Nonbelieved Memories in The False Memory Archive

Abstract

The False Memory Archive is a unique art collection containing hundreds of false memory
reports submitted by members of the general population. The current study aimed to analyse these reports. Specifically, we examined whether some of the memories reported in these submissions were better described as nonbelieved memories (NBMs). Furthermore, we investigated the reasons for why people decided that their memory was false and assessed the verification strategies that people used to validate their mental representation. Five hundred submissions were coded and more than half (53.4%) met the criteria for NBMs. Social feedback was the most frequently reported reason for reducing belief and asking family members was the most frequently mentioned memory verification strategy. Reports categorized as NBMs were more likely to include mention of memory verification strategies than were believed memories.

Full text

NONBELIEVED MEMORIES
1
Nonbelieved Memories in The False Memory Archive
Henry Otgaar1,2,3, Charlotte Bücken1, Glynis Bogaard1, Kimberley A. Wade4, A.R. Hopwood,
Alan Scoboria5 & Mark L. Howe3
1Maastricht University, The Netherlands
2Faculty of Psychology, Universitas Indonesia
3City, University of London, UK
4University of Warwick, UK
5University of Windsor, Canada
In Press, Journal of Applied Research in Memory and Cognition
Correspondence should be sent to: Henry Otgaar, Henry.Otgaar@maastrichtuniversity.nl,
Faculty of Psychology and Neuroscience, section Forensic Psychology, Maastricht University.
We dedicate the current work to Alan Scoboria, a brilliant memory researcher and dear friend,
who passed away too soon.

NONBELIEVED MEMORIES
2
Abstract
The False Memory Archive is a unique art collection containing hundreds of false memory
reports submitted by members of the general population. The current study aimed to analyse
these reports. Specifically, we examined whether some of the memories reported in these
submissions were better described as nonbelieved memories (NBMs). Furthermore, we
investigated the reasons for why people decided that their memory was false and assessed the
verification strategies that people used to validate their mental representation. Five hundred
submissions were coded and more than half (53.4%) met the criteria for NBMs. Social feedback
was the most frequently reported reason for reducing belief and asking family members was the
most frequently mentioned memory verification strategy. Reports categorized as NBMs were
more likely to include mention of memory verification strategies than were believed memories.
Keywords: False Memory; Nonbelieved Memory; Memory Verification

NONBELIEVED MEMORIES
3
General Audience Summary
False memories or memories for events/details that were not experienced can occur to everyone.
As part of an artwork called the False Memory Archive, the artist -Alasdair Hopwood- has been
collecting false memory reports from the general population. People who submitted these reports
were aware that these memories were false. We took a closer look at a random sample of these
statements and found that the majority could be categorized as nonbelieved memories.
Nonbelieved memories are memories for events of which people have reduced the belief that the
event actually occurred. Furthermore, we found that such nonbelieved memories were mainly the
result of other people telling the submitters that their memory was incorrect. Our results show
that belief-memory dissociations are quite normal in the general population.

NONBELIEVED MEMORIES
4
Nonbelieved Memories in The False Memory Archive
Scientific discoveries frequently inspire artists to develop creative artwork. In the current
paper, we do the reverse by presenting a scientific study that was inspired by art. We examined
an artwork called the False Memory Archive developed by the London-based artist Alasdair
Hopwood. The artwork contains a unique assemblage of vivid personal memories of events that
the contributors -the general public- claim never happened. Our aim was to describe the reasons
for why people decided that their memories might be false, and the verification strategies people
use to validate their memories.
When we retrieve a memory for an event, we generally also hold a strong belief that this
event occurred. However, studies have revealed a counterintuitive class of memories for events
where belief that the remembered event occurred is reduced or absent called nonbelieved
memories (NBMs; for a review, see Otgaar, Scoboria, & Mazzoni, 2014). Despite retaining vivid
mental representations of the event accompanied by feelings of re-experiencing, people report a
reduction in the belief that the remembered event actually occurred. NBMs might have started
out as false memories which are correctly rejected later, called refuted memories. However,
NBMs can also be incorrectly rejected true memories, called disowned memories (Mazzoni,
Scoboria, & Harvey, 2010).
Submissions to the False Memory Archive formed part of an international art project. The
artist, Alasdair Hopwood, was inspired by research demonstrating the malleability of memory.
Based on this research, Hopwood worked in collaboration with several international memory
researchers to develop a series of exhibitions that contain objects, text, videos, and photography,
as well as a written collection of personal accounts of experiences that never occurred. One aim
of the archive is to show the public how the past is constantly reconstructed which might lead to
“humorous, obscure and uncomfortable things people have misremembered” (Hopwood,

NONBELIEVED MEMORIES
5
http://www.falsememoryarchive.com). Visitors to the False Memory Archive exhibitions were
invited to describe a memory of an event of which they now know never happened. As such, the
False Memory Archive provides an ideal opportunity to analyze potential NBM reports from the
general population. One key advantage of analyzing these memory reports is that individuals
who contributed to the Archive were not aware that they were taking part in research. In most
prior studies of false memories and NBMs, individuals have been aware and were compensated
for their participation. Such awareness may have affected the information that these participants
chose to provide. For example, participants might have been more willing to tell a story -true or
false- than submitters of the False Memory Archive.
In the current study, we analyzed reports submitted to the False Memory Archive. By
gaining a better understanding of false memories and NBMs in the real world, we can help to
advance theories of memory. Specifically, we examined whether the reports submitted to the
archive referred to false memories and/or to NBMs, and examined the reasons that people
provided as to why belief in these memories was retracted.
Studies on Nonbelieved Memories
In the first empirical study of NBMs, Mazzoni and colleagues (2010) found that 20% of
those asked were able to retrieve a NBM. However, participants were explicitly cued to do so; a
procedure that might lead to an overestimation of the frequency of NBMs in the general
population. Scoboria and Talarico (2013) examined how frequently NBMs would occur without
participants being directly cued to report them. In three studies, participants retrieved
autobiographical experiences after which these were rated on belief in occurrence (i.e., truth
value attributed to an event) and recollection. They found that 3% to 3.8% of participants
reported autobiographical memories that scored higher on recollection than belief in occurrence
which they classified as NBMs.

