 1996 for review). High NFC is associated with greater recall of information ( Cacioppo et al., 1983 and Kassin et al., 1990), production of more task-related thoughts ( Lassiter et al., 1991 and Verplanken, 1993), more cognitive effort ( Batra and Stayman, 1990 and Cacioppo et al., 1983), and differences in information search during decision making ( Levin, Huneke, & Jasper, 2000). Of particular relevance to the current work are studies showing that NFC is related to more thorough processing of information. In particular, high NFC individuals are more likely to elaborate on, and derive meaning from, information than their low NFC peers. For example, high NFC participants in a study by Cacioppo et al. (1983) discriminated more between strong and weak arguments, and utilized argument quality more, when forming an evaluation of a fictitious university policy proposal. Nussbaum (2005) also found that high NFC participants generated more, and deeper, arguments in an online discussion task. When presented with a mock advertisement, those high in NFC were more likely to spontaneously generate conclusions to ads that did not explicitly state them (Stayman & Kardes, 1992). Similarly, Martin, Lang, and Wong (2004) found that high NFC participants responded more favorably when advertising messages only implied the superiority of a product. Together, these studies indicate that high NFC individuals approach information processing tasks differently, with substantive effects on performance. In the current research, two experiments were conducted to examine the impact of NFC on false memory in the DRM task. As discussed, processing manipulations and theories of DRM task performance indicate that processing during list encoding is a mechanism for the creation of false memories. It was therefore expected that due to their tendency to elaborately process information, high NFC individuals would be more likely to activate critical items during list presentation, and subsequently show higher levels of false memory for those items. 3. Experiment 1 In Experiment 1, participants who were high or low in NFC studied DRM word lists and were then given a word recognition task which included the critical items, studied items, and unrelated items. It was hypothesized that high NFC participants would show a greater rate of false recognition for critical items.