Bookbinder, S. H., & Brainerd, C. J. (2016) .

Psychological Bulletin
Emotion and False Memory: The Context–Content Paradox

S. H. Bookbinder and C. J. Brainerd
Online First Publication, October 17, 2016. http://dx.doi.org/10.1037/bul0000077
CITATION
Bookbinder, S. H., & Brainerd, C. J. (2016, October 17). Emotion and False Memory: The
Context–Content Paradox. Psychological Bulletin. Advance online publication. http://
dx.doi.org/10.1037/bul0000077
Psychological Bulletin © 2016 American Psychological Association
2016, Vol. 142, No. 10, 000 0033-2909/16/$12.00 http://dx.doi.org/10.1037/bul0000077
Emotion and False Memory: The Context–Content Paradox

S. H. Bookbinder and C. J. Brainerd
Cornell University
False memories are influenced by a variety of factors, but emotion is a variable of special significance,
for theoretical and practical reasons. Interestingly, emotion’s effects on false memory depend on whether
it is embedded in the content of to-be-remembered events or in our moods, where mood is an aspect of
the context in which events are encoded. We sketch the theoretical basis for this content-context
dissociation and then review accumulated evidence that content and context effects are indeed different.
Paradoxically, we find that in experiments on spontaneous and implanted false memories, negatively
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
valenced content foments distortion, but negatively valenced moods protect against it. In addition,
This document is copyrighted by the American Psychological Association or one of its allied publishers.
correlational data show that enduring negative natural moods (e.g., depression) foment false memory.
Current opponent-process models of false memory, such as fuzzy-trace theory, are able to explain the
content-context dissociation: Variations in emotional content primarily affect memory for the gist of
events, whereas variations in emotional context primarily affect memory for events’ exact verbatim form.
Important questions remain about how these effects are modulated by variations in memory tests and in
arousal. Promising methods of tackling those questions are outlined, especially designs that separate the
gist and verbatim influences of emotion.
Keywords: emotion, false memory, fuzzy-trace theory, mood, opponent processes
Over the past quarter-century, false memory has been one of the property destroyed) are infused with emotional content; they are
most extensively studied topics in psychology. Practical motiva- negative and arousing in the terminology of the circumplex model
tions, in particular, have abounded as there are some high-stakes of emotion (Russell, 1980, 1991). On the other hand, the moods of
situations in which the consequences of false memories are quite individual witnesses at the time may have been different. Some
serious (e.g., courtroom testimony, eyewitness identifications of may have been happy just prior to the crime, then became fearful
suspects, histories taken during psychotherapy, recountings of and eventually angry. The moods of witnesses may be different
battlefield events, histories taken during emergency room treat- still at the time of police interviews, although the emotional
ment, terrorism interrogations). The memories that are retrieved in content of the events remains relatively unchanged. The point is
those circumstances are affect-laden, and hence, one of the most that the emotional content of events and the emotional context that
enduring questions about false memory is how it is influenced by is supplied by moods are not the same thing. A key motivation for
emotional states that accompany past experience (e.g., Howe, this review is to consider whether, according to extant data, they
2007; Loftus, 1993; Loftus & Bernstein, 2004; Stein, Ornstein, affect false memory in the same manner.
Tversky, & Brainerd, 1997). As Laney and Loftus (2010) discussed in their review of jurors’
Emotion can figure in past experience in two broad ways. It can perceptions of testimony, the law provides a commonsense answer
be part of the content of events, in the sense that some events are to the emotion-false memory question—namely, that emotional
emotional in themselves, or it can be present in our moods as content inoculates memory against distortion, to the point that it is
events are experienced, which can be thought of as part of the
virtually impossible to have false memories of events whose
context of experience. Importantly, our moods may or may not
content is strongly emotional. This view is so prevalent that it
match events’ emotional content. For instance, consider witnesses
figures routinely in expert testimony in certain types of cases, such
to violent crimes who subsequently attempt to remember those
as when it is a core element in the defense of people who are
events during police interviews. On the one hand, many of the
accused of implanting false memories of traumatic experiences in
events (e.g., seeing someone threatened with a weapon, seeing
plaintiffs (for a review, see Brainerd & Reyna, 2005). However,
this view is known to be wrong empirically, it being well estab-
lished that people can remember a range of traumatic and near-
traumatic events that they did not experience, such as being sex-
S. H. Bookbinder and C. J. Brainerd, Institute of Human Neuroscience ually abused in a previous life (e.g., Spanos, 1996), being abducted
and Department of Human Development, Cornell University. by space aliens (e.g., Spanos, Cross, Dickson, & DuBreuil, 1993),
Preparation of this article was supported by National Institutes of Health committing embarrassing acts at public events (e.g., Hyman &
Grant 1RC1AG036915 and Department of Agriculture Grant NIFA
Pentland, 1996), suffering injuries requiring hospitalization (e.g.,
1003856 to the second author.
Correspondence concerning this article should be addressed to S. H. Garry, Manning, Loftus, & Sherman, 1996), and committing major
Bookbinder, Institute of Human Neuroscience and Department of Human crimes (e.g., Shaw & Porter, 2015). Consequently, the view that
Development, Cornell University, Martha Van Rensselaer Hall, Ithaca, NY that false memories of strongly emotional events are commonplace
14853. E-mail: sb978@cornell.edu has also figured in expert testimony—for instance, in defense of
1
2 BOOKBINDER AND BRAINERD
people who have been accused of bizarre or improbable forms of means that they do not appear in the table of effect sizes (see Table
sexual abuse (see Appelbaum, Uyehara, & Elin, 1997; Kassin, 1). The fact that effect sizes could not be computed for 40% of the
Ellsworth, & Smith, 1989; Loftus & Ketcham, 1996). literature militated against conducting a meta-analysis, and hence,
Although the data show that even highly emotional events are the present article is a narrative review.
prone to memory distortion, some basic uncertainties remain about
the emotion-false memory relation that must be resolved before
False Memory: Theory and Measurement
theory and research can proceed to more subtle questions. Three
elementary ones are these. First, does false memory vary in a To make progress on how emotion influences false memory and
uniform way as a function of the emotional concomitants of to generate theoretical hypotheses, it is first necessary to be clear
experience; that is, is there a simple directional relation such that about what false memory is, operationally speaking, and to con-
distortion consistently increases or decreases as emotion varies? sider the processes that are thought to be responsible for it. Those
Second, does the manner in which false memory reacts to emo- two topics are examined in the present section. Then, in the next
tional variation depend upon the quality of that variation—in section, we consider how emotion has been manipulated in false
particular, whether its valence is positive or negative or how memory experiments.