NONBELIEVED MEMORIES
6
Besides retrospective accounts of naturally occurring NBMs (see also e.g., Brédart &
Bouffier, 2016), researchers have also elicited NBMs experimentally. Otgaar, Scoboria, and
Smeets (2013) falsely told children and adults that they experienced a childhood hot air balloon
ride. During two interviews, a significant minority of participants (36%) indicated that they came
to remember this (false) event. Crucially, after the last interview, participants were informed that
the suggested event never actually occurred, and participants’ belief in the occurrence and their
memory for the false event were measured. Forty percent of those who reported remembering
the false event claimed to have a NBM after the debriefing.
Other studies have shown that NBMs can be experimentally created using a variety of
methodologies including doctored video clips of fake actions (Clark, Nash, Fincham, &
Mazzoni, 2012). When informed about the false actions, participants reported decreases in belief
in occurrence while maintaining a strong sense of recollection. Although this study concentrated
on inducing NBMs for false (non-performed) actions, studies have shown similar findings when
challenging memories for true experienced events (e.g., Mazzoni, Clark, & Nash, 2014; Otgaar,
Scoboria, Howe, Moldoveanu, & Smeets, 2016; Scoboria, Otgaar, & Mazzoni, 2018). Together,
the research to date shows that NBMs are not uncommon in the general population and can be
elicited and manipulated experimentally.
Reasons to Reduce Belief
Previous work has examined reasons why people decide to withdraw belief in the
occurrence of remembered events. Mazzoni et al. (2010) identified three types of reasons. The
first and most frequently mentioned reason involved other people telling the participants that the
memory was incorrect. For example, a sibling might have stated that the event actually happened
to another family member. The second category referred to events being too implausible to have
actually happened (e.g., recalling seeing a living Dinosaur). The last category concerned

NONBELIEVED MEMORIES
7
contradictory evidence regarding the memory such as discovering a photograph challenging
whether a certain event was indeed experienced (see Scoboria, Boucher, & Mazzoni, 2014 for
similar results).
Theoretically, research on the reasons for reducing belief in occurrence parallels
strategies that people use to verify memories that have been brought into question (Wade &
Garry, 2005; Wade, Nash, & Garry, 2014; Nash, Wade, Garry & Adelman, 2017). These studies
indicate that people evaluate the costs and reliability when choosing how to verify personal
memories and tend to prioritize lower cost (in terms of the effort required to pursue a memory
verification strategy) over reliability when picking a strategy. In these studies, participants
primarily reported that they would rely on other people to validate memories. The chief motive
for gleaning information from others is that relying on other people is a relatively cheap and easy
way to verify one’s memories. This might also clarify why social feedback is the primary reason
why people reduce their belief in the occurrence of events. In contrast, receiving or finding
nonsocial external evidence is less often reported as a verification strategy. Albeit a potentially
more reliable source, searching for nonsocial external evidence such as legal documents or
photographs requires more time and energy, making it less likely that people use this strategy
and hence it is less likely to play a role in belief withdrawal.
Method
Sample
Hopwood has been collecting false memory submissions online (via
https://www.arhopwood.com/fma) and in six European museums: [1] The Mead Gallery at the
University of Warwick, UK; [2] the Talbot Rice Gallery at the University of Edinburgh, UK; [3]
the Newlyn Art Gallery, UK; [4] The Exchange, UK; [5] The Freud Museum, London, UK; [5]
Carroll/Fletcher London; and [6] the Schunck Museum, Heerlen, Netherlands). Parts of the work

NONBELIEVED MEMORIES
8
have also been displayed at Warwick Arts Centre in the UK, ADM gallery in Singapore, and the
French Cultural institute in Boston, US. Total visitors to events and exhibitions on the UK tour
was 18191 visitors in 213 days of exhibition time.
For the current study, we only analyzed English submissions that were collected from
exhibitions in the UK and online. At the time of the current analysis, a total of 805 submissions
had been collected. Contributors to the Archive received the same instructions, regardless of
whether they were submitting online or in-person. Specifically, they were told:
“We are collecting false memories for a False Memory Archive. You can
anonymously submit a false memory by using the form overleaf or by going to:
falsememoryarchive.com. The false memory can be your own or it can belong
to someone else.”
Although the instructions were the same, the context was different. During the
exhibitions, people received more information about the phenomenon of false memories
than people who submitted their false memory report online.
The accounts in the False Memory Archive vary dramatically, ranging from impossible
memories of pre-birth experiences to possible memories of, for example, holidays. This
collection lends itself perfectly to the study of NBMs because the submissions, according to the
archive, follow a pattern that is linked to research on the reasons for why people reduce belief.
Specifically, according to the website of the False Memory Archive (see
https://www.falsememoryarchive.com/anthology): “a memory is described, only to be undone by
evidence that the recollection is faulty or by a suspicion that the experience never actually
occurred.” We were only interested in personal submissions and excluded submissions referring
to other people’s memories. We randomly selected 500 submissions to code for our analyses. All
coded responses are available on the Open Science Framework (https://osf.io/nk54r/).

