False memories and false judgement of frequency in a visual word-only DRM-paradigm
Jessica Carrión
Master thesis, 30 ECTS
The programme for Master of science in Psychology, 300 ECTS Autumn 2019
Supervisor: Ulrich Olofsson
Acknowledgements
I would most of all like to express my gratitude to Ulrich Olofsson for his expertise, guidance and patience in assisting me with the most challenging parts throughout the process of writing
this thesis. I would also like to thank the participants as well as those who aided me in the
recruitment process.
Abstract
Björklund & Sandström (2012) reported a false memory for the frequency in which words have appeared using a Deese-Roediger-McDermott paradigm (Deese, 1959; Roediger & McDermott, 1995). The purpose of this study is to recreate the false frequency test by Björklund & Sandström using a visual word-only condition. 31 Swedish speaking students from Umeå University were asked to study slideshows presenting words one by one
with a 3 second interval, some words repeating more than once. Afterwards the participants were asked to complete a recognition-questionnaire as well as estimate the amount of times a word has appeared. The results show a false memory effect as well as a false frequency effect when removing the correct rejections of critical lures. The participants estimated the critical lures to appear more than once. The results can be understood using
dual-process theories - Activation/Monitoring theory and Fuzzy-trace theory.
Abstrakt
Björklund & Sandström (2012) visade på falska minnen gällande den frekvens i vilken ord har dykt upp, med hjälp av Deese-Roediger-McDermott paradigmet (Deese, 1959; Roediger & McDermott, 1995). Syftet med denna studie var att återskapa testet av Björklund & Sandstöm men genom att istället presentera orden visuellt.
31 svensktalande studenter från Umeå Universitet ombads att studera PowerPoint-presentationer som visade ord en efter en med 3 sekunders intervall. Vissa ord repeterades mer än en gång. Därefter ombads deltagarna att fylla i ett igenkänningsformulär och även uppskatta hur många gånger orden har dykt upp. Resultatet visar på en
falsk-minneseffekt men också en falsk-frekvenseffekt efter att orden som deltagarna korrekt identifierade som falska tagits bort. Deltagarna uppskattade att de kritiska orden dök upp mer än en gång. Resultatet kan förstås
genom dual-process teorier - Activation/ Monitoring theory och Fuzzy-trace theory.
False memories
The concept of false memories has been a topic of interest for many years and even today. False memories refers to remembering things that never happened, or remembering things differently than how they truly happened. In the past, interest in this topic has been sparked by clients in therapy remembering previously unknown cases of abuse. This brought up the question of whether certain types of therapy can lead to the development of false memories (Roediger & McDermott, 1995). There have been numerous cases of recovered memories of childhood abuse that have been proven to be false and created through suggestions made by the therapist, or reading about similar events or through imagination (Lindsay & Read, 1994; Loftus, 1993a). Akhtar, Justice, Morrison & Conway (2018) suggest that memories before the age of two are fictional memories, and largely consist of imagined rather than experienced events.
Acquiring false memories through imagination is one of the many ways false memories are created. Elizabeth Loftus is one of many researchers interested in how false memories are acquired and has proven that people can have vivid memories of events that never occurred, simply by being told about them or imagining them (Loftus, 2005; Loftus &
Pickrell, 1995)
In modern research, false memories have branched out to several other research areas;
for example, there have been studies on how false memories can affect criminal cases. A
study by Cochran, Greenspan, Bogart & Loftus (2016) found that eyewitnesses could develop
false memories when presented with a fabricated version of their own memory report. They
fail to notice the manipulation, and their memory alters itself to be consistent with the
presented reports. Loftus has also shown that a person can recall eyewitness information that the person did not witness themselves, but instead was told about by another person. The told information becomes integrated with their own witnessed information (Loftus, 1979). False memories have practical implications for the criminal justice system. After giving a witness report, witnesses are often given a summary of their story and asked to sign it. If this summary contains errors, or if the person has been influenced by the media or other outlets, findings suggest that the errors will go unnoticed and their memory will be contaminated (Cochran et al, 2016).
This study’s area of interest is how a person's memory can be manipulated through association using the Deese- Roediger - McDermott paradigm.
The DRM-Paradigm
The Deese – Roediger – McDermott paradigm, pronounced as “the dream-paradigm,”
(Deese, 1959; Roediger & McDermott, 1995) is a method developed in order to study false memories in a scientific manner. The original study is from 1959, by American psychologist James Deese, where, after reading wordlists, participants falsely remembered critical lures (i.e a word that is not presented) that were not included in the wordlist in a single-trial free recall paradigm (Deese, 1959). Deese developed 36 lists with 12 words per list. The 12 words on the list were all associated with a critical lure; for example, if the critical lure on a list was
“needle”, the associated words might be thread, pin, eye, sewing etc. The findings showed that only some of these lists reliably made the test subjects produce the critical lure in free-recall. Deese’s main interest was why some lists produced this effect, while others did not. Deese’s findings did not have much scientific impact when they were initially presented and were largely forgotten until 1995, when Roediger & McDermott modified the method.
