Running head: FALSE RECALL The Effect of Warning and List Type on False Recall University of Minnesota Abstract This experiment (modeled after Watson, Balota, & Roediger, 2003) studied the effect of warning participants and list type on the DRM paradigm. Participants were split into warned and unwarned groups. Each participant was exposed to 12 lists containing four semantic, phonological, and hybrid list types and then asked to recall list items. It was found that hybrid lists produced the highest rate of false recall and warning participants about a critical lure reduced false recall in semantic and hybrid ists.
It was concluded that hybrid lists activate both phonological and semantic representations making it more difficult to identify the critical lure. Also, it was more difficult to identify the lure in the phonological condition regardless of warning status. The Effect of Warning and List Type on False Recall in the DRM Paradigm What are false memories? According to Roediger, a leading expert in false memory research, false memories include remembering events that never happened or a distorted recollection of the events that actually took place (Roediger & McDermott, 1995).
Interest in this subject has become increasingly popular with regard to the accuracy of eye witness testimony in court cases across the United States. It is important to research the phenomenon of false memories to better understand the conditions that cause false memories and how to reduce the formation of them. Although interest in this topic has recently become increasingly popular, the first research on the subject was conducted by Bartlett (1932; as cited in Roediger & McDermott, 1995). Bartlett had his participants read a passage titled “War of the Ghosts”, which went against traditional Western paradigms.
He found that as time passed, his participants retold the story increasingly inaccurately. During this time, Bartlett characterized reproductive memory as an accurate retelling of events and reconstructive memory as the act of filling in missing details, often inaccurately. Other studies conducted on false memory include the use of sentences (Bransford & Franks, 1971), thematic effects (Sulin & Dooling, 1974), and leading questions (Loftus & Palmer, 1974). In addition, false memory has been studied through utilization of the Deese, Roediger and McDermott (DRM) paradigm.
Although Deese’s work went argely unnoticed (Roediger & McDermott, 1995), his study was the first to look at semantic list types producing false memories. Decades later, Roediger and McDermott (1995) enhanced the Deese (1959) study giving us the DRM paradigm (Dehon & Bredart 2004). In these studies, participants observed a list of words (e. g. , bed, rest, snooze, etc. ) that are related to a nonrepresented word called the “critical lure” which in this case is sleep. These lists typically produce a high rate of false recall for the critical lure (Dehon & Bredart, 2004).
So, why do these types of lists produce false recall? One leading theory is that of spreading activation, which was proposed by the Collins and Loftus (1975) study. They suggested that humans have mental representations for concepts and when those representations are activated, related concepts are also activated (Collins and Loftus, 1975). For instance, if a participant read a list containing words such as cigar, fire, nicotine they would probably falsely recall the word smoke because it was activated three different times prior.
In addition to spreading activation, source monitoring can be applied to explain why participants falsely recall the critical lure. This theory states that one needs to discriminate between the different sources of information in order to avoid falsely recalling the critical lure (Watson, Balota, Roediger, 2003). In other words, the participant needs to distinguish between the critical lure and list items during recall. Watson, Balota, & Roediger (2003) looked at source monitoring through the incorporation of three different list types, semantic, phonological, and hybrid.
For the critical lure smoke, semantic lists contained items that were thematically related such as chimney, pipe, and fire. The phonological lists ncluded items that were similar in speech sounds like poke, smirk, stroke, and oak. The hybrid lists were an alternating combination of both list types (e. g. , chimney, poke, pipe, smirk, etc. ). Watson and colleagues found that the introduction of the hybrid lists greatly increased false recall of the critical lure.
They proposed that it was easier for participants to select against the critical lure in purely semantic and phonological lists because activation was limited. Specifically, semantic lists triggered conceptual cues but not perceptual cues as was the case with phonological ists. When the participant viewed a purely semantic list, they could use phonological information to distinguish between actual list items and the critical lure. When a purely phonological list was viewed, semantic information could be applied to differentiate between list items and the critical lure.