NONBELIEVED MEMORIES
9
Demographic information for those who submitted reports to the False Memory Archive is
unknown. However, there are some general demographic data on people visiting museums made
within the past year. In England, in 2017/2018, about half of the population (from all age-groups)
visited a museum or gallery (age range 16-74; 48.1% of 16-24 year olds, 53.2% of 25-44 year
olds, 52.3% of 45-64 year olds, 52.7% of 65-74 year olds; Statista, 2019). American data show
that most museum visitors are white and well-educated (Farrell & Medvedeva, 2010).
Predictions
Our predictions were as follows. We anticipated that the majority of false memory
submissions could be categorized as nonbelieved memories. This prediction is based on the idea
that if people submit a false memory, they are aware that they at some point believed that the
memory was genuine and then at some point their belief that the event occurred was reduced.
Also, in line with previous findings (Scoboria et al., 2015), we predicted that social feedback
would be the most frequently reported category for why participants decided that their memory
was false. Finally, based on previous work (Ost, 2017; Wade et al., 2014), we expected that
asking family members would be the most reported memory verification strategy, and we
explored whether previous categorizations of verification strategies fully captured the range of
strategies present in the archive or whether previously undocumented memory verification
strategies emerged.
Coding
Submissions were coded using an adapted coding scheme based on Scoboria et al.
(2017) and Ost (2017). The coding scheme contained three main components. First, the scheme
described how to code reports on whether the submitter believed and/or recollected the event.
Second, the scheme described how to code the reasons for why submitters reduced belief in the

NONBELIEVED MEMORIES
10
occurrence of events. Finally, details were provided to judge the presence of memory verification
strategies.
More specifically, we first determined the extent to which each contributor appeared to
believe and recollect the event described in the report. Categories were taken from Scoboria,
Boucher et al. (2015), Scoboria et al. (2017), and Ost (2017), and some were slightly changed for
the current investigation (i.e., we did not include all subcategories of social feedback from
Scoboria et al.). Categories included (1) Judged memory, (2) Judged belief, (3) Explicit
statement of having a memory, (4) Explicit statement of having no memory, (5) Explicit
statement of having a belief in occurrence, (6) Explicit statement of having no belief in
occurrence, and (7) Acceptance of Events. From these categories the quality of the submissions
in terms of memory and belief in occurrence was inferred, forming the categories of (1) Non-
believed memory (NBM; stating memory, stating no belief in occurrence), (2) Believed memory
(BM; stating memory, stating belief in occurrence), (3) Believed-not-remembered events (BNRs;
stating no memory, stating belief in occurrence), and (4) Non-believed-not-remembered events
(NBNRs; stating no memory, stating no belief in occurrence).
The following guidelines were used to categorize submissions: an event was judged to be
recollected (Judged memory = ‘Yes’) based on the presence of a positive and lack of negative
memory statement about the event, as well as the coder’s general impression that the memory
was associated with vivid recollection. Moreover, the coding of ‘acceptance’ was used as a
guideline for when belief in occurrence was or was not present. The extent to which an event was
accepted (to have occurred) was coded on a scale of outright rejection (0) to full acceptance (3).
The status of the report was based on whether any reports were present regarding belief in
occurrence, other information about the truth status, and the coder’s judgement. Reports were

NONBELIEVED MEMORIES
11
defined as believed to have occurred if the acceptance score was 2 or higher on the scale. Table 1
summarizes which combination of scores led to which memory-belief category.
Table 1
Memory-belief Cross-classifications
Judged memory: Yes (2)
Judged memory: No (0)
Acceptance of events: High (2/3)
Believed memory (1)
Believed-not-remembered event (3)
Acceptance of events: Low (0/1)
Nonbelieved memory (2)
Non-believed-non-remembered event (4)
In the second step of coding, we investigated the reduction or withdrawal of belief. Here
only submissions coded as NBMs and NBNRs were considered, since these represent all
submissions for which belief in occurrence decreased. The categories from Scoboria, Boucher et
al. (2015) were included complemented by two additional categories – the ‘Change of Context’
and ‘Noticing Deterioration’ categories identified by Ost (2017). Thus, the categories coded for
reasons for belief reduction were: (1) Social feedback [including the subcategories: a) being told
the event did not occur, b) being told the event happened differently, c) being told that the event
is impossible, d) being told the event happened to someone else, e) being told the event is
unlikely, f) disconfirming non-verbal feedback, g) being told that the submitter was not present
at the event, h) lack of corroboration, i) others unavailable (e.g., important other people are not
available to confirm the event), j) the submitter was pressured by another/others to stop believing
in the memory, k) another/others refused to discuss the event, l) other social feedback]; (2) Event
plausibility [a) general event plausibility (refers to how objectively plausible the event is to
happen in general, not taking into account information that applies only to the submitter

NONBELIEVED MEMORIES
12
specifically), b) specific event plausibility (refers to reports by the submitters as to how plausible
the event is for the submitter specifically, in his/ her life; e.g., if someone submitted a memory of
fighting with a sibling, this report would be generally plausible. However, if the person who
submitted this report did not have any siblings, then this event would not be plausible for this
person specifically)]; (3) Alternative attributions [a) internal/ asleep, b) internal/awake, c) other
mental state, d) external]; (4) General Beliefs about Memory [a) memory and age, b) about the
integrity of memory, c) about the ongoing influence of memory]; (5) Internal memorial
characteristics (qualities associated with the mental representation for the event); (6) Notions of
self/others [a) memory is incompatible with self-image or image of others, b) memory is
incompatible with image of others]; (7) External evidence [a) disconfirmatory evidence obtained,
b) confirmatory evidence not obtained]; (8) Personal motivation to reduce belief for the memory;
(9) Change of Context (refers to cases in which a change of social situation, location, or medical
treatment induced a change in belief), (10) Noticing deterioration (refers to cases in which
submitters realize that psychological treatment/counselling they are receiving is not leading to
improvement, leading to a reduction in belief for memories discussed during therapy), (11)
Other, and (12) No information regarding reasons for withdrawal of belief given (see Table 2 for
definitions of each reason). The lack of explanation for why belief was retracted in some
submissions created the need to include the final category; this is one important way in which the
convenience sample in the False Memory Archive differs from studies of NBMs in which the
method typically involves asking people to describe reasons for reducing belief.
In the third step, the coding scheme developed by Ost (2017) and the categories reported
by Wade and Garry (2005) were used to identify memory verification strategies that were
mentioned in submissions. Additionally, a check of whether a verification attempt was reported
in general was included before coding the subcategories. This was deemed necessary since not