Roediger & McDermott (1995) first replicated the 1959 study by Deese in a
single-trial, free-recall paradigm. Deese’s result showed that the participants were often very accurate in recalling lists, and only made a few intrusions (Deese, 1959; Roediger,
McDermott, 1995). Roediger & McDermott suggested that perhaps Deese’s method showed a way of studying false recollections in free-recall and were interested in generalising the method. Roediger & McDermott (1995) modified Deese’s method to extend it to recognition tests; after being presented with several lists, the participants were asked to either recall or recognise the list of words, and answer with how much confidence they recognized or didn’t recognize a word. In a second experiment, they attempted to generalise the material by testing other lists designed to create memory errors in a single-trial free-recall. The results were impressive, showing that people remember things that are non-existent through association, while being highly confident when it came to their estimations (Roediger & McDermott, 1995). This generalised method is called the Deese – Roediger – McDermott paradigm (Deese, 1959; Roediger & McDermott, 1995). The recollected memories are often powerful and difficult to distinguish from true memories that they can be considered “phantom
memories” (Brainerd, Payne, Wright, & Reyna, 2003). This can arguably be compared to the
proposed development of fictional memories prior to the age of two as mentioned by Akhtar et al (2018).
Being warned beforehand about the purpose of the study reduces but does not eliminate a false memory effect (McDermott & Roediger, 1998). Directed forgetting of the DRM-list has shown to increase the false memory effect (Kimball & Bjork, 2002; Seamon, Luo, Shulman, Toner, & Caglar, 2002) Other studies have explored whether feedback and redoing the DRM-test can eliminate the effects of false memories, with most studies reporting substantial error preservation (e.g., Butler & Roediger, 2005; Pashler, Cepeda, Wixted, &
Rohrer, 2005; Peeck, van den Bosch, & Kreupeling, 1985; Phye & Bender, 1989). Efforts to prevent false memory has had minimal effect. McConnell and Hunt (2007) reported both positive and negative effects on DRM following feedback. The positive being that false memories are reduced and correct recall is increased. The negative being that new errors were more likely to occur following feedback compared to without feedback.
The DRM-paradigm is an effective method in studying false memories, however it does not study false memories by the frequency that they believe something occurs, such as the critical lure “needle” appearing more than once, despite never being presented. A Swedish master thesis by Rebecka Björklund and Elin Sandström (2012) focuses on whether the DRM-Paradigm could lead to false judgement in frequency when it comes to false memories, meaning if it is possible to remember a word several times although it was never presented.
The study was done with 36 participants at Umeå University, who were instructed to listen to a series of lists of associated words where a selected number of words were repeated several times. They were afterwards instructed to try to estimate how many times (frequency) the words were presented. The participants remembered the critical lures on an average of 1.76 times. The participants were equally certain about the critical lures appearing in the list as with the presented ones.
The study by Björklund and Sandström (2012) shows that false memories are possible even in frequency; however the study has certain issues that need to be addressed. Two lists, matta (carpet) and hammare (hammer) included words that were too alluding to the critical lure, making it apparent to the participants that there was a manipulation attempt. This could possibly have led to an unwanted effect in the results, as knowing of the manipulation attempt might have affected their outcome, as results by Roediger & McDermott (1998) have shown.
The lists that led to this outcome will be exchanged to other lists with high intrusion rates.
This will be explained in detail in the material section.
Visual DRM-tests
Israel & Schacter (1997) proposed the idea that participants would be able to reduce false memory by using the distinctiveness of the format in which the items were presented.
One group was given a presentation where the words were presented visually and auditorily
(word-only condition) simultaneously. A second group was given a presentation where the
words were presented auditorily and were shown black and white line drawings (picture +
word). False recognition was significantly lower in the second group. The word-only group
did not get a significant result compared to a control group where the words were simply presented auditorily (classic DRM-test). The authors argue that the visual pictures led to more perceptually distinct memories than the study of words only due to the pictorial information.
They speculate that participants most likely relied less on perceptual information in the word-only condition. The extra information provided by the written words was not helpful in reducing false memories. Whereas Israel & Schacter were getting a significant effect only from using pictorial materials, Smith and Hunt (1998) argued that presenting words visually would be enough to reduce a false memory effect, compared to the effect shown when testing DRM in the original auditory manner. It shows that the false recall for a visual word
presentation of DRM is dramatically lower than the recognition for the auditory version. A 50% reduction was shown by simply changing the manner of presentation from auditory to visual. The theory presented was that the visual memories lead to more distinctive item processing. They propose that the discriminability between the visually studied items and the internally generated events is greater than the discriminability between the auditory events and the internally generated event. The study by Smith and Hunt (1998) is the first to study the DRM paradigm visually without audio, but there have been many studies of using a visual presentation alone to study false memories.
Gallo, McDermott, Percer & Roediger (2001) mention that there is plenty of evidence suggesting that false recall is sensitive to study modality, however there has been little
evidence suggesting that there is a dramatic reduction is false recognition following a word-only visual DRM-test. They argue that several studies have found robust level of false recognition, for example Arndt & Hirshman (1998) and Seamon et al. (1998). But there are also many studies suggesting that there is no difference between an auditory and visual (word-only) condition (eg. Miller & Gazzaniga, 1998 and Israel & Schacter, 1997). Maylor and Mo (1999) did both a visual and auditory DRM-test with an immediate item recognition after each test. Similarly to Smith & Hunt (1998), Maylor & Mo (1999) also found a dramatic reduction of 50%, however they received the opposite result - visual false recognition was greater than auditory. The authors argue that auditory presentation leads to more perceptually detailed and temporarily distinct memories, and were able to use reality monitoring to reduce false recognition. This result is the opposite of Smith and Hunt (1998) where the results found that false recognition was lower following a visual study. Both pairs of authors argue that the memory that caused the least false memory effect was more distinct than the other.