However, in hybrid lists both conceptual and perceptual cues were activated which made it more difficult for participants to distinguish between the list items and the critical lure (Watson et al. , 2003). In addition to the above studies, Gallo, Roberts, & Seamon (1997) examined alse memory but incorporated warning participants about the critical lure. They split participants into 1 of 3 conditions, forewarned about the critical lure, not warned about the critical lure, and urged to minimize mistakes. Participants viewed semantic lists and were asked to complete a recognition test.
It was found that compared to not warned and urging participants to minimize mistakes, forewarning participants significantly reduced false recognition of the critical lure (Gallo, Roberts, & seamon, 1997). In the current study, we predicted that hybrid lists would produce a higher rate of alse recall then the semantic or phonological lists because both semantic and phonological cues would be activated. In addition, we believed that warning participants about the critical lure would produce a lower rate of false recall because the participants would identify the critical lure and then select against it during recall.
Methods Participants The participants in this experiment were 52 University of Nebraska-Omaha undergraduates who were enrolled in PSYC 314 Research Methods. Materials Each list contained 12 words all of which were associated to one critical lure word either a) phonologically (e. . , code, called, fold, etc. ) b) semantically (e. g. , chill, hot, warm, etc. ) or c) both phonologically and semantically (i. e. , hybrid lists, e. g. , chill, called, warm, etc. ) The critical lure was not represented on any of the lists.
The hybrid lists consisted of alternating phonological and semantic word associates, and their position on the hybrid list was consistent with the original list. The three types of lists were created for each of the 12 critical lures resulting in 36 different lists (adapted from Watson, Balota, & Roediger, 2003). Three different files of 12 lists ontaining four phonological, semantic, and hybrid lists types were created and each participant was exposed to one of these files. These materials were counterbalanced across participants so the critical lure was represented equally in each list type.
The following three file patterns represent the order in which each list was presented: a) semantic (Lists 1, 4, 7, 10), phonological (Lists 2, 5, 8, 11), hybrid (Lists, 3 6, 9, 12) b) phonological (Lists 1, 4, 7, 10), hybrid (Lists 2, 5, 8, 1 1), semantic (Lists 3, 6, 9, 12) c) hybrid (Lists 1, 4, 7, 10), semantic (Lists 2, 5,8,1 1), phonological (Lists Equipment A microcomputer, controlled by the experimenter was used to store list types. A projector and projector screen connected to the microcomputer were used to display the lists to the participants.
A recall sheet was distributed by the experimenter to record participant responses. Procedure Research participants were split into two groups and tested during the usual class time. One group of participants was warned there would be a critical lure for each list, while the other group was not warned. Instructions were displayed on a projector screen and then verbally summarized by the experimenter. Both groups were notified that the experiment was an investigation of word list recall. Participants were informed that a total of 12 lists of words would appear on the screen, one word at a time each for 2 seconds.
They were instructed to do their best to remember the words. After the list finished, “Recall NOW’ would appear on the screen and participants would have 2 minutes to write down as many of the words they remembered from the list. The warned group received further instructions that each of the 12 lists were based on a trick word (i. e. , the critical lure) not appearing on he list such as “smoke”. They were told the list might sound like the trick word (e. g. , poke, cloak, smirk). The list might be related to the trick word (e. g. , fire, nicotine, cigar) or the list may contain both (e. . , poke, fire, cloak, nicotine). At that time, the warned participants were advised to be cautious at the time of recall to avoid “remembering” the critical lure. Lastly, all participants had the opportunity to ask questions. Following these instructions, the experimenter distributed a recall sheet to each participant. Then, the experiment proceeded to precisely follow what the articipants were told. After the last round of recall, the participants were debriefed about false recall and the purpose of the study. Design The design was mixed and consisted of two independent variables.