NONBELIEVED MEMORIES
13
all submitters reported attempting verification. Therefore, the following categories were
included: (0) General verification, if yes: (1) Checking with family, (2) Checking with others, (3)
Searching for additional cues, (4) Searching for physical evidence, and (5) Cognitive techniques
(see Table 3 for definitions of each strategy).
The second author coded the reports using this coding scheme. The full coding scheme
including all details on coding definitions and categories, a list of abbreviations, and a list of the
decision rules used are available on the Open Science Framework (https://osf.io/nk54r). A
second rater coded the memory/belief quality of 50 reports. Interreliability was substantial:
Cohen’s kappa = 0.69 (Landis & Koch, 1977).
Table 2
Reasons for Belief Withdrawal
Reason for Belief Withdrawal
Definition/ Explanation
Social Feedback (SF)
Exchanges with other people lead to
invalidating information and this feedback is
the reason to reject that the remembered
events occurred
SF1
Feedback that the remembered events did not
occur
SF2
Feedback that the remembered events
happened differently
SF3
Feedback that the remembered events are
impossible

NONBELIEVED MEMORIES
14
SF4
Feedback that the remembered event(s)
happened to someone else
SF5
Feedback that the occurrence of the
remembered event(s) is unlikely
SF6
Non-verbal disconfirming feedback that the
memory might be false
SF7
Feedback that the submitter was not present at
the recollected event
SF8
Feedback that the memory cannot be
corroborated/ confirmed by others
SF9
Other people important to the remembered
event(s) are not available to give feedback
SF10
The submitter was pressured by others to stop
believing in the memory and starts to mistrust
that person
SF11
Others refuse to discuss the events with the
submitter
SF12
Other disconfirming social feedback was
obtained
General Event Plausibility
The occurrence of the event(s) is not possible
in general, not taking the specific position of
the submitter into account

NONBELIEVED MEMORIES
15
Specific Event Plausibility
The occurrence of the event(s) is not possible
for the submitter specifically, even if in
general these events could have happened to
someone else
Alternative Attributions (AA)
The recollection is allocated to an origin in a
source other than real life experience
AA1
Events attributed to internal, mental images
while asleep (e.g., dreaming)
AA2
Events attributed to an internal source while
awake (e.g., imagination)
AA3
Events attributed to fabrication while in
another mental state (e.g., hallucination,
intoxication, exhaustion)
AA4
Events attributed to confusion between
external sources and experiences (e.g., books,
TV)
General Beliefs (GB)
Changed belief due to general beliefs about
memory and memory ability
GB1
Beliefs about memory during childhood (e.g.,
memory cannot occur before a certain age,
childhood memories are unreliable)
GB2
Beliefs about the integrity of memory (e.g.,
memory can be false)

NONBELIEVED MEMORIES
16
GB3
Beliefs about influence of memory (e.g.,
memories should have an ongoing influence
on behavior)
Internal Features of Event Representation
Internal features such as sensations, images,
emotions are abnormal for a memory
External Evidence (EE)
Found external evidence (not from a social
exchange) indicating that the recollected
event(s) could not have occurred
EE1
Evidence is found that disconfirms the
memory events
EE2
Evidence to confirm the memory cannot be
found
Notions of Self/ Other (NSO)
Occurrence of event(s) is incompatible with
submitter’s self-concept or image of another
person
NSO1
Occurrence of event(s) is incompatible with
the submitter’s self-image
NSO2
Occurrence of event(s) is incompatible with
the submitter’s image of another person
Personal Motivation
Personal motivation to invalidate the
recollection for some self-benefit
Change of Context
Change in belief due to change in social
situation or location

NONBELIEVED MEMORIES
17
Noticing Deterioration
Change of belief after submitter noticed that
treatment they are getting is not improving
their condition, so that memories from during
the treatment time are questioned
Other
Belief is changed due to some other reason
not listed above
No information
No information was given to indicate why
belief was withdrawn
Table 3
Verification Strategies
Verification Strategy
Explanation/ Definition
General Attempt at verification
The submitter actively tried to verify their
memory (e.g., by actively searching for more
information)
VS1
Asked family members whether details/ or the
complete remembered event(s) are true
VS2
Asked people other than family members
whether details/ or the complete remembered
event(s) are true (e.g., friends, acquaintances)