arousing it is? Third, does the manner in which false memory

reacts to emotional variation depend upon where that variation is
What Is False Memory?
localized—in particular, whether it is inherent in the content of
events or whether it is a feature of the context in which they are False memory merely refers to situations in which subjects
experienced? We consider findings on these questions in this recollect events that, in fact, they did not experience. For instance,
review, and as an advance organizer, it will turn out that the if a friend asks what you ate at a baseball game a week ago and
answer to the last two questions is yes while the answer to the first what you drank at lunch yesterday, you may say hot dog and milk,
is no. Paradoxically, we shall see that whether emotion distorts although you consumed neither. This illustrates three features of
memory or inoculates memory against distortion depends upon false memories as they are normally measured. First, misremem-
whether it is localized in the content or the context of experience. bered events are not ones that subjects have never experienced,
This content-context paradox is one of the main themes of the such as being abducted by space aliens or winning the lottery, so
present review. that they are false in the narrow sense of not being part of a
particular context that is specified in the experimental design
Structure of the Review and Method of (baseball game, lunch). Second, misremembered events are usually
familiar: Hot dogs, unlike baklava, are a common food, and milk,
Literature Search
unlike suanmeitang, is a common drink. Third, misremembered
The article begins with a brief overview of method and theory in events fit the gist of the target context (hot dogs are baseball food,
false memory research— of current accounts of factors that influ- milk is a luncheon beverage). Thus, the false memories that are
ence false memory, including manipulations and measures that are measured in the modal experiment are semantic errors that are
used to test those accounts. We then review findings from false rooted in strong meaning resemblance to actual events.
memory experiments in which emotional content and mood were Although these are modal features that hold for most published
manipulated, with attention to methodological differences that may experiments, none is universal. Researchers occasionally study
explain why a single, clear pattern for emotional influences has not false memories of events that subjects have never experienced
yet emerged. We conclude with a working explanation that an- (e.g., being in a traffic accident, being lost in a mall), that are
swers the direction, quality, and location questions and proposes unfamiliar or even bizarre, or that do not share semantic content
near-term targets for research on emotion and false memory. with the experimental context (e.g., Santa Claus in a baseball game
Our search method began with Web of Science, searching for video, a gorilla in a ballet video; for a review, see Brainerd &
entries containing the terms “emotion” and “false memory.” We Reyna, 2005). Nevertheless, the bulk of what we know about how
then followed up using “affect,” “mood,” “valence,” or “arousal,” emotion affects false memory is for familiar, previously experi-
as the first term and “false memory” or “misinformation” as the enced events that preserve the meaning of the events to which
second. We then performed the same searches with the Google subjects are exposed. That still leaves very wide latitude with
Scholar and PSYCinfo databases. Using the resulting articles, we respect to how false memories are induced and what types of
conducted two snowball searches: First, we consulted the reference events subjects are exposed to.
lists of the articles and searched for citations of referenced articles Spontaneous false memory. In our example of eating hot
in Web of Science, and second, we did likewise with the reference dogs and drinking milk, suppose that your memory errors were
lists of recent unpublished and in press articles. The latter articles pursuant to recall and recognition probes such as: What did you eat
were secured by searching conference proceedings and by contact- at the game? Did you eat a hot dog at the game? What did you
ing colleagues. Ultimately, we located 46 peer-reviewed articles drink at lunch? Did you drink a glass of milk at lunch? Apparently,
reporting research that met three inclusion criteria: (a) The depen- such errors must be attributable to spontaneous, endogenous dis-
dent variable was a form of false memory, and emotion was either tortion processes that are a normal part of how episodic memory
(b) manipulated in the form of content variations (e.g., valence/ operates; that is, they are natural concomitants of trying to remem-
arousal of words, pictures) and/or context variations (e.g., valence/ ber familiar events that fit with the gist of events that were actually
arousal of music, videos), or (c) measured (e.g., scores on depres- experienced.
sion scales) and correlated with false memory. Of these 46 articles, Over the decades, these spontaneous false memories have been
19 did not report sufficient data to compute effect sizes, which measured with materials as varied as narratives (Bartlett, 1932),
Table 1
Overview of Studies With Effect Sizes
Study n Paradigm Population Type of test Groups Mean Comparison Cohen’s d
Budson et al. (2006) 58 DRM Younger adults Recognition Negative .37 .08
Neutral .39
Older adults Recognition Negative .49 .11
Neutral .46
Adults with AD Recognition Negative .63 .18
Neutral .68
Howe, Toth, and Cicchetti 284 DRM Maltx 6–9 yos Recall Negative .17 4.34
(2011) Neutral .11
Maltx 10–12 yos Recall Negative .15 10.38
Neutral .10
Non-maltx 6–9 yos Recall Negative .16 4.06
Neutral .11
Non-maltx 10–12 yos Recall Negative .14 10.60
Neutral .10
El Sharkawy et al. (2008) 32 DRM Young adults Recognition Negative 7.88 .67
Neutral 6.56
Howe et al. (2010; Exp. 40 DRM Young adults Recognition Negative .67 1.28
1) Neutral .51
Recall Negative .27 1.44
Neutral .40
Howe et al. (2010; Exp. 60 DRM Children Recognition Negative .67 .45
2) Neutral .59
Recall Negative .23 .73
Neutral .34
Howe et al. (2010; Exp. 60 DRM Young adults Delayed recall Negative .20 .95
3) Neutral .30
Howe et al. (2010; Exp. 30 DRM 5 & 8 yos Delayed recall Negative .10 1.82
4) Neutral .25
Howe et al. (2010; Exp. 80 DRM 7 & 11 yos Delayed recall Negative .20 .77
EMOTION AND FALSE MEMORY
5) Neutral .31
Dehon et al. (2010; Exp 36 DRM Young adults Recognition Negative .42 Neg vs. neu .71
1) Positive .60 Pos vs. neu .77
Neutral .59
Dehon et al. (2010; Exp 54 DRM Young adults Recall Negative .37 Neg vs. neu .70
2) Positive .39 Pos vs. neu .81
Neutral .19
Gallo, Foster, and 24 Pictures with labels Younger and older Recognition Negative .14 Pos vs. neu .82
Johnson, (2009) adults Positive .18 Neg vs. neu .70
Neutral .09 Pos vs. neg .65
Choi et al. (2013; 48 Pictures with Young adults Recognition Negative .25 Neg ⫽ pos vs. neu 1.41
Exp. 2) thematic labels Positive .30
Neutral .25
Bookbinder & Brainerd 68 Categorized pictures Young adults Recognition Immediate Immediate test
(2016) Negative .67 Neg vs. pos 2.35
Positive .61 Pos vs. neu 3.14
Neutral .53 Delay test
Neg vs. pos 4.00
(table continues)
3
Table 1 (continued) 4
Delay Neu vs. pos 3.00