However Gallo et al (2001) argues that the study by Smith and Hunt (1998) does fit with the theory that false memory is caused by thoughts, in which case it is reasonable to think that visual presentation would be more discriminable from such effects. In an attempt to test if the modality does affect the outcome of false recognition on a visual drm-test, Gallo et al (2001) replicated the study by Smith & Hunt (1998) but obtained a smaller effect result.
Several studies (eg. Cleary & Greene, 2002; Gallo et al., 2001; Gallo & Roediger,
2003; Kellogg, 2001; Pierce, Gallo, Weiss & Schacter., 2005; Smith & Hunt, 1998; Smith,
Lozito, & Bayen, 2005) have shown that false memories are reduced following a visual
DRM-study as explained in Smith, Hunt and Gallagher (2008). The authors elaborate by
noting that although those effects appear to be reliable, several experiments have gathered a
small effect size compared to auditory tests (e.g., Gallo et al., 2001; Gallo & Roediger, 2003) or gathered none at all (McKone & Murphy, 2000).
Although there have been results showing that there is a false memory reduction when testing visually compared to auditory, the effect has not been uniformly demonstrated. Smith, Hunt, and Gallagher (2008) tested three different boundary conditions to see if the results are dependent on the modality. The results point towards a false memory reduction when tested visually relative to an auditory version only if a recall test preceded a recognition test. A second finding is that the reduction in false recognition recurred for non-recalled critical items. This is important to note because the latter result suggests that the false recognition effect is not based on the preceding recall test or fewer critical items, but rather that the preceding recall test influences processing during recognition. A third point is that there is a reduction in false recognition when not preceded by a recall test if the participants were told to carefully scrutinize their own memory. They add that the order of the recognition items was not important (Smith, Hunt & Gallagher, 2008).
There are few examples studying DRM visually using word-only presentation.
However, there have been many studies studying DRM visually using imagery (e.g. Israel &
Schacter, 1997; Schacter, Cendan, Dodson, & Clifford, 2001; Schacter, Israel, & Racine, 1999; Baioui, A., Ambach, W., Walter, B., Vaitl, D., & Alain, C, 2012; Bourscheid, F., Pinto, L., Knijnik, L., & Stein, L, 2014), and the effects have not been eliminated completely. This study is focusing on a word-only presentation of visual DRM. As of 2020, there has not to my knowledge been a single study examining the frequency effect of false memories on visual DRM, meaning if the participants will falsely believe that a word appeared more than once. This study will focus on whether the false memory effect, although shown to be smaller than the auditory versions, will also show a false frequency effect.
False memory for frequency
Björklund & Sandström (2012) suggested that their results were accounted for by the activation monitoring theory (Roediger et al., 2001). Activation-monitoring are two processes that produce false memories, and suggests a reason why it is difficult to tell apart the critical lures from the studied ones in a DRM-test. Activation processes are generally (however not exclusively) the result of encoding circumstances that makes a word's representations activate, which then increases the risks of false memories when related words are presented.
Monitoring processes are generally (not exclusively) activated at retrieval and are enhanced by the encoding of associated words, and aid the participants in judging the source of a memory. In a DRM-test the encoding of strongly associated words leads to participants falsely assuming a word has appeared, misjudging the source of a word based on the strong associations (Ardnt & Gould, 2006, Björklund & Sandström, 2012). The
Activation-Monitoring theory is activated through studying a word and its origin is determined. The memory then spreads to nearby similar or related words which then
increases the risk of false memories. Björklund & Sandström suggests that this might be the
reason why they acquired their result, that an error in encoding led to them believing that
more words appeared, and also more than once. The participants mistake that the critical lures have appeared due to the strong associations and are misled about the word’s origin. The authors believe that the process is the same as that in the original DRM-test, despite the words being read multiple times, unlike the original test. The Activation-Montoring as a theory for false memory reduction gains support from visual experiments on DRM and also suggests that the fact that the participants were more certain about the non-critical lures (lures not related to the presented lists of associated words) also adds weight to the theory
(Björklund, Sandström, 2012). The items on the DRM-list become more distinct when presented alongside images, reducing the false memory effect. (Schacter, Cendan, Dodson &
Clifford, 2001).
They also suggest the Fuzzy-trace theory (Brainerd et al., 1995; Reyna & Brainerd, 1995) also mentioned by Ardnt & Gould (2006). According to the Fuzzy trace theory, memory is processed on a continuum between verbatim traces and gist traces. The verbatim memory records perceptual features while the gist memory stores more meaning-based elements of memory (Ardnt & Gould, 2006; Arndt, 2012). A gist-memory stores the meaning, the elaborations of a memory and the commonalities between a set of similar experiences, which leads to higher risk of creating false memories. The false memory occurs when the multiple associated words in a DRM-test activates a similar gist-representation (Ardnt & Gould, 2006; Ardnt, 2012). The verbatim memory is better at discriminating false memories due to its attention to detail (Arndt, 2012). Björklund and Sandström (2012) agree that the fuzzy trace theory is a likely reason why their results occurred. The participants used the gist memory to process the information and struggle to identify the details of their
memory. Then instead used the strong associations to that word and falsely assumed the word had appeared several times. Björklund and Sandström believe that even though their study is studying the frequency of false memories and not just simply if they appeared or not, the theory is still applicable (Björklund, Sandström, 2012).
Purpose
This study is a continuation of a study by Rebecka Björklund and Elin Sandström (2012) where the aim was to study if the DRM-paradigm would lead to false memories in frequencies if words are repeated.
The purpose of this study is to see if the visual word-only DRM-paradigm will create a false memory effect, and to study if a false frequency effect will be observed.