NONBELIEVED MEMORIES
18
VS3
Sought out locations or cues to cue recall and
verify their memory (e.g., returning to the
event location)
VS4
Searched for physical evidence compatible
with the memory (e.g., scars, newspaper
articles)
VS5
Purposefully used cognitive techniques (e.g.,
trying to remember more of the event) to
verify memory events
Results
Nonbelieved Memories
As predicted, more than half of the submissions were categorized as nonbelieved
memories (n = 267, 53.4%). An example of a NBM from the archive is: “I remember getting lost
in a national park as a 6 year old child. I even remember the conversation with my parents.
There was a hill, I said I'd go round one way and they'd go the other and we'd meet on the other
side. Only when I reached the other side they weren't there […]. I was picked up by a park
ranger, rode in the back of his open back truck. Was taken to a shop[…]. I remember hearing
notice of a lost child going over the tannoy for my parents to come and pick me up, which they
did. I remembered and believed all this for over 30 years until one day I asked my parents if they
remember it. They both […] swore blind that it never happened. I believe them, they would
definitely remember something like that […] I now believe I'd dreamt it.”
A large percentage of the reports were coded as believed memories (n = 208, 41.6%). An
example is the following: “When I was approximately 6 years old, I was visiting my older

NONBELIEVED MEMORIES
19
cousin’s house. I recall brawling with him in his room with my older sister watching from the
bunk bed. I then had both of my front teeth knocked out of my mouth. His mother then told me I
would get £2 from the tooth fairy etc. and I placed my teeth near a sink, which they later fell
down/went missing; and I was given £2. Later I was told by my parents they don’t remember it
[…]. [My cousin and aunt do] not remember either. However, to this day I believe it’s real […]”.
A minority of reports were coded as believed-not-remembered events (n = 11; 2.2%) or not-
believed-not-remembered events (n = 14, 2.8%).
Figure 1. Reasons for Belief Reduction
Reasons for Belief Reduction
To analyze the reasons for belief reduction, we focused on memory reports in which
belief was reduced or relinquished and hence focused on nonbelieved memories or non-believed-
not-remembered events (n = 281). As expected, and consistent with Scoboria, Boucher and
Mazzoni (2014) social feedback was the most frequently mentioned category (n = 94, 33.5%; see
Figure 1). Also consistent with their study, within the social category being told that the event
0
5
10
15
20
25
30
35
40
Social-feedback
Event-plausibility
Alternative-attributions
General-beliefs
Internal-features
Notions-of-self-or-others
External-evidence
Personal-motivation
Change-of-context
Percentage-of-reasons-for-belief-
reduction

NONBELIEVED MEMORIES
20
did not occur was the most frequently mentioned sub-category, 44.7% (42/94); n = 42. For
example, one individual wrote: “I can vividly remember attending a wedding, aged around 5,
where there was a grape juice fountain built out of stone in the corridor. [...] I have always
counted this as one of the weddings I had been to until a couple of years ago (aged about 25) my
mum said it had never happened.”.
Other reasons were the following (most to least frequent): Alternative attributions was
stated as a reason in 19.6% (n = 55) of cases, specific event plausibility in 12.8% (n = 36) of
cases, external evidence in 11% (n = 31) of cases, general event plausibility in 10.7% (n = 30) of
cases, general beliefs in 2.8% (n = 8) of cases, internal features in 2.1% (n = 6) of cases, notions
of self and others in 1.4% (n = 4) of cases, change in context in 0.1% (n = 2) of cases, and
personal motivations in 0.4% (n = 1) of cases. Of those who reduced belief due to general beliefs
(n = 8), 87.5% (n = 7 out of 8) did so because of general beliefs about the functioning of memory
(e.g., “[...] I remember my mum bringing home my baby brother from [the] hospital when I was
18 months old, and me prodding him in his baby basket. I remember it was in the living room
and where he was placed on the floor, but I was too young to remember this happening [...]”)
and 12.5% (n = 1 out of 8) due to general beliefs about memory integrity.
For some of the primary categories, several subcategories were identified. For example,
for the alternative attributions category, 52.7% (n = 29 out of 55) of submitters indicated
attributing the memory to an internal source while asleep (e.g., dreaming, “When I was about 10
I had a memory of having my arm stitched at a local hospital [...]. Thinking about it, it must have
been a very vivid dream that I confused with ”reality”) while 32.7% (n = 18) reduced belief
because they attributed the memory to an external source (e.g., movie, photograph). Also, 14.5%
(n = 8) viewed their memory as originating from an internal source while being awake (e.g.,
imagination, “[...] A little later one of the kids reported seeing someone in the woods near the

NONBELIEVED MEMORIES
21
cabin. A teacher got into the minibus, put the headlights on and slowly drove towards the woods
where this person had supposedly been. I remember walking alongside the minibus holding onto
the wing-mirror. Suddenly a man in a Halloween mask jumped out from behind a tree 10 feet in
front of me, screamed, then ran off. [...] I can only think now that I'd imagined [...] the whole
thing”) and one person (1.8%) ascribed their memory to another mental state (intoxication).
With regards to retracting a belief because of nonsocial external evidence, 83.9% (n = 26)
did so because they received nonsocial external evidence disconfirming the authenticity of their
memory (e.g. legal documents, pictures) and 16.1% (n = 5) of the people retracted belief because
they failed to find any evidence confirming memory (e.g., lack of scars, “I thought I remembered
putting a garden fork through my foot [...]. My memory was attempting to use the fork to dig but
accidentally striking my wellington boot and into my foot. However, I have no scars! [...]”).
Comparison with Scoboria et al. (2015) and Ost (2017). To examine whether our data
on the reasons for belief reduction mirrored previous work, we compared our observed
percentages with those found in Scoboria et al. (2015) and Ost (2017; see Table 5). The most
notable result was that social feedback was the most frequently mentioned reason for why people
changed belief.
Table 5
Reasons for Belief Retraction
Reasons for Belief
Retraction
Scoboria et al.
(2015)
Ost (2017)
False Memory
Archive
Social Feedback
53.0% (n = 198)
19.6% (n = 31)
33.5% (n = 94)
Event Plausibility
35.0% (n = 132)
1.8% (n = 3)
23.0% (n = 66)
Alternative attributions
30.0% (n = 108)
1.2% (n = 2)
19.6% (n = 55)