Negative .53
Positive .45
Neutral .51
Mirandola et al. (2014) 53 Scripted material in Young adults (who did Recognition Negative .25 Gap-filling errors .15
pictures not elaborate Gap-filling errors Neutral .28 Causal errors .45
Causal errors Negative .32
Neutral .44
Porter et al. (2010) 40 Misinformation with Young adults Recognition of major Negative .82 .38
pictures misinformation Positive .92
Porter et al. (2014) 44 Misinformation with Young adults Recognition of major Descript. at encoding .42
ambiguous misinformation Descript. at recall .32
pictures Description of Negative .51
picture at None .35
encoding
Description of Negative .33
picture at recall None .45
Van Damme and Smets 53 Misinformation with Young adults Recognition of Positive/low .20 Positive/low vs. negative/low 6.75
(2014) pictures peripheral details Positive/high .21 Positive/high vs. negative/high 7.50
Negative/low .47 Ambiguous/low vs. ambiguous/high 5.52
Negative/high .59
Ambiguous/low .33
Ambiguous/high .58
Storbeck and Clore (2005; 100 Music induction; Young adults Recall Negative 5.40 Pos vs. neg .17
Exp 1) DRM Positive 6.00 Neg vs. control .34
Control 7.00
BOOKBINDER AND BRAINERD
Storbeck and Clore (2005; 119 Music induction; Young adults Recall Negative 8.00 Pos vs. neg .15
Exp 2) DRM Positive 9.60 Neg vs. control .13
Control 9.20
Corson and Verrier (2007) 222 Music induction; Young adults Recall Happy .34 Recall
DRM Angry .36 Happy vs. serene .60
Serene .24 Angry vs. sad .96
Sad .20 Recognition
Control .25 Happy vs. serene .40
Recognition Happy .66 Angry vs. sad .38
Angry .64
Serene .57
Sad .55
Control .54
Storbeck (2013; Exp. 2) 86 Music induction; Young adults Recognition Negative .60 Neg vs. pos .79
DRM Positive .73 Neg vs. neu .44
Control .68
Storbeck (2013; Exp. 3) 70 Music induction; Young adults Word Recognition Negative .44 Word
DRM Positive .60 Neg vs. pos .74
Control .59 Neg vs. neu .65
(table continues)
Table 1 (continued)
Word ⫹ picture Negative .43 Word ⫹ picture