1. Will a false memory effect be observed, i.e., more estimates that are greater than zero for critical lures than non-critical lures?
2. Will a “false frequency” effect be observed, i.e., will there be more estimates that are greater than one than equal to one, for critical lures?
Method
Sample
The sample for this study was drawn from students attending Umeå University, from Facebook groups and from word of mouth recruitation. The total number of participants were 31 with 52% women and 48% men. All participation was voluntary. The participants were between the age span of 21-46. The cut out criteria was that the participants needed to be between 18-50, should understand and speak Swedish as well as not being psychology students - with the exception of first term and third term. The reason for this criteria is that the students attending the 1st and 3rd year have not been tested using DRM as part of a course.
Material
The wordlist used in this experiment was the same as in Björklund and Sandström (2012) with some minor changes, based on the reason that some lists caused the participants to be aware of the manipulation. The lists, matta (carpet) and hammare (hammer), were exchanged to söt (cute) and lång (tall), in an attempt to avoid this outcome. The lists used were based on the lists by Johansson and Stenberg (2002). These lists are a Swedish version of wordlists translated using the words from Kent-Rosanoff Word Association Test as reported by Russell Jenkins (1954), which in turn was used by Deese and Roediger and McDermott. They compiled lists of the most associated words to the 100 translated words (Johansson, Stenberg, 2002).
The lists chosen had an associative strength between 0,75 and 0,95, with the critical lures arbete (work), flicka (girl), flod (river), grov (coarse/ rough), söt (cute), kvadrat
(square), mjuk (soft), nål (needle), rädd (scared), tjuv (thief), tobak (tobacco), and lång (tall).
The 12 lists will be divided into two groups of 6 lists each, group 1 and 2. Each list contained 9 unique words, and each word was repeated 1, 2, or 3 times. Three variations of each list had been made to ensure every word is repeated an equal amount of times, i.e., that each word is on one variation of a list 1 time, another list 2 times, and 3 times in a third list. This makes a total of 32 variations of lists.
A 13 page stapled handout was constructed, wherein the first two pages shows general information about the test, and a form where participants were asked to fill out their age, gender, and consent to participate. The following 6 pages contained mathematical questions the participants will be asked to complete between each viewed list. Page 9-11 had an answer sheet with information, a word recognition test, and questions regarding their confidence in their recollection of the words they had seen. The options were “Very Uncertain”,
“Uncertain”, “Certain” and “Very Certain”. Page 10 has two questions the participants were asked to answer before more information about the test is provided. The two questions were;
“Have you done anything similar to this before? If so, what?” and “Did you notice anything
about the lists, if so, what?”. The reason for these two questions was to see if the participants
had noticed the manipulation, to be certain whether my changes had prevented this outcome.
The last page shows more information about the test, and informed the participants that they had participated in a false memory test.
Procedure
The participants were divided into two groups, Group 1 and Group 2. They were given written information in a group 1 or group 2 specific handout stating that they are to participate in a visual memory test, and that they will study a powerpoint presentation that will present a series of words from 6 of the designed lists, one word at a time. They were informed that some words might appear more than once. After each list, they were asked to complete mathematical questions that are included on a separate page in the handout and were given 15 seconds to complete each. The same instructions were given to all participants.
After reading the instructions, they were asked to fill out their gender, age, if Swedish is their first language, and if they have any visual impairments or any other impairment that could affect their performance. The participants were instructed to try to remember as many words as possible, and that they were not allowed to write any words down while viewing the screen.
After viewing the Powerpoint presentation, participants were asked to recall as many words as they can remember, and state how many times they appeared on the Powerpoint, as well as answering questions regarding their certainty of the frequency of how many times the words appeared. They were given 6 minutes to complete the recollection part of the test.
Ethical considerations
The data collection has been anonymised and treated according to the Swedish ethical guidelines. The participants were informed about the confidentiality of the study, that
participation is voluntary, and that withdrawing consent is possible at any point of the process. The data was stored on a password-protected computer, only accessible from the computers at the university.
Many participants were recruited through convenience selection. This means that the author has a personal connection to some of the participants, however, due to the quantitative nature of this test, the effect of personal relationships should be minimal.
Result
The result is based on data from all participants (n=31) and a p-value of 0,05 was set.
Some participants mentioned that the words were related to a word that did not appear. These
participants have not been excluded because of the reason that a false memory effect will not
be eliminated, although reduced, even though the participants are aware of the manipulation
(Roediger & McDermott, 1998). The mean value on answers on the words that appeared 1, 2
or 3 times, as well as the critical lures and non-critical lures were calculated. See Table 1
Table 1. Descriptive data for the variables in the experiment.
Note : Kön= Gender. Ålder= Age. V= Variable. F= Frequency. K= Confidence. T= Critical lures. D = Non-critical lures.
In order to see whether there was a false memory effect, the number of estimates that were greater than zero were compared for critical lures ( M = 3.9, SD = 1.68) and non-critical lures ( M = 0.68, SD = 1.17) which proved to be significant, t(30) = 3.9, p < .001. Next, in order to see whether there was a “false frequency” effect, the number of estimates for critical lures that were greater than one ( M = 2.58, SD = 1.8) was compared to the number of
estimates that were exactly one ( M = 1.32, SD = 1.33). This difference also proved to be significant, t(30) = 2.6, p < .01
The results show that when there are false memories, the mean value is closer to 2 than 1. This means that it is more common to estimate that the critical items appear 2 times than 1, if the participant believes the word appeared. There were 116 cases of false memories for critical lures compared to 64 correct rejections.