NONBELIEVED MEMORIES
22
General beliefs
18.0% (n = 67)
0.6% (n = 1)
2.9% (n = 8)
Internal features of
event presentation
16.0% (n = 61)
2.5% (n = 4)
2.1% (n = 6)
Notion of self or others
12.0% (n = 45)
0.0%
1.4% (n = 4)
External evidence
10.0% (n = 40)
27.2% (n = 43)
11.1% (n = 31)
Personal motivation
4.0% (n = 16)
1.8% (n = 3)
0.4% (n = 1)
Change of context
-
12.0% (n = 19)
0.7% (n = 2)
Noticing deterioration
-
2.5% (n = 4)
0.0%
Memory Verification Strategies
We found that 10.6% (n = 53) of the sample mentioned that they attempted to verify their
memory. Of these people, 52.8% (n = 28) indicated trying to verify their memory by asking
family members (e.g., “I remember meeting Elton John at a bar in New York when I was about
8. I remember going to ask him for napkins and an autograph b[ut] I’ve asked my mother and
she says that it never occurred. [...]”).
Furthermore, 9.4% (n = 5) asked others (e.g., friends), 7.5% (n = 4) searched for cues
(e.g., returning to the event location or sought out situations to cue recall, “[my memory is] being
in China and going to visit a huge Buddha temple. The temple was impressive. Few years later I
[visited] China. And I went to [the] place where the temple was [...].” or “I too believed I could
fly as a small child of about 2-3. I tried in vain to re-establish this ability until quite a bit
older…. [...]”), 39.6% (n = 21) looked for physical evidence (e.g., photos), and 5.7% (n = 3)
used cognitive strategies (e.g., trying to remember more of the event or engaged in reasoning
about the event, “About thirty years ago a friend and I were climbing a mountain in Switzerland.
[...] I recall we stayed roped since I recall thinking that I should jump to the opposite side of the

NONBELIEVED MEMORIES
23
ridge if [my companion] slipped - he was leading. [I think we stayed roped] because these days
it might seem prudent to unrope as then only one might fall but there is still debate over this -
current thinking has influenced his memory. I am certain that my recollection is the correct
one”). The percentages do not add up to 100% because some submissions included more than
one verification strategy.
We also examined whether the use of memory verification strategies differed between
believed and nonbelieved memories. We found that in reports categorized as nonbelieved
memories, memory verification strategies were mentioned statistically more frequently (n = 40)
than in those categorized as believed memories (n =12; c2 (1) = 10.18, p = .001, Cramer’s V =
0.15).
Comparison with Wade and Garry (2005) We compared our data with previous
memory verification work (Wade & Garry, 2005; see Table 6). An important observation was
that in our data and Wade and Garry’s (2005) paper, asking family members was the most
frequently mentioned strategy. Furthermore, looking for physical evidence was a seldom used
strategy in Wade and Garry’s (2005) work while we found that it was often mentioned in the
false memory submissions.
Table 6
Memory Verification Strategies
Verification Strategies
Wade and
Garry (2005)
False Memory
Archive
Asking family member
56.7% (n = 20)
52.8% (n = 28)
Asking others
15.0% (n = 5)
9.4% (n = 5)

NONBELIEVED MEMORIES
24
Looking for cues
3.0% (n = 1)
7.5% (n = 4)
Looking for physical evidence
3.0% (n = 1)
39.6% (n = 21)
Using cognitive strategies
28.0% (n = 10)
5.7% (n = 3)
Discussion
We analyzed reports submitted to the False Memory Archive to examine why people
reduced belief in memories and the strategies that they used to verify them. We found that many
archival reports could be classified as NBMs. Furthermore, social feedback was provided as the
main reason for why people viewed their memory as being false and relying on family members
to validate the memory was the most common verification strategy.
The instruction to submit a false memory led many people to submit a NBM. Of course,
NBMs are believed to be ‘false memories’ by the person possessing the memory, although these
memories might actually refer to either false or truly experienced events. The fact that they are
aware that the memory is false might have resulted in people reducing or surrendering belief that
the remembered event took place despite retaining a sense of recollection for the event.
Nonetheless, a large percentage of the remaining reports were classified as believed memories
(41.6%) which might be regarded as surprising in a source of reports that people provide about
false memories. Thus, although people who submitted these reports assumed that their memory
was false, their reports indicated there was still a strong belief that the event took place. Perhaps
these memories were actually NBMs, but because submitters were not asked specifically to
provide reasons for why their memory was false, they did not mention them. The consequence is
that these reports “look” like believed memories due to the language that people chose to use
when describing them (Otgaar et al., 2013). Although these findings might be considered
surprising, they fit well with previous research on different types of NBMs (Scoboria et al.,