recognition Positive .48 Neg vs. pos .21
Control .46 Neg vs. neu .13
Van Damme (2013; Exp. 301 Music induction, Young adults Recognition at end Serene .81 Serene vs. happy 2.33
1) DRM of all lists Happy .74 Sad vs. angry 3.00
Sad .82 Neutral vs. control 1.33
Angry .76
Control1 .81
Control2 .77
Van Damme (2013; Exp. 297 Music induction, Young adults Recognition after Serene .71 Serene vs. happy .57
3) DRM each list Happy .68 Sad vs. angry 2.25
Sad .71
Angry .62
Control1 .70
Control2 .71
Knott, Threadgold, and 72 Video induction, Young adults Recall Negative .19 Neg vs. pos .80
Howe (2014) DRM Positive .33 Neg vs. control 1.00
Control .40
Yang et al. (2015) 78 Real-life mood Young adults Recognition Positive 42.5 .72
induction, DRM Neutral 56.5
Ruci, Tomes, and 93 Narrative induction, Young adults Recall Positive mood- .49 Recall
Zelenski (2009) emotional DRM positive CL Positive mood-positive CL vs. 1.26
Positive mood- .18 positive mood-negative CL
negative CL Positive mood-positive CL vs. .80
Positive mood- .27 positive mood-neutral CL
neutral CL Negative mood-negative CL vs. 1.47
Negative mood- .55 negative mood-positive CL
negative CL Negative mood-negative CL vs. 1.04
Negative mood- .13 negative mood-neutral CL
EMOTION AND FALSE MEMORY
positive CL Recognition
Negative mood- .22 Positive mood-positive CL vs. .89
neutral CL positive mood-negative CL
Recognition Positive mood- .86 Positive mood-positive CL vs. 1.17
positive CL positive mood-neutral CL
Positive mood- .60 Negative mood-negative CL vs. .96
negative CL negative mood-positive CL
Positive mood- .51 Negative mood-negative CL vs. 1.12
neutral CL negative mood-neutral CL
Negative mood- .86
negative CL
Negative mood- .58
positive CL
Negative mood- .50
neutral CL
Knott and Thorley (2014) 48 Video induction; Young adults Delayed recognition Negative mood- .89 Negative mood-negative CL vs. 5.63
emotional DRM negative CL negative mood-neutral CL
Negative mood- .55
neutral CL
(table continues)
5
Table 1 (continued)
Bland et al. (2016) 83 Video induction; Young adults (18 year Recognition Fearful mood- .93 Fear-fear vs. fear mood-anger list 5.69
emotional DRM olds) fearful CL
Fearful mood- .57 Fear-fear vs. fear mood-neutral list 5.74
angry CL
Fearful mood- .53 Angry-angry vs. angry mood-fear 2.14
neutral CL list
Angry mood- .78 Angry-angry vs. angry mood- 3.99
angry CL neutral list
Angry mood- .63
fearful CL
Angry mood- .48
neutral CL
Van Damme and 300 Video induction; Young adults Recognition Serene .11 Serene vs. sad 4.96
Seynaeve (2013) misinformation Happy .11 Happy vs. angry .97
Sad .15
Angry .09
Control1 .13
Control2 .12
Howe and Malone (2011) 48 Emotional DRM Depressed subjects and Recognition Depression- Depression-related lists: depressed 1.10
controls related lists vs. controls
Depressed .75
Controls .47
BOOKBINDER AND BRAINERD
Toffalini et al. (2014) 60 Scripted material in Depressed subjects and Recognition Depressed .52 Negative causal errors: depressed 1.65
pictures controls Controls .15 vs. controls
Bremner, Shobe, and 63 DRM PTSD diagnosis, Recognition PTSD .95 PTSD vs. abuse .51
Kihlstrom (2000) history of abuse but Abuse .78 PTSD vs. female controls .50
no PTSD diagnosis, Female control .79 PTSD vs. male controls .31
controls Male control .86
Brennen, Dybdahl, and 100 DRM PTSD and control Recall PTSD .34 War-related: PTSD vs. control .65
Kapidzic (2007) Control .19
Hauschildt et al. (2012) 62 Videos PTSD, trauma-non- Sensitivity PTSD 53.94 .70
PTSD, control Control 60.61
Goodman et al. (2011) 93 DRM Adults and adolescents Recall CSA-related .15 CSA CLs vs. positive CLs .46
with and without Negative .16 Negative CLs vs. positive CLs .50
histories of CSA Positive .06
and PTSD Neutral .11
EMOTION AND FALSE MEMORY 7
sentences (Bransford & Franks, 1971; Kintsch, Welsch, Schmal- interview about that material, with suggestions being delivered as
hofer, & Zimny, 1990), lists of semantically related words or leading questions. For instance, suppose that the target material is
pictures (Brainerd & Reyna, 2007; Howe, 2005), live staged events a video of a convenience store robbery. During the misinformation
(e.g., Holliday & Hayes, 2000; Lampinen, Copeland, & Neuschatz, phase, the leading questions might include: How long was the
2001), and videos of real-life events (e.g., Bjorklund, Bjorklund, knife he was holding? (The robber’s hands were empty.) What
Brown, & Cassel, 1998; Bjorklund et al., 2000). When recall tests color was his baseball cap? (The robber was wearing a ski cap.)
are administered, the raw intrusion rate for unpresented events is What type of candy did he grab from the counter? (The robber
the false memory index, and when recognition tests are adminis- grabbed cigarettes.) On subsequent memory tests, recall/recogni-
tered, the false memory index is the false alarm rate for such tion of such suggested details supply the false memory measure,
events, which normally incorporates a correction for response bias, and recall/recognition of actual details from the video supply the
such as the signal detection statistics a= and d= (see Snodgrass & true memory measure.
Corwin, 1988). Subjects respond to three types of probes when Beyond interpolated suggestions, a key feature of the typical
recognition tests are administered: targets (presented items; true misinformation experiment is that the target materials are (a)
memory measures), related distractors (unpresented meaning- realistic depictions of real-life events that (b) revolve around a
sharing items; false memory measures), and unrelated distractors theme that is either obviously forensic or is at least forensically
(unpresented items that do not share meaning with presented related. By virtue of the latter property, negative emotion is a
material; bias controls). Bias correction is essential because re- routine feature of misinformation experiments. In Loftus’ (1975)
sponse bias can differ dramatically for different subject popula- original work, for instance, the target materials consisted of a slide
tions and conditions (e.g., Brainerd, Reyna, & Forrest, 2002). show of a traffic accident that resulted in a pedestrian injury, and
Currently, the dominant procedure for studying spontaneous misinformation focused on details (traffic signs) that determined
false memories is a word list paradigm, the Deese/Roediger/Mc- driver responsibility. On memory tests, subjects could no longer
Dermott (DRM; Deese, 1959; Roediger & McDermott, 1995) distinguish between critical details in the slide show and details
illusion. Subjects study short lists of related words (e.g., town, that were only suggested to them.
state, capital, streets, subway, village, . . .) from which a critical The misinformation literature is vast, and several reviews of
word (city) is missing, followed by recognition or recall tests. portions of that literature have been published (e.g., Brainerd &
Recall or recognition of this missing word (usually called a critical Reyna, 2005; Ceci & Bruck, 1993; Goodman, 2006; Reyna &
distractor or critical lure) is the false memory measure. The key Titcomb, 1997; see various chapters in Bjorklund, 2000). Although
feature of the DRM illusion for this review is that it is easily there have been occasional controversies as to how best to measure
adapted to study the effects of emotion. With respect to emotional the distortive effects of memory suggestion (e.g., McCloskey &
content, Budson et al. (2006; see also Pesta, Murphy, & Sanders, Zaragoza, 1985; see various chapters in Doris, 1991), those effects
2001) pointed out that the DRM illusion can be compared for are routinely observed experimentally, and when coupled with
negatively valenced lists (e.g., mad, fear, hate, rage, temper, . . .; reports of false memories in legal cases in which witnesses were
critical distractor ⫽ anger), positively valenced lists (e.g., child, subjected to suggestive interviewing (for reviews, see Kassin et al.,
cute, infant, mother, doll, . . .; critical distractor ⫽ baby), and 2010; Wells et al., 1998), there is little doubt that suggestion
neutral lists (e.g., blouse, sleeves, pants, tie, button, . . .; critical reshapes memory for the events of our lives.
distractor ⫽ shirt). With respect to emotional context, Storbeck In the present review, the crucial point about misinformation
and Clore (2005) pointed out that the DRM illusion can be com- research is that owing to its forensic slant, it provides another
pared for subjects who were in different mood states when lists existence proof that events that are accompanied by negative
were studied or tested. emotion, at least, can be systematically misremembered. Modifi-
Implanted false memory. Returning to our example of eating cations to the basic misinformation procedure are then required to
hot dogs and drinking milk, suppose that your false memories were answer more subtle questions about whether suggestion also dis-
pursuant to probes such as: You ate a hot dog at the game, didn’t torts memory when positive emotion is present, about the relative
you? You drank a glass of milk at lunch, didn’t you? Now, you are susceptibility to distortion of events that are accompanied by
confronted with clear suggestions about what you ate and drank, positive versus negative versus neutral emotion, and about whether
which are hallmarks of lawyerly questioning and police interviews. emotion effects are different when emotion is embedded in content
Presumably, you will be more likely to misremember than you versus context (see below).
were before. If so, it can no longer be assumed that endogenous
distortion processes are responsible because external distortion is
Theories of False Memory
present.
The basic procedure for studying implanted false memories, the If emotional content or context influence false memory, the
misinformation paradigm, was pioneered by Loftus (1975; Loftus basic memory processes that control such errors must be affected.
& Palmer, 1974) and was designed to emulate the suggestive There have been two historical stages in the evolution of theoret-
questioning that is integral to police interviews and interrogations ical hypotheses about the nature of those processes—namely, early
(see Inbau, Reid, Buckley, & Jayne, 2001). This procedure in- one-process accounts, followed by contemporary opponent-
volves the same two steps as spontaneous false memory designs, process theories (Brainerd & Reyna, 2002). One-process theories,
plus another step that is interpolated between them. During the which emerged in the 1970s and 1980s, and their associated
interpolated step, the misinformation phase, erroneous suggestions literatures, were reviewed by Reyna and Brainerd (1995). The
are presented about the target material that subjects have encoded. most influential examples are constructivism (Bransford & Franks,
To emulate police investigations, it traditionally consists of an 1971), schema theory (Alba & Hasher, 1983), and the source-
monitoring framework (Johnson, Hashtroudi, & Lindsay, 1993).
Increases when true memory
Increases for positive moods

