Creating false memories using real-world objects: an attempt to expand upon a classic study

This study aimed to investigate the relationship between different types of stimulus presentation and their effects on both memory confidence and memory accuracy through researching the phenomenon known as false memories. Investigating ways to raise memory confidence and to still retain accurate memories can help lead to new developments in therapy to assist individuals with memory issues. In addition to this, more accurate memories can lead to more reliable recall of memories for the average person. In the current literature, there have been very few investigations into the creation of false memories outside of the auditory medium. Prior research has found that false memories are developed in both auditory and visual memories, but none of the research has compared the conditions against one another. In this study, we aimed to fill this gap in the literature by comparing these different conditions of stimuli (auditory vs. visual) against one another. We drew 34 participants from the University of Utah's research participant pool. We hypothesized that visual memories would have higher judgments of confidence that the critical lure was not presented as well as more accurate recall. The data revealed that certain lists of items had significant differences compared to the same items in a different stimulus presentation condition, both in memory confidence and accuracy of recall for the critical lure. These findings were consistent with our hypotheses that visual memory recall would be more accurate and have greater confidence judgments. It was found that there was no significant difference in confidence averaged across conditions, but there was for certain lists of stimuli. It was found that for 3 out of the 5 lists of stimuli, the confidence judgments of the participants were more accurate for the visual condition, and 1 out of the 5 lists of stimuli elicited a more accurate recall for participants. This implies that visual information is less susceptible to false memories than auditory information.

CREATING FALSE MEMORIES USING REAL-WORLD OBJECTS: AN ATTEMPT TO EXPAND UPON A CLASSIC STUDY by Dallin Woolford A Senior Honors Thesis Submitted to the Faculty of The University of Utah In Partial Fulfillment of the Requirements for the Honors Degree in Bachelor of Science In Psychology Approved: ______________________________ Trafton Drew Thesis Faculty Supervisor _____________________________ Jonathan Butner Chair, Department of Psychology _______________________________ Lisa G. Aspinwall Honors Faculty Advisor _____________________________ Sylvia D. Torti, PhD Dean, Honors College April 2023 Copyright © 2023 All Rights Reserved ABSTRACT This study aimed to investigate the relationship between different types of stimulus presentation and their effects on both memory confidence and memory accuracy through researching the phenomenon known as false memories. Investigating ways to raise memory confidence and to still retain accurate memories can help lead to new developments in therapy to assist individuals with memory issues. In addition to this, more accurate memories can lead to more reliable recall of memories for the average person. In the current literature, there have been very few investigations into the creation of false memories outside of the auditory medium. Prior research has found that false memories are developed in both auditory and visual memories, but none of the research has compared the conditions against one another. In this study, we aimed to fill this gap in the literature by comparing these different conditions of stimuli (auditory vs. visual) against one another. We drew 34 participants from the University of Utah’s research participant pool. We hypothesized that visual memories would have higher judgments of confidence that the critical lure was not presented as well as more accurate recall. The data revealed that certain lists of items had significant differences compared to the same items in a different stimulus presentation condition, both in memory confidence and accuracy of recall for the critical lure. These findings were consistent with our hypotheses that visual memory recall would be more accurate and have greater confidence judgments. It was found that there was no significant difference in confidence averaged across conditions, but there was for certain lists of stimuli. It was found that for 3 out of the 5 lists of stimuli, the confidence judgments of the participants were more accurate for the visual condition, and 1 out of the 5 lists of stimuli elicited a more accurate recall for ii participants. This implies that visual information is less susceptible to false memories than auditory information. iii TABLE OF CONTENTS ABSTRACT ii INTRODUCTION 1 METHODS 3 RESULTS 6 DISCUSSION 8 REFERENCES 12 APPENDIX 14 iv INTRODUCTION False memories are defined as an individual remembering something that never happened or remembering something differently than the way it happened (Roediger & McDermott, 1995). It is important to study false memories because it will help lead to a better grasp of knowledge retention (in the sense of more accurate memories). Previous studies have indicated some methods of educating individuals are more effective than others (Haddock, 1976), with a combination of auditory and visual stimuli leading to better performance in participants. In addition to this, previous studies have also found that memory confidence is not a good predictor of retention of information (Conway et al., 1991). This study could be used as a baseline to determine new ways to present information to students to increase the retention of information presented. If one presentation condition (either auditory or visual) has a higher accuracy rate than the other, further research could investigate new ways to develop instructional methods that have a higher retention rate. The current research examines false memories in the context of auditory and visual attention. While prior literature has been conducted on false memories, only a few have used the visual medium, and none have researched false memories in the real-world using 3D objects (Dechterenko et al., 2021). The brain uses different methods to encode auditory and visual information (Chaplin et al., 2018), which can hypothetically lead to differences in the recall of the same memories. While all individuals are prone to false memories, it is unclear which brain structure encourages the formation of false memories (Patihis, 2018). 