NONBELIEVED MEMORIES
25
2017). In that research, people sometimes have NBMs they still largely believe, but are only
modestly unsure about (i.e., “grain of doubt” NBMs).
Alternatively, the high percentage of believed memories in this archive may be linked to
cognitive dissonance theory. Here, when there is disagreement between different sources of
information, people attempt to resolve the discrepancy (Festinger, 1957). This can transpire in
several ways (Scoboria et al., 2014). First, people might distrust their memory leading to the
formation of a NBM. Second, they might reject the contradictory information thereby defending
their memory (see also Scoboria, Otgaar, & Mazzoni, in press; Sheen, Kemp, & Rubin, 2001).
Third, they might decide to appraise the remembered event as less important, independently of
whether they revise the belief that the event occurred.
Regarding the reasons that people noted to reduce belief, social feedback was the most
often reported reason. This is in line with previous research by Scoboria et al. (2015) and Ost
(2017). Interestingly, when comparing the percentages of our study with those reported in
Scoboria and colleagues’ and Ost’s study (see Table 5), our results are most in line with Scoboria
et al.’s study (2017). That is, in Ost’s study, external evidence was mentioned relatively more
often than in our study and in Scoboria et al.’s study. Of course, the explanation for this finding
is that Ost’s sample concerned highly serious and negative events (i.e., sexual abuse). Ost
explained that in his sample, retractors encountered external evidence in the form of, for
example, newspaper articles that were critical about the experiences that the retractors reported.
Similarly, event plausibility was rarely mentioned in Ost’s study while this was frequently
reported in our study and in Scoboria et al’s study. Here too, this might be related to the fact that
in our and in Scoboria et al.’s samples, certain stories were highly bizarre and hence, implausible
(e.g., memories of flying). In Ost’s sample, statements were predominantly about traumatic
events (sexual abuse) that are much more plausible than some of the stories of our and Scoboria

NONBELIEVED MEMORIES
26
et al.’s work. Likewise, in our study, we found substantial evidence that submitters searched for
physical evidence to verify their memories while this was minimally mentioned in Wade and
Garry (2005)’s study. One plausible explanation for this result might be the context of the
different studies. That is, the submissions in our study were part of an art project in which
submitters were exposed to physical sources such as photos and videos related to false memory
creation which might have triggered submitters to mention physical evidence in their statements.
Regarding people’s attempts to verify their memories, people preferred to use more
cheap-and-easy strategies (i.e., asking their family members) than strategies that required more
investment of time and/or energy (e.g., asking people other than family, searching for physical
evidence). Moreover, NBMs were more likely to contain memory verification strategies than
were believed memories. Perhaps people tried to verify the authenticity of the memory causing
them to reduce belief in the occurrence of the memory. For example, people might have asked
one of their family members to validate a memory and when this member suggested that the
memory was false, people might have altered belief for the remembered event.
Findings concerning the reasons for relinquishing belief and verification strategies
broadly fit within the source monitoring framework (Nash, Wade, Garry, & Adelman, 2017).
Source monitoring, in part, refers to the processes that people use to differentiate between events
that truly happened from events that were, for example, imagined (Johnson, Hastroudi, &
Lindsay, 1993). Work in this area has focused on mental heuristics such as the qualities of
memories (e.g., strength of visual details) that are frequently diagnostic of experienced events
(e.g., D’Argembeau, van der Linden, d’Acremont, & Mayers, 2006; Destun & Kuiper, 1999;
Johnson, Foley, Suengas, & Raye, 1988). The work reported here relates to a systematic form of
source monitoring in which people might actively search for evidence (e.g., memory

NONBELIEVED MEMORIES
27
verification) for the authenticity of their mental representations. Because of this search, people
may come to reappraise their mental representation as false which might lead to NBMs.
Blank (2017) suggests that there are many different examples of dissociations of belief
and recollection. For example, déjà vu can be considered a phenomenon where people have a
sense of recognition accompanied by a feeling that this sense is inaccurate (Brown, 2003).
Another example is a state of memory distrust in which people lack confidence in their own
experienced events which has sometimes been linked with the occurrence of false confessions
(Gudjonsson, 2014; Van Bergen, Jelicic, & Merckelbach, 2009). According to Blank, these
recollection-belief dissociations show that before people stop believing in their recollections,
they first go through a “reality check” that may include seeking evidence that a memory reflects
an accurate representation of what happened. When such reality checks fail, dissociations
between recollection and belief might occur. Perhaps the NBMs in archival accounts analyzed
here were the result of such failed reality checks.
Although the archive does not contain any demographic data concerning the submitters
(e.g., age, gender), we do know that the archive concerns submissions from the general
population. On the one hand, the archive may therefore be more diverse than previous studies
focusing on college students or MTurk participants who are provided compensation (Mazzoni et
al., 2010; Scoboria et al., 2015) and hence may be more generalizable to real-words settings. On
the other hand, demographic data on people visiting museums show that they are largely white
and well-educated but do seem to be quite equally divided among different age groups.
A limitation of the current experiment is that although substantial agreement was found
between different raters, this agreement might have been higher when using a shorter scoring
form. That is, in the current study, raters used an extensive scoring form which might have made
the scoring of submissions challenging. Of course, even with this detailed scoring form, our

NONBELIEVED MEMORIES
28
results are in line with previous results in this area (e.g., Scoboria et al., 2015). Nevertheless,
future research might attempt to use shorter and perhaps more simple scoring forms. Second, in
the current study, people who submitted a false memory account received little guidance on what
they should exactly report. The likely consequence of this is that the submissions varied much in
content. In previous work on the reporting of nonbelieved memories, participants received more
guidance on the events that occurred in their childhood. For example, participants were
specifically asked to think about events from their childhood and had to insert a short description
of each event which could serve as memory cue of the event (Scoboria & Talarico, 2013).
In sum, the current study focused on reports from the False Memory Archive. Many of
these reports were judged to be NBMs. Submitters described that social feedback led them to
believe that their memory was false and that they often approached family members to verify
their memory. This study emphasizes the flexibility that exists in autobiographical belief and
supports the view that autobiographical belief and recollection reflect distinct underlying
processes (Scoboria et al., 2014).
References