Supported by familiarity but
False memory predictions
Reduced for central details
via relational processing

Increased for peripheral or
The hallmark of these theories is the assumption that true and false
compared to positive
memories are products of a single process, although the specific
Increased for negative

of emotional event
process was different for different theories: Constructivism as-
not recollection
neutral details
sumed that true and false memory are products of meaning-driven
interpretation and elaboration of target events; schema theory
increases
valence
assumed that target events are encoded into preexisting interpre-
tive structures (schemata), with episodic memory relying on re-
trieval from those structures; and the source-monitoring frame-
work assumed that episodic memory relies on source attributions,
which are attributions about the contexts in which events were
experienced.
Increased for central emotional details;
Increases for negative moods via item

Negative valence and arousal increase
In connection with these theories, the common thread that true
reduced for peripheral or neutral
Increased for positive compared to

and false memory are attributable to a common process lead to the
Increases when mood is same at

True memory predictions

straightforward predictions that for a given type of false memory
Supported by recollection and

Increases when false memory
task, (a) variability in true and false memory over subjects and
encoding and retrieval

conditions should be positively correlated, and (b) manipulations
that increase true memory should also increase false memory and
negative valence
conversely (see Brainerd, Reyna, & Ceci, 2008). However, data
TM for details
from early studies persistently failed to support either prediction.
processing
familiarity
increases
Several investigators examined the correlation between target hit
details
rates and related distractor false alarm rates, finding that the two
are usually uncorrelated (e.g., Brainerd, Reyna, & Kneer, 1995;
Reyna & Kiernan, 1994, 1995; Tussing & Greene, 1999). Also,
various manipulations were identified that dissociate true from
false memory, in the sense that they affect one but not the other
(single dissociation) or they drive them in opposite directions
Emotional valence strengthens gist and

negative valence strengthens it more
peripheral or non-emotional details
congruence between study and test

(double dissociation; e.g., Israel & Schacter, 1997; Lampinen et
Emotion enhances memory accuracy
Mood affects item versus relational

together for true memory but in
A single process controls true and
Recollection and familiarity work

al., 2001; Payne, Elie, Blackwell, & Neuschatz, 1996; Schacter,
opposition for false memory
Memory is enhanced by mood

Israel, & Racine, 1999).
Emotion reduces attention to

Opponent-process accounts grew out of such findings, and they
Central idea
than positive valence

have two major characteristics (see Table 2 for a summary). First,
they echo the dominant contemporary perspective on episodic
memory, the dual-process approach, which is exemplified by many
false memory
Central Ideas and Predictions of Theories of Emotion and False Memory
theories in the mainstream memory literature (e.g., Jacoby, Begg,
processing
& Toth, 1997; Parks & Yonelinas, 2007; Rotello, Macmillan, &
Reeder, 2004). Their common feature is the principle that remem-
bering an item, such as the word bagpipe from a study list, can be
attributable to either of two retrieval operations. One, which is
most often called recollection, brings realistic details of the item’s
presentation to conscious awareness, which may include psycho-
logical details (a visual image of a bagpipe or kilt) as well as
Emotional content
Emotional content
Emotional content
Emotional mood
Emotional mood
physical ones (the visual appearance of bagpipe on a computer

False memory
False memory
Domain
screen). The other, which is most often called familiarity, produces

strong feelings that the item was presented but without awareness
of accompanying details.
Dual-process approaches to false memory posit that the two
processes operate in opposition when it comes to distortion. Fa-
miliarity supports false memories because the salient properties
that related distractors (e.g., chair, robin) share with targets (e.g.,
Easterbrook; central/peripheral
couch, sparrow) produce strong feelings that they were presented,

Emotional enhancement of
but recollection suppresses false memories because it consciously

Network theory of affect
reinstates verbatim content, eliminating a variety of similar-but-

Affect as information
Theory
not-verbatim possibilities. Fuzzy-trace theory (FTT; Reyna &

Fuzzy-trace theory
Opponent-process
Brainerd, 1995) is the original example of such a theory and the

One-process
one that has most often figured in emotion-false memory research.

memory
tradeoff
In FTT, subjects are assumed to store two distinct traces of target