List Type (semantic, phonological, hybrid) was manipulated within subjects, and Warning Condition (warned, not warned) was manipulated between subjects. The proportion of false recall served as the dependent variable. Results The average proportion of false recall by warning condition and list type is presented in Table 1. The proportion of false recall was analyzed with a two-factor mixed esign ANOVA. It was determined that there was main effect of List Type, F(2, 100) = 22. 33, p < . 05 and of Warning Status, F(l, 50)= 13. 24, p < . 05. As well as an interaction between list type and warning status, F(2, 100) = 5. 6, p < . 05. Moreover, this study established that the unwarned group given hybrid lists had the highest proportion of false recall with 54% of the subjects' recall being false words. The group with the lowest rate of false recall was the warned group given the semantic lists, with a mere 5% of events considered to be false recall. To determine if there were significant ifferences in the rate of false recall between each list type, post hoc paired t-tests were performed. The alpha value was set at . 05 for each of the post hoc t-tests and corrected for family wise error according to the Bonferroni correction method.
The difference in rate of false recall created by the semantic and phonological lists was not significant, t(51)=1. 57. However, the difference in false recall rate produced by the semantic and hybrid lists was significant, t(51)= 6. 31, p < . 05. Likewise, the rate of false recall produced by the phonological condition differed from the hybrid condition, t(51)= 4. 29 , p < . 5. Now, because there was an interaction between List Type and Warning Status, independent t- tests were performed evaluating the rate of false recall between the warned and unwarned groups for all three types of lists.
The difference in the rate of false recall for the participants in the phonological condition was not significant, t(50)= . 39. The difference in false recall between warned and unwarned participants was significant in the semantic condition, t(50)= 3. 39, p < . 05 and the hybrid condition t(50)= 3. 58 p < . 05. Discussion We found that semantic and phonological lists produced a similar rate of false recall, hile hybrid lists produced a considerably higher rate of false recall.
The rate of false recall was substantially reduced in hybrid and semantic lists for the participants who were warned of the critical lure, but warning had little impact on the rate of false recall in the phonological condition. These findings were consistent with the prediction that semantic and phonological lists would produce similar rates of lists. As stated earlier, the semantic and hybrid lists had significant reduction in the rate of false recall when participants were warned, References Bransford, J. D. , & Franks, J. J. (1997). The abstraction of linguistic ideas.
Cognitive Psychology, 2, 331-350. Collins, A. M. , & Loftus, E. F. (1975). A spreading-activation theory of semantic processing. Psychological Review, 82, 407-428. Dehon, H. , & Bredart, S. (2004). False memories: Young and older adults think of semantic associates at the same rate, but young adults are more successful at source monitoring. Psychology and Aging, 19, 191-197. Gallo, D. A. , Roberts, M. J. , and Seamon, J. G. (1997). Remembering words not presented in lists: Can we avoid creating false memories? Psychonomic Bulletin & Review, 4, 271-276 Loftus, E. F. & Palmer, J. C. (1974). Reconstruction of automobile destructions: An example of the interaction between language and memory. Journal of Verbal Learning and Verbal Behavior, 13, 585-589. Neuschatz,J. S. , Benoit, G. E. & Payne, D. G. (2003). Effective warnings in the Deese-Roediger-McDermott false-memory paradigm: The role of identifiability. Journal of Experimental Psychology: Learning, Memory, and Cognition, 29, 35-41. Roediger, H. L. Ill, & McDermott, K. B. (1995). Creating false memories: Remembering words not presented in lists. Journal of Experimental Psychology:
Learning, Memory, and Cognition, 21, 803-814. Sulin, R. A. , & Dooling, D. J. (1974). Intrustion of a thematic idea in retention of prose. Journal of Experimental psychology, 103, 255-262. watson,J. M. , Balota, D. A. , & Roedtger, H. L. Ill. (2003). Creating false memories with hybrid lists of semantic and phonological associates: Over-additive false memories produced by converging associative networks. Journal of Memory and Language, 49, 95-118. Table 1 Effect of List Type and Warning Status Warning Condition Unwarned Warned List Type Semantic Phonological Hybrid . 21 . 54 . 05 . 19 . 26