As can be seen in Table 1, the participants were less confident in their estimates for critical lures than for non-critical lures, t(30) = 4.86, p < .001. A one-way repeated measures ANOVA on presented words (frequency 1-3) and critical lures, yielded a significant effect of confidence, F(90) = 11.9, p < .001. Bonferonni post-hoc tests (see Table 2) showed that this was due to lower confidence ratings for critical lures than for presented words.
Table 2. Bonferonni post hoc comparisons of mean differences in confidence ratings for presented words and critical lures
F2 F3 CL
F1 0.114 0.051 0.384*
F2 -0.063 0.270*
F3 0.333*
Note . * = p < .01
Similar to the study by Björklund and Sandstöm (2012) the non-critical lures have the lowest standard deviation ( SD = 0.31) and the critical lures the highest (SD = 0.88). When looking at specific items of critical lures on the lists a large variation is found. Tobak
(Tobacco) has the lowest mean estimate (0.40) and Tjuv (Thief) the highest (2.07). Ten of the
12 lists have a mean frequency estimate that is less than 1.5.
Discussion
The purpose of this study is to see if the visual word-only DRM-paradigm will create a false memory effect, and to study if a false frequency effect will be observed. The results show that there is a false memory effect. The participants believe that words have appeared even though they did not. A false frequency effect for visual word-only DRM was found, there were more estimates that were greater than one, than exactly one, for critical lures. In other words, the results showed that when a false memory was created, the participants were more likely to estimate its frequency as greater than one. This could be interpreted as when false memories are created, a false memory of the words frequency is also likely. There were 116 instances of false memories compared to 64 where the participants’ memories did not get manipulated, nearly twice the amount.
The participants were more confident regarding the non-critical lures compared to the critical lures. This is not surprising considering the non-critical lures were not related to the associative words, and are easier to discriminate from the other words. One interesting thing to note is that the participants were more confident regarding the associative words compared to the critical lures. This implies that there was an aspect of doubt regarding whether the critical lures appeared or not. A possible reason for this is that the test was presented visually, which could have led to a decrease in false memory.
The participants were also asked if they had taken part in a similar test and to
comment on the nature of the test, if they noticed anything strange. The purpose for this was to see if the participants were aware of the manipulation after doing the test. None of the participants had taken part in the DRM-test in the past. The participants did notice a theme and a third of the participants mention noticing that certain words did not appear. It is unclear whether the participants are referring to the non-critical lures or the critical lures, which is an error that was noticed afterwards. Only a couple mentioned specifically that they associated to a word that didn’t appear or that they noticed I was studying false memories. A few participants mentioned having memory difficulties or other conditions, which might have affected their results.
Tobak (Tobacco) was a word that was frequently mentioned as a word that did not appear by participants in both groups, even though it was a critical lure in test 2. One possible reason behind this is that some of these categories are used less in current society, for
example tobak (tobacco). The source of the categories used (Johansson & Stenberg, 2002) is 18 years old.
Theoretical discussion
There are many studies on the DRM-paradigm both auditory and visually using imagery. However there are fewer studying the DRM-paradigm using word-only
presentations. There have been inconsistent findings regarding if word-only presentations
does have a reduced false memory effect, and several have had little to no effect (e.g. Gallo et
al, Miller & Gazzaniga, 1998 and Israel & Schacter, 1998), or the opposite effect (Maylor &
Mo, 1999).
In this study there is a clear false memory effect from the visual test. Although this study did not compare the results between a visual and auditory DRM-test, compared to the first made study testing out false memory for frequency by Björklund & Sandström (2012) where the results show that the average mean value for frequency is 1.76, our result 1.31 of indicates that there is a reduced false memory effect. However there would be a need for an additional test to confirm if the reduced effect is significant. There is to my understanding no study that has studied visual DRM-test with a word-only condition on frequency. The results show that there is an indication of a false frequency effect for a visual word-only DRM when comparing the instances of false memory. A possible reason for the smaller effect compared to Björklund & Sandström (2012) is that the visual memory is more detailed than the
auditory, as speculated by Smith & Hunt (1998), and Gallo et al (2001). However the results between the studies have not been consistent, and it has been suggested that this only applies to pictorial visual DRM-tests and that word-only tests does not provide extra useful
information compared to an auditory test (Israel & Schacter, 1997), or only if it’s preceded by a recall test or been asked to scrutinize their memory (Smith, Hunt, and Gallagher, 2008). In this test there is a false memory effect, as well as a false frequency effect, however there is a need for additional tests to compare the effects with auditory tests.
The result of this study shows that if the correct rejections are removed and only the instances of false memory are studied, there is a clear false frequency effect. It does appear that the participants estimate the frequency to be 2 rather than 1. This is interesting
considering numerous studies (eg. Cleary & Greene, 2002; Gallo et al., 2001; Gallo &
Roediger, 2003; Kellogg, 2001; Pierce et al., 2005; Smith & Hunt, 1998; Smith, Lozito, &
Bayen, 2005) have shown that a visual test does lower the false memory effect. This result shows when a false memory is created, there is also a false memory for frequency. The reason why this happens is unclear and it would be interesting to study the reason behind this effect. This is also interesting considering that even though the visual memory is considered by some to be more detailed (Smith & Hunt, 1998; Gallo et al 2001), there is still a false frequency effect.