Table 2
items in parallel: (a) verbatim traces of targets’ surface form (e.g.,

the word “couch” and the word “sparrow”), along with (b) gist
traces of their meanings (e.g., “household furniture,” “bird”). Re- been tied to variability in valence versus arousal (Colibazzi et al.,
trieval of verbatim traces is the basis for what is usually called 2010). Arousal has been most consistently linked to amygdala
recollection, supporting true memories (“yes, I clearly remember activation (e.g., Anderson, Christoff, Panitz, De Rosa, & Gabrieli,
studying couch”) and suppressing false ones (“no, I didn’t study 2003), especially when information is negatively valenced (An-
chair because I clearly remember studying couch instead”). Re- derson & Phelps, 2001). Valence, on the other hand, has most
trieval of gist traces is the basis for what is commonly called often been linked to activation in the prefrontal and cingulate
familiarity, supporting true memories (“yes, I studied couch be- cortices (Beauregard, Levesque, & Bourgouin, 2001; Colibazzi et
cause I know there was furniture on this list”) and false memories al., 2010).
for items that share that gist (chair). Of course, there are influential theoretical conceptions other
Summing up, if false memory is affected by emotional content than the circumplex model, such as the theory of discrete emotions
or context, opponent-process theories say that this would happen in (e.g., Ekman, 1992) and psychoevolutionary theory (e.g., Plutchik,
either or both of two ways: (a) by influencing recollection-based 2001). According to the former, there is a small number of discrete
suppression of false memories or (b) influencing familiarity-based emotions, which have evolved for distinct universal purposes, each
acceptance of false memories. We shall see that the literature of which is grounded in a specific neurological system with a
contains examples of both of these hypotheses. Here, experimen- specific physiological expression. Facial expression research fo-
tation on how the DRM illusion is affected by negative emotion is cuses on the universality and physiological distinctiveness of dis-
illustrative. Storbeck and Clore (2005) hypothesized that during crete emotions (e.g., Ekman & Friesen, 1971; Ekman, Levenson,
the study phase, the presence of negative emotion in subjects’ & Friesen, 1983). Psychoevolutionary theory (Plutchik, 2001)
moods enhances the processing of targets’ surface forms, produc- combines the discrete and continuous views of emotion and has
ing stronger verbatim traces and enhanced suppression of false stimulated multiple models. According to one, there are eight basic
memories. In contrast, Brainerd, Stein, Silveira, Rohenkohl, and emotions, with each being treated as a dimension that combines
Reyna (2008) hypothesized that the presence of negative emotion with the others to produce emotional states that are mixtures of
in targets themselves enhances the processing of semantic relations these dimensions. Studies of computer recognition of facial ex-
among them, producing stronger gist traces and higher levels of pressions of emotion have been used to investigate this particular
false memory. model (e.g., Susskind, Littlewort, Bartlett, Movellan, & Anderson,
2007).
Despite the existence of these other models, the circumplex
Emotion-False Memory: Method and Theory
approach has dominated research on emotion-false memory. No
We now sketch methodologies that have been used to measure doubt, that is largely attributable to the availability of multiple
and manipulate levels of emotion, and we consider the theoretical pools of normed materials that allow the dimensions of valence
hypotheses that have most often guided research on how emotion and arousal to be manipulated under controlled conditions, a
might influence false memory. Then, in the next two sections, we matter to which we now turn.
review findings from extant studies of how false memory is Manipulating emotional content. Emotional word lists are
influenced by the emotional content of target events and by sub- the most common method of manipulating the affective content of
jects’ moods. target events. We noted earlier that Budson et al. (2006; cf. Table
3) originated this technique by modifying DRM lists so that some
consisted of emotional words and emotional critical distractors
Manipulating and Measuring Emotion
(e.g., cough, fever, ill, flu, and vomit; with the critical distractor
Most emotion-false memory research has relied on procedures sick) and others did not (e.g., shoe, hand, toe, kick, and sock; with
that implement the circumplex model of emotion (Russell, 1991), the critical distractor foot). Budson et al.’s emotional lists are
according to which emotional states vary continuously along in- actually mixtures of emotional words (e.g., violate, isolated, sad)
dependent dimensions of valence and physical arousal. Russell and neutral ones (e.g., man, quiet, tissue). Other than the emotional
defined valence as varying from pleasant to unpleasant and arousal content manipulation, the Budson et al. procedure is a standard
as varying from calm to excited. Other conceptualizations have DRM design in which subjects study the lists in Table 3 followed
defined valence as varying from approach to withdrawal (Lang, by recognition or recall tests to measure true and false memory
Bradley, & Cuthbert, 1998) or from positivity to negativity (Tel- (e.g., Howe, 2007; Howe, Candel, Otgaar, Malone, & Wimmer,
legen, Watson, & Clark, 1999). The literature contains a number of 2010).
standardized tools—such as the Affective Norms for English There are three complexities with these materials that have been
Words (ANEW; Bradley & Lang, 1999), the International Affec- topics of subsequent methodological refinement (Brainerd, Holli-
tive Pictures System (IAPS; Lang, Bradley, & Cuthbert, 2008), day, Reyna, Yang, & Toglia, 2010; Brainerd, Stein, et al., 2008).
and the Geneva Affective Picture Database (GAPED; Dan-Glauser First, whereas the emotional lists are emotional, it is inaccurate to
& Scherer, 2011)—that supply investigators with pools of familiar say that the others are nonemotional. Some of the nonemotional
words or photographs that have been normed for valence and critical distractors (sweet, fruit, soft, and sleep) are clearly emo-
arousal (also for a third dimension, dominance.) tional words, as are some of the words on the corresponding lists
In the circumplex approach, any specific emotion can be con- (candy, treat, apple, orange, nap, dream). The key difference
ceived of as a pair of values on continuous scales of valence and between emotional versus nonemotional lists is valence rather than
arousal. The separability of those dimensions is assumed to be emotion per se: Words such as rape and alone are negative,
attributable to the activation of distinct neurological systems (Feld- whereas words such as sweet and soft are positive. Second, there is
man Barrett & Russell, 1998). Here, different brain regions have an arousal difference between the two groups of lists. Many of the
Table 3
Emotional Word Lists That Were Used in Various Studies of Emotional Content Effects on False Memory
Budson et al. (2006)

Neutral Chair Foot Fruit Girl Sleep
Table Shoe Apple Boy Bed
Sit Hand Vegetable Dolls Nap
Desk Toe Orange Female Dream
Seat Kick Kiwi Young Snore
Couch Sock Citrus Dress Awake
Sofa Arch Ripe Pretty Doze
Wood Heel Pear Caring Pillow
Cushion Ankle Banana Pink Yawn
Slow Soft Sweet Teacher Window
Fast Fur Candy School Door
Time Touch Sugar Classroom Glass
Boring Feather Bitter Student Pane

Stay Satin Sour Learn Shade

Wait Fleece Treat Quiz Ledge
Steady Cotton Pastry Test Sill
Turtle Smooth Taste Grade House
Stop Gentle Salty Lecture Open
Negative Alone Anger Cry Danger Hell
Single Mad Tears Fear Devil
Isolated Rage Sad Caution Satan
Solitary Annoyed Tissue Trouble Evil
Apart Furious Sorrow Warning Damned
Separate Bothered Eyes Risk Sin
Quiet Wrath Weep Hazard Demon
Detached Hate Sob Alarm Heaven
Self Mood Unhappy Help Judgment
Hungry Lie Rape Sick Thief
Food Fib Sex Cough Steal
Starve Cheat Man Fever Robber
Famine Truth Violate Ill Crook
Empty False Blame Flu Burglar
Stomach Mislead Struggle Vomit Money
Poor Trick Date Doctor Cop
Eat Fake Force Health Purse
Pangs Betray Shame Dizzy Mugger
Palmer & Dodson (2009)

Neutral Chair Needle Sleep
Table Thread Doze
Rocking Haystack Bed
Recliner Injection Snooze
Stool Sewing Tired
Desk Knitting Snore
Sit Prick Pillow
Sofa Sharp Dream
Sitting Thorn Relax
Bench Point Quiet
Legs Eye Blanket
Negative Kill Pain Sad
Assassinate Suffer Depressed
Slay Hurt Melancholy
Slaughter Ouch Cheerless
Murder Anguish Somber
Execute Pleasure Miserable
Massacre Harm Lonely
Stab Distress Upset
Behead Back Gloomy
Homicide Death Hopeless
Shoot Blood Desolate
Positive Beautiful Happy Love
Gorgeous Glad Adore
Stunning Elated Affection
Picturesque Content Passion
Breathtaking Joyful Heart
(table continues)
Table 3 (continued)
Pretty Pleased Kiss