1 Currently, the hallmark study on false memories found that when presented with spoken lists of similar items, participants would remember related items as being present when they were not (Roediger & McDermott, 1995). These items, which belong to the same schema but are not actually on the list, are known as critical lures. Researchers came to these findings by reading lists of words to participants, and then asking them which words they had recognized from the original list. They found that false memories were created, with the critical lures being inaccurately identified as being present. However, these researchers only presented auditory stimuli to their subjects. These findings have been replicated in a number of studies (Calvillo et al., 2018, Steffens & Mecklenbraucker, 2007). In light of these findings, our research question is this: by presenting physical objects instead of auditory stimuli, would individuals be less susceptible to false memories? An important study regarding false memories looked into autobiographical discrepancies when it comes to false memories. It has been found that individuals can be led to remember autobiographical events differently than reality, and even remember events that did not happen (Loftus & Pickrell, 1995). This study suggests that events can be fabricated through false memories. In the current literature, there is one study that recreates false memories using visual stimuli instead of auditory. They used 2D visual stimuli instead of real-world, 3D visual stimuli, and found that these 2D visual stimuli were also susceptible to false memories (Dechterenko et al., 2021). Previous literature has also found that exposing individuals to multimodal stimuli during encoding raises the memory recall rates of the stimuli (Heikkila et al., 2015). The more dimensions and information a stimulus conveys 2 to an individual, the more accurate their recall should be. With the evidence presented in prior literature, false memories should still be formed with visual stimuli, but the visual stimuli should elicit more accurate memory recall. The current study aims to investigate if visual and auditory stimuli can fabricate memories at the same efficacy level. In this study, we presented lists of items to participants as well as physical objects which mapped to the words on the lists. After the presentation, we conducted a free recall test to assess participants’ confidence that a stimulus was originally present as well as their memory accuracy in the post-presentation phase. We hypothesized that the visual stimuli would elicit more accurate memory recall, as well as more accurate memory confidence. This study provided important insights related to memory confidence and memory accuracy and how they relate to different modalities of stimuli. METHODS Participants Participants for this study were drawn from the University of Utah’s Psychology participant pool. When recruited through this pool, participants volunteered to participate for course credit. A total of 34 participants were enrolled in this study. Participants from the overall pool identified as following: 32.31% as male, 66.38% as female, .43% as transgender, .47% as other, and .004% declined to answer. Age-wise, identification is as follows: .58% were under the age of 18. Approximately 67.41% were between the ages 18-35, 2.31% were between the ages of 35-65, .021% were over the age of 65, and .60% declined to answer. Racial demographic data are presented as follows: 10.02% identified as Asian, 1.95% identified as Black, 10.28% identified as Latino, 6.09% identified as 3 Multiracial, 68.95% identified as White, 1.84% identified as other, and .88% declined to answer. Procedures Researchers administered 5 lists with 15 items on them. Lists consisted of items that were similar to objects in the same list translated to auditory and visual forms. Participants were randomly assigned to the visual or auditory condition to reduce confounding variables. Of the 34 participants, half were exposed to auditory cues used in previous literature (Roediger & McDermott, 1995), and the other half were exposed to visual cues. Auditory cues consisted of reading the words on lists aloud to participants, and the visual cues consisted of physical items of the words from the same lists being presented. The auditory cues used a previous study’s methodology (Roediger & McDermott, 1995), while the visual form used custom methods, closely replicating the Deese-Roediger-McDermott (Deese, 1959; Roediger & McDermott, 1995) task. Measures Auditory and Visual Stimuli Presentation. Participants were randomly assigned the order of the lists which they were presented with. Lists were made up of different topics, namely Furniture, Animals, States, Dishes, and Fruits. The critical lures for each list were Chair, Tiger, Texas, Spoon, and Apple, respectively. Participants were given each word for a 1.5 second interval. For the physical item stimulus, researchers presented the subjects with objects