NONBELIEVED MEMORIES
29
Clark, A., Nash, R. A., Fincham, G., & Mazzoni, G. (2012). Disowned recollections: Denying
true experiences undermines belief in autobiographical events. PLoS One, 7, 1-7 e32998.
Blank, H. (2017). Recollection, belief and metacognition: A reality check. Memory, 25, 869-875.
Brédart, S., & Bouffier, M. (2016). Nonbelieved memories in middle-aged and older people.
Consciousness and Cognition, 42, 352-357.
Brown, A. S. (2003). A review of the déjà vu experience. Psychological Bulletin, 129, 394-413.
D’Argembeau, A., Van der Linden, M., d’Acremont, M., Mayers, I. (2006). Phenomenal
characteristics of autobiographical memories for social and non-social events in social
phobia. Memory, 14, 637-47.
Destun, L. M. & Kuiper, N. A. (1999). Phenomenal characteristics associated with real and
imagined events: The effects of valence and absorption. Applied Cognitive Psychology,
13, 175-186.
Farrell, B., & Medvedeva, M. (2010). Demographic transformation and the future of museums.
The AAM Press, American Association of Museums.
Festinger, L. (1957). A theory of cognitive dissonance. Stanford, CA: Stanford University Press.
Foley, M.A. (2017). Reflecting on how we remember the personal past: Missing components in
the study of memory appraisal and theoretical applications. Memory, 26, 634-652.
Gudjonsson, G. (2014). How I got started: From memory distrust to false confessions. Applied
Cognitive Psychology, 28, 809-811.
Johnson, M. K., Foley, M. A., Suengas, A. G., & Raye, C. L. (1988). Phenomenal characteristcis
of memories for perceived and imagined autobiographical events. Journal of
Experimental Psychology: General, 117, 371-376.
Johnson, M. K., Hashtroudi, S., & Lindsay, S. (1993). Source monitoring. Psychological
Bulletin, 114, 3-28.

NONBELIEVED MEMORIES
30
Landis, J.R., & Koch, G.G. (1977). The measurement of observer agreement for categorical data.
Biometrics, 33, 159-174.
Maran, M. (2010). My lie: A true story of false memory. San Francisco, CA: Jossey-Bass.
Mazzoni, G., Clark, A., & Nash, R. A. (2014). Disowned recollections: Denying true experiences
undermines belief in. Acta Psychologica, 145, 139-146.
Mazzoni, G., Scoboria, A., & Harvey, L. (2010). Non-believed memories. Psychological
Science, 21, 1334-1340.
Nash, R. A., Wade, K. A., Garry, M., & Adelman, J. (2017). A robust preference for cheap-and-
easy strategies over reliable strategies when verifying personal memories. Memory, 25,
890-899. doi: 10.1080/09658211.2016.1214280
Ost, J. (2017). Adults' retractions of childhood sexual abuse allegations: high stakes and the
(in)validation of recollection. Memory, 1-10.
Otgaar, H., Scoboria, A., Howe, M.L., Moldoveanu, G., & Smeets, T. (2016). Challenging
memories in children and adults using an imagination inflation procedure. Psychology of
Consciousness: Theory, Research, and Practice, 3, 270-283.
Otgaar, H., Scoboria, A., & Mazzoni, G. (2014). On the existence and implications of
nonbelieved memories. Current Directions in Psychological Science, 23, 349-354.
Otgaar, H., Scoboria, A., & Smeets, T. (2013). Experimentally evoking non-believed memories
for childhood events. Journal of Experimental Psychology: Learning, Memory, &
Cognition, 39, 717-730.
Scoboria, A., Boucher, C., & Mazzoni, G. (2015). Reasons for withdrawing belief in vivid
autbiographical memories. Memory, 23, 545-562.

NONBELIEVED MEMORIES
31
Scoboria, A., Jackson, D. L., Talarico, J., Hanczakowski, M., Wysman, L., & Mazzoni, G.
(2014). The role of belief in occurrence within autobiographical memory. Journal of
Experimental Psychology: General, 143, 1242-1258.
Scoboria, A., Mazzoni, G., & Boucher, C. (2016). Nonbelieved memories: A review of findings
and theoretical implications. In R. A. Nash, & J. Ost, False and distorted memories.
online: psychology press.
Scoboria, A., Nash, R. A., & Mazzoni, G. (2017). Sub-types of nonbelieved memories reveal
differential outcomes of challenges to memories. Memory, 25, 876-889.
Scoboria, A., Otgaar, H., & Mazzoni, G. (2018). Defending and reducing belief in memories: An
experimental laboratory analogue. Memory & Cognition, 1-17. DOI: 10.3758/s13421-
018-0800-1
Scoboria, A., & Talarico, J. M. (2013). Indirect cueing elicits distinct types of autobiographical
event representations. Consciousness and Cognition, 22, 1495-1509.
Scoboria, A., Wade, K. A., Lindsay, D. S., Azad, T., Strange, D., Ost, J., & Hyman, I. (2017). A
mega-analysis of memory reports from eight peer-reviewed false memory implantation
studies. Memory, 1-18.
Statista. (2019). Share of adults who visited a museum or gallery in the last year in England from
2012/13 to 2017/18, by age. Retrieved from
https://www.statista.com/statistics/418323/museum-galery-attendance-uk-england-by-
age/
Sheen, M., Kemp, S., & Rubin, D. (2001). Twins dispute memory ownership: A new false
memory phenomenon. Memory & Cognition, 26, 779-788.

NONBELIEVED MEMORIES
32
Van Bergen, S., Jelicic, M., & Merckelbach, H. L. G. J. (2009). Are subjective memory
problems related to suggestibility, compliance, false memories, and objective memory
performance? The American Journal of Psychology, 122, 249-257.
Wade, K., & Garry, M. (2005). Strategies for verifying false autobiographical memories. The
American Journal of Psychology, 118, 587-602.
Wade, K. A., Nash, R. A., & Garry, M. (2014). People consider reliability and cost when
verifying their autobiographical memories. Acta Psychologica, 146, 28-34