Lovely Ecstatic Like
Exquisite Laugh Attraction
Striking Enjoyment Care
Attractive Satisfied Devoted
Elegant Enjoyable Admire
Pesta et al. (2001)

Neutral Peach Rink Shave Digit Hook Park
Beach Link Slave Widget Book Bark
Leach Mink Stave Midget Look Dark
Teach Sink Shove Bridget Cook Hark
Reach Wink Share Fidget Nook Lark
Poach Pink Have Dibot Rook Mark
Peak Rank Shade Divvy Took Nark

Perch Risk Shake Dimwit Hock Pack

Peace Ring Shale Digest Honk Perk
Preach Rick Shame Gadget Hood Pork
Peal Fink Shape Dig Hoof Spark
Negative Bitch Hell Penis Rape Slut Whore
Ditch Bell Venus Cape Slug Chore
Hitch Dell Genus Nape Slum Bore
Batch Fell Penal Tape Slur Wore
Pitch Jell Peevish Ripe Slot More
Itch Sell Penance Rope Slue Tore
Botch Tell Venice Race Shut Pore
Mitch Hall Zenith Rapt Slit Sore
Butch Hill Pennies Rake Smut Horn
Birch Hull Punish Rare Glut Shore
Witch Shell Pianist Raze Scut Core
Moritz et al. (2005)

Neutral Window
Light
View
Air
Glass pane
Clean
Open
Curtain
Cover
Door
Flowers
Bird
Transparent
Negative Betrayal (Delusion-relevant) Loneliness(Depression-relevant)
Disappointment Sad
Deceit Alone
Backstabbing Silence
Trust Sorrow
Mistrust Isolation
Loyalty Emptiness
Intrigue Longing
Enemy Boring
Secret Depression
State Dull
Dishonest Cold
Spy Hermit
Positive Holidays
Sun
Beach
Sea
Relaxation
Recovery
Leisure
Fly
(table continues)
Table 3 (continued)
Vacation
Palms
Travel
Read
Adventure
critical distractors in the emotional group (e.g., rape, anger, dan- Mood-induction. Here, subjects participate in conditions
ger, hell, cry, lie, thief) refer to states of excitement, as do several that place them in different emotional states either before or
of the list words (e.g., violate, rage, fear, damned, sob, cheat, after presentation of target material. The former can affect
steal). In contrast, none of the critical distractors in the nonemo- encoding processes—and, potentially, retrieval and consolida-
tional group seem arousing, nor do most of the list words. There- tion processes—whereas the latter can affect retrieval and con-
fore, third, the actual differences between the two groups of lists solidation processes. The subjects in all mood conditions are
are that (a) one consists of negatively valenced materials while the exposed to the same target materials, and occasionally, the
other consists of a mixture of positively valenced and neutral materials’ emotional content is also manipulated. During the
materials, and (b) one is more arousing than the other. Conse- test phase, subjects respond to the usual recognition or recall
quently, if the two groups of lists produce different false memory tests. In addition, a mood manipulation check is administered at
rates, this difference cannot simply be attributed to emotional some point, to verify that induction was successful.
versus nonemotional content, and it is unclear whether it is attrib- This brings us to how moods are induced and checked. With
utable to the valence difference or the arousal difference. respect to induction, one of the most widely used techniques (e.g.,
These complexities can be removed by revised lists for which Storbeck & Clore, 2005, 2011) is a music and guided imagery
valence and arousal values are known, which can be done by procedure that was introduced by Mayer, Allen, and Beauregard
relying on word pools such as the ANEW (Bradley & Lang, 1999). (1995). Subjects are asked to adopt a specific mood, and then they
Using such word pools, DRM lists can be constructed for which listen to music that matches that mood, and finally, they read some
valence and arousal levels of both critical distractors and targets mood-matching vignettes. The Mayer et al. materials have been
are precisely counterbalanced. A pool of 32 counterbalanced lists normed for their ability to induce happy, angry, fearful, and sad
that have been administered in some recent experiments appears in moods. A second method consists of viewing emotional pictures,
Table 4. Lists of this sort allow two types of experiments to be
especially from the IAPS, which has the key feature that pictures’
conducted. First, whether false memory is affected by valence per
valence and arousal levels can be varied systematically. A third
se can be studied by comparing performance on positive, negative,
method (Gross & Levenson, 1995) uses videos that represent eight
and neutral lists whose arousal levels have been equated (Brainerd,
discrete emotions (amusement, anger, contentment, disgust, fear,
Stein, et al., 2008). Second, by manipulating valence and arousal
neutral, sadness, and surprise).
factorially, it is possible to study whether false memory is affected
Subjects’ moods are measured to check the effectiveness and
by both valence and arousal and whether the two interact (Brainerd
durability of induction. The modal method is to administer the
et al., 2010).
Positive and Negative Affect Schedule (PANAS; Watson, Clark,
Another technique for manipulating the emotional content of
& Tellegen, 1988), which asks subjects to indicate the degree to
target materials is via pictures that have been normed for valence
and arousal, such as the IAPS (Lang et al., 2008), which contains which they have felt emotions that are described by a series of
color photographs of people, animals, objects, and scenes, rated on words (e.g., excited, scared, irritable, inspired) during a specified
valence and arousal using the same scales as the ANEW. Another period of time (e.g., right now, today, in the past week). The
example is the GAPED (Dan-Glauser & Scherer, 2011), which is PANAS was developed for nonclinical populations (Crawford &
a similar picture database. An important feature of the GAPED is Henry, 2004), and disorder-specific instruments are available for
that it contains thematic groups of pictures (e.g., several babies, clinical populations (e.g., the Hamilton Depression Scale; Ham-
several dangerous animals) that differ in valence and arousal. ilton, 1960). Other mood measures have occasionally been
Similar to the DRM lists in Table 4, these picture norms can be administered in false memory experiments—such the Satisfac-
used to construct lists in which valence and arousal are counter- tion with Life Scale (Diener, Emmons, Larsen, & Griffin, 1985;
balanced. Koo & Oishi, 2009), the Brief Mood Introspection Scale (BMIS;
A final, rarely used, technique involves pictures of human faces Mayer & Gaschke, 1988), and the Self-Assessment Manikin (Bradley
that express various emotions. There are many databases of such & Lang, 1994)— but the PANAS predominates.
pictures, but most have not been normed. A few have been normed Correlational studies. Here, measures of naturally occurring
for valence (e.g., Langner et al., 2010), but none has been normed moods are added to otherwise standard designs, and the scores are
for arousal. then correlated with measured levels of false memory. This has
Manipulating and measuring mood. Both experimental and been done for two broad types of populations: the usual subjects in
correlational designs have been used to investigate how false memory research (undergraduates) and special populations that are
memory is influenced by mood. In experimental work, mood- characterized by particular moods [for example, depressed indi-
induction tasks are administered to distinct groups of subjects, viduals, individuals with Posttraumatic Stress Disorder (PTSD)].
whereas in correlational work, subjects’ natural moods are mea- As in mood-induction experiments, the PANAS has been the most
sured. We sketch the basic characteristics of each type of design. common instrument for measuring naturally occurring moods.
Table 4
Emotional DRM Lists (Arousal Controlled) From Brainerd et al., 2010
Negative/High arousal Negative/Low arousal Positive/High arousal Positive/Low arousal
Anger Cold Board Dirt Baby Beach Bird Fish