for a 1.5-second interval. After the 1.5-second interval passed, the stimulus was removed and stowed out of sight, preventing the participant from observing it again. Once all the items in the list were presented to the participant, they engaged in a free recall task where they recalled the stimuli they had observed by writing 4 them down. Participants were encouraged to write down the stimuli they remembered from last to first, and then writing down any they remembered once their memory failed them regarding the order of the items. They were encouraged not to guess any stimuli. This free recall task took 2.5 minutes to ensure memories were encoded into long-term memory. This process was repeated 5 times, once with each list in the set. Accuracy and Confidence in Detecting Critical Lures. For the visual condition, once the recall test had been performed, the participant was presented with items consisting of the physical stimuli, false lures, and critical lure for a total of 18 presented objects. For the auditory condition, a document was provided with the auditory stimuli, false lures, and critical lure. In this experiment, false lures were objects only tangentially related to the presented items. While similar in schema, they are not quite as similar as the other items presented in the list. Critical lures, however, were the most representative items in their categories, determined by the researchers. These lures, both false and critical, were not presented along with the original stimuli. Prior studies have found that objects similar to the presented items would be falsely recalled (Doss et al., 2016). When the items or list of words was presented, the participant was asked to rank each one on a Likert scale of 1 to 4, reflecting their memory confidence of each stimulus listed. A value of “1” on the scale represented “Sure the stimulus was not originally present,” and a “4” on the scale represented “Sure the stimulus was originally present.” These scale items were also dummy coded to indicate participants’ accuracy at identifying critical lures. Values of “1” and “2” were coded as 0 to indicate that the participant thought the stimulus was not originally present, and values of “3” and “4” were coded as 1 to indicate the participant did think the stimulus was originally present. 5 With the data being coded as such, a lower value for accurate confidence and memory accuracy would mean that the participant was more accurate in their response. RESULTS We used SPSS to conduct a series of two-tailed t-tests to determine which group (auditory or visual) had a more accurate confidence in their recall. To do this, group condition was entered as a predictor and the accuracy of the confidence in the recall of the critical lures were tested as separate outcomes. Additionally, we conducted a series of chi-square tests to determine which group (auditory or visual) was more accurate in their recall. Our first hypothesis was that the visual group would have more accurate confidence in their recall, which would be indicated by a lower mean response on the Likert scales, compared to the auditory group. Our second hypothesis was that the visual group would have more accurate recall than the auditory group, which would be indicated by a higher likelihood of a “0” in the dummy coded response. Participants were assessed on their performance on 5 individually presented groups of stimuli, and overall performance was assessed by computing a composite variable accounting for their performance overall. Due to the nature of the stimuli, participant performance on each list was tested and performance was assessed as a function of the condition in which stimuli were presented. Differences in Memory Confidence Our hypothesis was that the presentation of critical lures in the visual condition would elicit a higher memory confidence accuracy. We found the following results. The average of the participants’ self-reported memory confidence accuracy was measured across all critical lures to determine a difference in the different stimuli presentation 6 styles. Results indicated that there was not a statistically significant difference in the groups’ overall accurate confidence (t(32)= -1.77, p=.09) (see Figure 1). Following this overall analysis, additional t-tests were conducted to assess the condition effects on each individual critical lure. It was found that groups did not significantly differ on their performance for the Fruits or Furniture lures [(t(32)=1.21, p=.236), and (t(32)=1.46, p=.885), respectively]. See Figure 2 for the mean of the confidence ratings for Fruits and Figure 3 for the mean of the confidence ratings for Furniture. It was found that groups did significantly differ on their performance for the Animals, Dishes, and States lures [(t(32)=-2.94, p=.008), (t(32)=-2.55, p=.019), and (t(32)=-2.38, p=.03), respectively]. See Figure 4 for the mean of the confidence ratings of the Animals, Figure 5 for the mean of the confidence ratings of the Dishes, and Figure 6 for the mean of the confidence ratings for the Furniture. Results indicated that participants have a more accurate memory confidence in the visual condition when presented with stimuli relating to the listed groups of Animals, States, and Dishes in comparison to the auditory condition. Differences in Accuracy Additionally, chi-squared tests were conducted to determine if there was any significant difference in memory accuracy between conditions for each list. In this regard, we hypothesized that participants in the visual condition would be more accurate in recall of the critical lures than the auditory condition. To measure this, we dummy coded scores we used for the previous test into a bivariate response, with a response of “0” indicating the participant thought the item was not present, and a score of “1” indicating the participant thought the item was present. Similar to our previous hypothesis, we hypothesized that the visual presentation group would have a higher accuracy of recall. 