DRM, bellek yanılması, duygu, değerlik, uyarılmışlık
Elif Yüvrük, Haki Turan, Aycan KapucuDuygu İçerikli Sözcüklerden Oluşan Türkçe DRM Listelerinin Geliştirilmesi
Elif Yüvrük, Haki Turan, Aycan KapucuThe Deese-Roediger-McDermott (DRM) paradigm (Deese, 1959; Roediger & McDermott, 1995), is frequently used to experimentally create false memories. In a typical procedure, semantically-related word lists are studied, each of which is associated with one non-presented critical theme word, followed by a recall or a recognition test. Participants frequently report non-presented critical words at comparable levels with studied words (Roediger & McDermott, 1995). Even if participants are warned about the nature of the DRM effect, false memories for critical lures are almost inevitable (McCabe & Smith, 2002; McDermott & Roediger, 1998). The fact that the created memory illusion is quite powerful has led to the development of various sets of DRM lists in different languages.
One of the most frequently used DRM lists was created by Stadler, Roediger, and McDermott (1999). Some researchers have used Stadler et al.’s (1999) DRM lists by translating them into their own language (e.g., Anastasi, Rhodes, Marquez, & Velino, 2005, Spanish; Mısırlısoy, 2004, Turkish; Stein & Pergher, 2001, Brazilian Portuguese; Zeelenberge & Pecher, 2002, Dutch). In fact, translating DRM lists into different languages could potentially impair associations within list words (Anastasi, Rhodes, Marquez, & Velino, 2005). To overcome this problem, researchers have developed various original DRM lists in accordance with unique word characteristics in their own language (e.g., Anastasi, De Leon, & Rhodes, 2005, Spanish; Cadavid & Beato, 2017, Spanish; Göz, 2005, Turkish; Iacullo & Marucci, 2016, Italian). Yet, while some word characteristics (e.g., frequency of use, concreteness, etc.) were controlled in most of the developed lists, emotional valence and arousal of the words have often been ignored. Extensive research has demonstrated that these two dimensions of emotion substantially effect false memories (Dewhurst, Anderson, & Knott, 2012; Gong, Xiao, & Wang, 2016). To our knowledge, currently there are no emotional DRM lists available for use with native Turkish speakers. To this end, we aimed to develop Turkish DRM lists including positive, negative and neutral words that are matched on arousal.
Method
Ninety-two undergraduate psychology students (79% women) with an average age of 21.14 years (SD = 1.83) participated in this study. Participants were not informed about the nature of the DRM task; neither before the study phase nor before the test phase. DRM lists were created by using the word norms of Tekcan and Göz (2005) for imaginability, concreteness, associations, and frequency, and norms of Kapucu, Kılıç, Özkılıç and Sarıbaz (2018) for valence and arousal. Ten DRM lists were created in each emotion condition (negative, positive, neutral); each list included ten list words related to one critical lure. To prevent high memory load, each participant was presented with five randomly selected DRM lists for each emotion condition, instead of all ten lists. All participants studied positive, negative, and neutral DRM lists in a counterbalanced order followed by a recognition test including studied/old words, critical lures and noncritical lures. For each test item, participants decided whether each item was old or new.
Results
A 3 (emotion: positive, negative, neutral) x 3 (word type: old, critical lure, noncritical lure) two-way repeated measures ANOVA was conducted on participants’ “old” response rates (see Figure 1). As expected, false alarm rates (“old” responses to lures) for critical words (X = 0.61, SS = 0.19) were remarkably higher than those for noncritical words (X = 0.20, SD = 0.13), showing false memory for critical words was successfully created in participants in each emotion condition (F(2, 182) = 510.12, p < .001, η2 = .85). We then analyzed levels of false memories created by each DRM list, separately. To this end, we calculated the percentage of participants who gave an “old” response to each test word. A 3 (emotion: positive, negative, neutral) x 3 (word type: old, critical lure, noncritical lure) two-way between-subjects ANOVA was conducted on these old response rates (see Figure 2). Again, the significant main effect of word type (F(2, 138) = 204.09, p < .001, η2 = .74) showed that old response rates for critical lures (X = 0.63, SD = 0.17) were higher than those for noncritical lures (X = 0.20, SD = 0.13). Sixty-six percent of the negative critical lures, 65% of the positive critical lures and 58% of the neutral critical lures were recognized by the participants as old words.
Discussion
Emotional meanings of words are frequently ignored in false memory literature despite well-documented emotion-related effects. To our knowledge, this is the first study to develop Turkish emotional DRM lists including positive, negative and neutral words that are controlled on several word characteristics. Levels of false memories created by these newly developed DRM lists were satisfactory on each emotion condition and were comparable with existing studies in the literature (e.g., Stadler et al., 1999; Ulatowska & Olszewska, 2013). We believe that these emotional DRM lists will be useful for researchers working in different areas of psychology who are interested in investigating emotion and false memories in native Turkish-speakers.