Björklund & Sandström (2012) suggested that the possible reason behind why a false frequency effect is created when words are repeated more than once is either the
Activation-Monitoring theory or the Fuzzy-trace theory, as described by Arndt & Gold
(2006). According to the Activation-Monitoring theory, the words spread to similar related
words (in this case the critical lures) and the risk of false memories is increased. Similarly to
Björklund and Sandström (2012) a false effect for frequency was noticed. It is a possibility
that the reason for this outcome is based on an error in the encoding phase which leads to the
participants believing a critical lure appeared more than once. Björklund and Sandström
suggested that the theory is supported because participants were certain about the non-critical
lures. The false recognition can according to monitoring processes be based on the fact that
the words were activated during encoding, but the participants did not realise the origin of the
word. This leads to the participants thinking a word appeared based on the strong associations
(Björklund, Sandström, 2012). It is possible this is also the case when it comes to the word-only condition, but perhaps to a lower extent. Activation-Monitoring has shown to reduce the effect of false memories on visual pictorial tests (Schacter, Cendan, Dodson &
Clifford, 2001). However, this test is a word-only test and does not use pictures or drawings the participants can use in the encoding phase. So far there is conflicting information
regarding if word-only tests lead to a significant reduction in false memories compared to auditory tests, with many studies showing a clear reduction (Cleary & Greene, 2002; Gallo et al., 2001; Gallo & Roediger, 2003; Kellogg, 2001; Pierce et al., 2005; Smith & Hunt, 1998;
Smith, Lozito, & Bayen, 2005), a small reduction ( Gallo et al., 2001; Gallo & Roediger, 2003) or no reduction (McKone & Murphy, 2000). There is a need for more research on how word-only tests compare to auditory.
Another suggestion to why this effect happened is the fuzzy trace theory where information is processed on a continuum between verbatim traces and gist traces. The verbatim memory records perceptual features and the gist memory stores more meaning based-information (Ardnt, 2006). This leads to the verbatim memory being better at discriminating false information from true information, while the gist memory stores elaborations of a memory and the commonalities between a set of similar experiences. This leads to a higher risk of false memories (Ardnt, 2012). DRM makes use of the human ability to associate words to other similar related words or experiences, to create false memories.
Similar to how the gist-memory works. The results unsurprisingly suggests that the gist-memory is used when encoding the words of the lists, which then leads to the participants falsely believing the critical lures have appeared.
Both the Activation-Monitoring theory and the Fuzzy-trace theory are probable and should be considered as suggested by Björklund and Sandström (2012).
Were the participants aware of the manipulation?
A majority (20) of the participants noted that the words in the list followed a theme but did not necessarily notice that there was a word that was missing. A third of the
participants (10) mentioned that they either noticed that I was trying to recreate false memories or noticed that there are non-appearing words. If they noticed that some words were non-appearing before or after they had seen the word lists in the test is unclear. The list that were frequently mentioned were “nål” (needle), as they believed the associating words were too clearly alluding to that word. Another word that was mentioned was “tobak”
(tobacco), which is a word many participants mentioned did not show up. One interesting
thing to note is that even though some participants mentioned that it was clear that the lists
were leading them to associate to a certain specific word, several of the words they guessed
as critical lures were incorrect. Knowing what the test is about has shown a reduction in false
memories, but has not eliminated them (Roediger & McDermott, 1998). It is unclear whether
there is a reduction based on this, however there is still a clear false memory effect. Perhaps if
a reduction based on this was the case, the false memory effect could have been higher than
the results have shown.
A few participants (4) mentioned that they reacted more intensely to certain words that were presented. They said they were more emotionally charged and easier to remember.
After the tests, many of the participants expressed that it was easier to remember which words that did not appear compared to guessing how many times a word did appear. This could mean that the participants are guessing their answers rather than actually believing their own estimation. However in this study the area of interest are the frequency in which the participants believed the critical lures appeared, meaning if they appeared more than once despite not appearing at all. It is not of interest whether their estimations on the associated words are correct, as long as there is a false memory effect on the critical lures.
Another thing to note is that when it comes to the participants noting that it was easier to remember the words that did not appear, there is a possibility that it was easier for the participants to remember which words that did appear due to the repetitions of some of the words of each list.
Limitations
One main issue reported by the participants is that the 15 second time limit they had to answer the mathematical questions was too small. This was an issue that was noticed early.
Some participants appeared to have lost the motivation to answer the questions, felt stressed by them and showed signs of agitation when not being able to answer many questions before the time was up. Some mentioned that they noticed that the mathematical part of the test was not an important aspect of my study. However the main purpose of the mathematical
questions was to provide a distraction for the participants between the different lists. Even though the mathematical questions did not provide a distraction in the form that was
originally intended, there is a possibility the stressful factor of a short time limit still had the desired effect of preventing the participants from thinking about the prior list. It is possible however that it is an issue in the cases where the participants lost the motivation to answer the questions, and that a new attempt of this study would need an increased time limit.
Another error that surfaced was when replacing the lists Hammare (hammer) and Matta (carpet) for the lists Lång (tall) and Söt (cute), one word from the list, Tor (Thor), associated to Hammare was not replaced with a word from Söt in test nr 2. This means that there was an uneven amount of associated words in test nr 1 compared to test nr 2. Since Tor was not associated with the rest of the words that appeared in the lists, it functions as an extra non-critical lure.
A third limitation is the number of participants for this study. The ideal number of participants for this study would have been a minimum of 36. The reason for that number is to be able to balance the test conditions. It was difficult to find participants due to time limits and students’ commitments.