7 For the chi-squared tests, the Fruit, Animals, States, and Furniture lists all tested as not statistically significant between conditions [(x2(1, N=34)=1.62, p=.20), (x2(1, N=34)=.118, p=.73), (x2(1, N=34)=.81, p=.37), and (x2(1, N=4)=2.13, p=.15), respectively]. For the Dishes group, there was more accurate memory recall in the visual condition than the auditory one, x2(1, N=34)=5.86, p=.02. We found that 100% of participants in the visual stimuli condition correctly identified the critical lure as being absent, while only 70.6% of the auditory stimuli condition participants correctly identified the critical lure as being absent. See Table 1 for the measures of accuracy across all lists. The data suggest that individuals in the visual group are more accurate than those in the auditory group, when presented with Dishes stimuli. The chi-squared tests indicated that the conditions for stimuli relating to Fruits, States, Animals, and Furniture were not statistically different from one another. DISCUSSION The primary results from the study suggest that auditory vs. visual stimulus presentation does not result in different accurate memory confidence rates overall , but does differ for different stimuli. This suggests that accurate memory confidence cannot be predicted by the condition in which stimuli were presented. The condition of presentation acts as a moderating variable to elicit different memory confidence judgements for certain types of stimuli With our findings stated above, our results were in partial support of our hypotheses. Our first hypothesis, stating that the visual condition would have more accurate memory confidence, was supported, with most lists having a more accurate memory confidence in the visual condition. Previous literature states that presenting 8 multimodal stimuli shows evidence of having better recall, which explains why the memory confidence was higher for the visual condition (Heikkila et al., 2015). Our second hypothesis stated that the visual condition would have more accurate recall than the auditory condition. Our findings do support this hypothesis for certain stimuli, but not for the overall condition. These findings are in alignment with a previous study that found that memory confidence does not predict memory accuracy, which accounts for why there was a difference in accurate memory confidence when compared to memory accuracy (Nielsen, et al., 2022). Previous literature indicates that multimodal stimuli should elicit more accurate memory recall (Heikkila et al., 2015). However, because our study only partially supports these findings, participants may have found a way to encode these stimuli the same way. For example, when presented with visual stimuli, a participant may have encoded the stimuli as words and recalled the stimuli as words, making it similar to how one would encode and recall auditory stimuli. Limitations There were many limitations to this study. To begin with, there was a rather small sample size, with N=34. This small sample size could have played into the statistical significances we found, with the size contributing to lower statistical power for our analyses. A study with a larger sample size could generate more power and provide data that could represent more objects and categories. Another limitation that needs to be considered is the role human error played into the study. Variables such as exposure length were susceptible to human error. Another major difficulty that arose during the study was that the visual stimuli had more variability than the auditory stimuli. There were different ways of presenting a real-world 9 object, such as orientation of the object relative to the participant. This provides an opportunity for another confound to affect the dependent variable. Stimuli in the current study were modeled after previous literature (Deese, 1959; Roediger & McDermott, 1995). However, with previous studies involving verbal presentation of stimuli, there was difficulty in translating those stimuli into stimuli that could be presented in both auditory and physical representations, leading to researchers to create their own lists of stimuli. If an auditory stimulus represents a conceptual word, such as “tall” or “bright,” it can be hard for researchers to translate that into a visual stimulus. Future research into this topic should find stimuli that can translate well between the auditory and visual condition, with a focus on nouns. Overall, this study has low validity, both internal and external. For the internal validity, many variables could have served as confounds, such as those listed above. For the external validity, the participants themselves were drawn from the University of Utah’s psychology participant pool, which consists mostly of undergraduate psychology students. This severely limits the samples drawn from the population, making the data not easily generalizable to the broader world. Due to the way the study was conducted and measured, the external validity itself was also low. The lists presented are not easily generalizable outside of the schema the lists presented. Implications This study implies that different conditions of stimuli do not affect the accuracy of overall recall or levels of memory confidence. Evidence points towards the type of stimuli in combination with the modality of presentation (i.e., items presented) affecting the accuracy of recall and accurate memory confidence for each condition. Observing 10 stimuli in the visual condition was associated with greater accurate memory confidence