There is a limitation regarding the structure of the confidence-judgement where the
participants were asked to answer how certain they were about their own judgement
regarding how many times a word appeared. By using a single confidence-judgement this
way there is a risk the participants misinterpreted the instructions as being “answer how
confident you were that the word appeared” instead of how certain they were about their estimation of the frequency. Several participants expressed that it was easier to determine whether a word appeared or not, but not how many times they appeared. Despite this several participants answered that they were very sure about their estimations. If this estimation is based on whether they think the word appeared or if they are certain about their own estimation is unclear. A solution to this could be to have two confidence-related questions, one regarding whether the word appeared or not, and another about their certainty of their estimation.
The participants were asked to answer whether they noticed something strange about the test. Several participants mentioned noticing that certain words did not appear. It is unclear whether the participants are referring to the non-critical lures or the critical lures. As the purpose of the non-critical words are too unrelated to the associated words, it is not surprising that the participants noticed that the words stood out. It is however an issue since it is difficult to tell which participants were aware of the manipulation attempt, as was the purpose of the question.
Conclusion
A false frequency effect for false memories was detected in a visual word-only
DRM-test. Compared to the results by Björklund and Sandström (2012) where they tested the false memory for frequency on an auditory version of the DRM-paradigm, there was a
significant result on frequency on a visual word-only version. Provided that there was a false memory effect, the results showed that its estimated presentation frequency was more likely to be greater than one.
As this to my knowledge is the second study studying memory in frequency, the research area is limited. There is much more research to be done in this area. This study is the first to use DRM visually in a frequency experiment. This means that the categories
associative strength is based on the lists as described by Björklund & Sandström (2012) and
Johansson & Stenberg (2002). There is a high possibility the associative strength would differ
between an auditory and a visual test. It would be interesting to study which lists would
provide a high false memory effect in a visual word-only test and then reattempt this study
using those lists. It would be interesting to make wordlists more suited for a visual version of
the DRM-test, using only the highest associative lists - to see if that would change the results
as well as comparing the results with an auditory test.
References
Akhtar, S., Justice, L. V., Morrison, C. M., & Conway, M. A. (2018). Fictional First Memories. Psychological Science, 29(10), 1612–1619. https://doi.org/10.1177/0956797618778831
Arndt, J., & Hirshman, E. (1998). True and False Recognition in MINERVA2: Explanations from a Global Matching Perspective. Journal of Memory and Language, 39(3), 371–391. https://doi.org/10.1006/jmla.1998.2581 Arndt, J., & Gould, C. (2006). An examination of two-process theories of false recognition. Memory, 14(7), 814–833.
https://doi.org/10.1080/09658210600680749
Arndt, J. (2012). The influence of forward and backward associative strength on false recognition. Journal of Experimental Psychology: Learning, Memory, and Cognition, 38, 747-756.
Baioui, A., Ambach, W., Walter, B., Vaitl, D., & Alain, C. (2012). Psychophysiology of False Memories in a
Deese-Roediger-McDermott Paradigm with Visual Scenes (Psychophysiology of Visual False Memories). PLoS ONE, 7(1), e30416. https://doi.org/10.1371/journal.pone.0030416
Björklund R., Sandström E. Falska minnen och frekvens i DRM-paradigmet: Leder upprepning av enskilda ord i en ordliska till falska frekvensbedömningar? (master’s thesis) (2012). Retrieved from Ulrich Olofsson at Umeå University.
Brainerd, C. ., Payne, D. ., Wright, R., & Reyna, V. . (2003). Phantom recall. Journal of Memory and Language, 48(3), 445–467. https://doi.org/10.1016/S0749-596X(02)00501-6
Butler, A., & Roediger, C. (2008). Feedback enhances the positive effects and reduces the negative effects of multiple-choice testing. Memory & Cognition, 36(3), 604–616. https://doi.org/10.3758/MC.36.3.604
Cochran, K., Greenspan, J., Bogart, R., & Loftus, L. (2016). Memory blindness: Altered memory reports lead to distortion in eyewitness memory. Memory & Cognition, 44(5), 717-726.
Cleary, A. M., & Greene, R. L. (2002). Paradoxical effects of presentation modality on false memory. Memory, 10, 55-61.
Deese, J. (1959). On the prediction of occurrence of particular verbal intrusions in immediate recall. Journal of Experimental Psychology, 58, 17-22.
Fábio Rodrigo Bourscheid, Luciano Haussen Pinto, Luiza Feijó Knijnik, & Lilian Milnitsky Stein. (2014). Falsas Memórias e o Paradigma DRM: Uma Abordagem por Meio de Fotos Emocionais Associadas. Psicologia: Teoria e Pesquisa, 30(2), 163–170. https://doi.org/10.1590/S0102-37722014000200005
Gallo, D. A., McDermott, K. B., Percer, J. M., & Roediger, H. L., III (2001). Modality effects in false recall and false recognition. Journal of Experimental Psychology: Learning, Memory, & Cognition, 27, 339-353
Gallo, D. A., & Roediger, H. L., III (2003). The effects of associations and aging on illusory recollections. Memory &
Cognition, 31, 1036-1044
H. L. (2001). Modality Effects in False Recall and False Recognition. Journal of Experimental Psychology: Learning, Memory, and Cognition, 27(2), 339–353. https://doi.org/10.1037/0278-7393.27.2.339
Hall, L., Johansson, P., Tärning, B., Sikström, S., & Deutgen, T. (2010). Magic at the marketplace: Choice blindness for the taste of jam and the smell of tea. Cognition, 117, 54–61. doi:10.1016/j.cognition. 2010.06.010
Israel, L., & Schacter, D. (1997). Pictorial encoding reduces false recognition of semantic associates. Psychonomic Bulletin
& Review, 4(4), 577–581. https://doi.org/10.3758/BF03214352Gallo, D. A., McDermott, K. B., Percer, J. M., &
Roediger,
Johansson, M., & Stenberg, G. (2002). Inducing and reducing false memories: A Swedish version of the Deese–Roediger–McDermott paradigm. Scandinavian Journal of Psychology, 43(5), 369-383.
Kellogg, R. T. (2001). Presentation modality and mode of recall in verbal false memory. Journal of Experimental Psychology: Learning, Memory, & Cognition, 27, 913-919
Kimball, D. R., & Bjork, R. A. (2002). Influences of Intentional and Unintentional Forgetting on False Memories. Journal of Experimental Psychology: General, 131(1), 116–130. https://doi.org/10.1037/0096-3445.131.1.116
Lindsay, D. S., Read, J. D., Grossman, Lisa R., & Pressley, Michael. (1994). Psychotherapy and memories of childhood sexual abuse: A cognitive perspective. Applied Cognitive Psychology, 8(4), 281–338.
https://doi.org/10.1002/acp.2350080403
Loftus, E. F. (1979). Eyewitness testimony. Cambridge: Harvard University Press
Loftus, E. F. (1993a). The reality of repressed memories. American Psychologist, 48, 518–537.
Loftus, E. F. (2005). Planting misinformation in the human mind: A 30-year investigation of the malleability of memory.
Learning & Memory, 12, 361–366
Loftus, E. F., & Pickrell, J. E. (1995). The formation of false memories. Psychiatric Annals, 25, 720–725
Maylor, E. A., & Mo, A. (1999). Effects of study-test modality on false recognition. British Journal of Psychology, 90, 477–493.
Mcconnell, M., & Reed Hunt, D. (2007). Can false memories be corrected by feedback in the DRM paradigm? Memory &
Cognition, 35(5), 999–1006. https://doi.org/10.3758/BF03193472
McDermott, K. B., & Roediger, H. L. (1998). Attempting to avoid illusory memories: Robust false recognition of associates persists under conditions of explicit warnings and immediate testing. Journal of Memory and Language, 39, 508-520.
McKone, E., & Murphy, B. (2000). Implicit false memory: Effects of modality and multiple study presentations on long-lived semantic priming. Journal of Memory & Language, 43, 89-109
Miller, M. B., & Gazzaniga, M. S. (1998). Creating false memories for visual scenes. Neuropsychologia, 36, 513–520.
Pashler, H., Cepeda, N. J., Wixted, J. T., & Rohrer, D. (2005). When Does Feedback Facilitate Learning of Words? Journal of Experimental Psychology: Learning, Memory, and Cognition, 31(1), 3–8.
https://doi.org/10.1037/0278-7393.31.1.3
Peeck, J., van Den Bosch, A. ., & Kreupeling, W. . (1985). Effects of informative feedback in relation to retention of initial responses. Contemporary Educational Psychology, 10(4), 303–313.
https://doi.org/10.1016/0361-476X(85)90028-1
Phye, G. D., & Bender, T. (1989). Feedback complexity and practice: Response pattern analysis in retention and transfer.
Contemporary Educational Psychology, 14(2), 97–110. https://doi.org/10.1016/0361-476X(89)90028-3 Pierce, B. H., Gallo, D. A., Weiss, J. A., & Schacter, D. L. (2005) The modality effect in false recognition: Evidence for
test-based monitoring. Memory & Cognition, 33, 1407-1413.
Roediger, H. L., & McDermott, K. B. (1995). Creating false memories: Remembering words not presented in lists. Journal of Experimental Psychology: Learning Memory and Cognition, 21, 803-814.
Russell, W. A., Jenkins, J. J., & University of Minnesota. (1954). The complete Minnesota norms for responses to 100 words from the Kent-Rosanoff word association test. Minneapolis: Dept. of Psychology, University of Minnesota.
Seamon, J. G., Luo, C. R., & Gallo, D. A. (1998). Creating false memories of words with or without recognition of list items:
Evidence for nonconscious processes. Psychological Science, 9, 20–26.
Seamon, J. G., Luo, C. R., Shulman, E. P., Toner, S. K., & Caglar, S. (2002). False memories are hard to inhibit: Differential effects of directed forgetting on accurate and false recall in the DRM procedure. Memory, 10(4), 225–237.
https://doi.org/10.1080/09658210143000344
Schacter, D. L., Cendan, D. L., Dodson, C. S., & Clifford, E. R. (2001). Retrieval conditions and false recognition: Testing the distinctiveness heuristic. Psychonomic Bulletin & Review, 8, 827-833.
Schacter, D. L., Israel, L., & Racine, C. (1999). Suppressing false recognition in younger and older adults: The distinctiveness heuristic. Journal of Memory & Language, 40, 1-24.
Smith, R. E., & Hunt, R. R. (1998). Presentation modality affects false memory. Psychonomic Bulletin & Review, 5, 710-715
Smith, R., Hunt, E., & Gallagher, R. (2008). The effect of study modality on false recognition. Memory & Cognition, 36(8), 1439–1449. https://doi.org/10.3758/MC.36.8.1439
Smith, R. E., Lozito, J. P., & Bayen, U. J. (2005). Adult age differences in distinctive processing: The modality effect in false recall. Psychology & Aging, 20, 486-492.