in the Dishes, States, and Animals lists (meaning that their greater memory confidence was accurate). We could also say that observing stimuli in the visual condition was associated with higher rates of memory accuracy in the Dishes list. With these conclusions drawn from the data, stimuli relating to Dishes should be presented in a visual format to see higher levels of memory accuracy. The findings suggest that an individual’s levels of memory confidence cannot be taken as a sign that those memories are accurate, which is similar to previous findings (Neilsen, et al., 2022; Saraiva, et al., 2020). The findings also suggest that the type of stimuli presented in combination with the modality of presentation also influences the accuracy of the memory confidence and accuracy of recall of the critical lure. On the larger scale of things, however, we cannot say whether one condition is more accurate than the other or if one condition elicits a higher memory confidence than the other. In the future, it would greatly benefit this study if there was an increased sample size, with a more reliable way to present stimuli to an individual. What this study has revealed is that the condition in which stimuli were presented may create more accurate confidence in the visual condition and may have more accurate recall in the visual condition. While neither of these findings is consistent across all forms of stimuli, this does suggest that future research should be done with larger, more representative sample sizes as well as an operationalized protocol that eliminates the potential for human error. False memories are resilient and are able to be formed in a variety of different modalities and situations, and further investigation is warranted into the creation of false memories. 11 REFERENCES Dechterenko, F., Lukavsky, J., Stipl, J. (2021). False memories for scenes using the DRM paradigm. Vision Res., Vol 178, pg. 48-59. Roediger, H., McDermott, K. (1995). Creating false memories: Remembering words not presented in lists. Journal of Experimental Psychology: Learning, Memory, and Cognition, Vol. 21(4), pg. 803-814. Loftus, E., Pickrell, J. (1995). The Formation of False Memories. Psychiatric Annuls, Vol. 25 (12). Pg. 720-725. Heikkila, J., Alho, K., Hyvonen, H., Tiippana, K. (2015). Audiovisual Semantic Congruency During Encoding Enhances Memory Performance. Experimental Psychology, Vol. 62(2). Doss, M., Jamila, P., David, G. (2016). The Dark Side of Context: Context Reinstatement Can Distort Memory. Psychological Science, Vol. 29(6), pg. 914-925. Neilsen, N., Gehrt, T., Bentsen, D. (2022). Individual differences in autobiographical memory predict memory confidence but not memory accuracy. Journal of Applied Research in Memory and Cognition. Advance online publication. https://doi.org/10.1037/mac0000082. Deese, J. (1959). On the prediction of occurrence of particular verbal intrusions in immediate recall. Journal of Experimental Psychology, Vol. 58(1), pg. 17-22. Saraiva, R., Hope, L., Horselenberg, R., Ost, J., Sauer, J., van Koppen, P. (2020). Using metamemory measures and memory tests to estimate eyewitness free recall performance. Memory, Vol. 28(1), pg. 94-106. 12 Haddock, M. (1976). Effects of an auditory and an auditory-visual method of blending instruction on the ability of prereaders to decode synthetic words. Journal of Educational Psychology, Vol. 68(6), pg. 825-831. Conway, M., Cohen, G., Stanhope, N. (1991). On the very long-term retention of knowledge acquired through formal education: Twelve years of cognitive psychology. Journal of Experimental Psychology, General, Vol. 120(4), Pg. 395409. Chaplin, T., Posa, M., Lui, L. (2018). Auditory and Visual motion Processing and Integration in the Primate Cerebral Cortex. Frontier Neural Circuits, Vol. 26. Patihis, L. (2018). Why there is no false memory trait and why everyone is susceptible to memory distortions: The dual encoding interference hypothesis (Commentary on Bernstein, Scoboria, Desjarlais, & Siucie, 2018). Psychology of Consciousness: Theory, Research, and Practice. Vol. 5(2), pg. 180-184. Calvillo, D., Flores, A., Gonzales, L. (2018). A brief mindfulness induction after encoding decreases false recognition in the Deese-Roediger-McDermott paradigm. Psychology of Consciousness: Theory, Research, and Practice. Vol. 5(2), pg. 131-139. Steffens, M., Mecklenbrauker, S. (2007). False memories: Phenomena, theories, and implications. Zeitschrift fur Psychologie, Vol. 215(1), pg. 12-24. Underwood, B., (1965). False recognition produced by implicit verbal responses. Journal of Experimental Psychology. Vol. 70, pg. 122-129. 13 APPENDIX Table 1 Accuracy Percentages for Critical Lures in the Auditory and Visual Conditions in Different Lists List Presented Visual Condition Auditory Condition Fruits 70.6% 88.2% States 100% 88.2% Dishes 100% 70.6% Animals 88.2% 76.5% Furniture 47.1% 52.9% 14 A 1.00 score on the y-axis indicates complete confidence in the correct response of identifying that the critical lure was not present. 15 A 1.00 score on the y-axis indicates complete confidence in the correct response of identifying that the critical lure was not present. 16 A 1.00 score on the y-axis indicates complete confidence in the correct response of identifying that the critical lure was not present. 17 A 1.00 score on the y-axis indicates complete confidence in the correct response of identifying that the critical lure was not present. A 1.00 score on the y-axis indicates complete confidence in the correct response of identifying that the critical lure was not present. 18 A 1.00 score on the y-axis indicates complete confidence in the correct response of identifying that the critical lure was not present. 19 Name of Candidate: Dallin Woolford Date of Submission: May 2, 2023 20