Mad Hot Wood Mud Child Sand Nest Water
Fear Snow Chalk San Cry Sun Parrot Swim
Hate Warm Surf Clean Small Ocean Eagle Sea
Rage Winter Room Soil Bottle Water Feather Scales
Temper Ice Game Dirty Cute Towel Animal Smell
Fury Wet Nail Filth Infant Ball Wings Cod
Ire Frigid Plane Ground Mother Summer Canary Food
Wrath Chilly School Bath Crib Girls Cage Trout
Happiness Heat Ship Black Doll Swim Sing Gold
Fight Weather Walk Grime Little Fun Dog Ocean
Hatred Freeze Two-by-four Bike Boy Wave Flight Fry

Hostile Air Black Brown Adult Blanket Song Shark

Calm Shiver Boat Grease Diaper Bum Blue Jay Tank
Emotion Arctic Build Mess Girl Coast Tree Bass
Enrage Frost Unclean Young Short Robin Bowl
Dead Hurt Fall Fat City God Grass Green
Alive Pain Spring Skinny Town Jesus Green Grass
Gone Cut Leaves Thin Crowded Heaven Cut Yellow
Cold Ouch Winter Cat State Love Lawn Blue
Funeral Cry Down Ugly Capital Almighty Mow Color
Asleep Injury Autumn Lady Streets Church Hopper Red
Bury Life Hurt Big Subway Lord Pot Leaf
End Sick Frosty Cow Country Religion Mower Arrow
Grave Fall Summer Slim New York Creator Weed Clover
Sad Harm Catch Albert Village All-knowing Grow Frog
Still Hit Cool Slob Metropolis Bible High Vegetables
Animal Sore Trip Boy Big Damn Roots Apple
Cemetery Stop Drop Man Chicago Devil Smoke Bean
Deceased Wound Rise Pig Suburb Faith Turf Elf
Die Upset Stand Diet County Good Yard Go
Duck Ache Trees Large Urban Money
Needle Rough Glass Sick Hug Love Nice River
Thread Smooth Window Ill Kiss Hate Mean Water
Pin Bumpy Cup Well Love Kiss Sweet Stream
Eye Road Drink Cold Squeeze Like Good Lake
Sewing Tough Break Hospital Embrace Happy Bad Mississippi
Sharp Sandpaper Clear Vomit Affection Heart Kind Boat
Point Jagged Jar Doctor Hold Care Pleasant Tide
Prick Ready Crystal Bed Cuddle Admire People Canoe
Thimble Coarse Cut Flu Arms Adore Friendly Flow
Haystack Uneven Bottle Illness Bear Close Smile Run
Thorn Riders Fragile Fever Caress Friendship Happy Barge
Harm Rugged Plastic Nurse Tight Girl Friend Creek
Injection Sand Sharp Medicine Touch Happiness Pretty Brook
Syringe Beard Eye Nausea Warmth Hug Nasty Bend
Cloth Ground House Virus Clinch Kindness Delightful Bridge
Knitting Gravel Juice Pills Clutch Life Polite Wind
Spider Thief Shy Trash Music Pretty Sleep Soft
Web Steal Quiet Garbage Note Ugly Bed Hard
Insect Robber Outgoing Waste Sound Girl Rest Light
Bug Crook Timid Can Piano Beautiful Awake Pillow
Fright Burglar Bashful Refuse Sing Cute Tired Plush
Fly Money Scared Sewage Radio Adorable Dream Loud
Arachnid Cop Bold Bag Band Attractive Wake Cotton
Crawl Police Person Junk Melody Flower Snooze Fur
Tarantula Rob Introvert Rubbish Horn Gorgeous Blanket Touch
Poison Jail People Sweep Concert Hot Doze Fluffy
Bite Gun Me Scraps Instrument Model Slumber Warm
Creepy Villain Loud Pile Symphony Pink Snore Furry
Black widow Crime Afraid Dump Jazz Appealing Nap Downy
Monkey Bank Boring Landfill Orchestra Sweet Peace Kitten
Feelers Bandit Coy Debris Art Charming Yawn Skin
Criminal Extrovert Litter Rhythm Lovely Drowsy Tender
Target materials may be neutral, or their emotional content can a crime, attention to the weapon is obligatory but is optional to
be varied. An advantage of the latter procedure is that a basic other details (e.g., the appearance of the perpetrator).
memory law, the encoding-specificity principle (Tulving & Thom- More recently, Kensinger, Garoff-Eaton, and Schacter (2007)
son, 1971), should operate, according to which episodic traces of proposed that memory for negative elements of negative events is
targets include contextual details from the study phase. Mood is a enhanced but memory for neutral events is impaired. They also
salient contextual detail, and by varying targets’ emotional con- proposed a verbatim/gist tradeoff such that emotion enhances
tent, it can match or mismatch emotional content. According to memory for the gist of experiences but impairs memory for ver-
encoding specificity, that should produce mood congruency effects batim details. The general pattern that is expected is that (negative)
(better memory when mood and content match; Bower, 1981). emotional content will reduce false memory for certain compo-
Such effects have been investigated in correlational studies of both nents of an event, which can be specified on theoretical grounds,
nonclinical and clinical samples, typically by administering emo- but will increase it for other components of the same event. Mather
tional DRM lists. Mood congruency effects could just as easily be (2007) extended this line of reasoning, proposing that the organi-
investigated in mood-induction experiments, by varying the emo- zation of an emotional event determines its effects on memory.
tional content of targets, and we shall see that the literature Emotional content is assumed to attract processing attention, and
contains a few experiments of that ilk. any details that are “considered to be in the same object as the
arousing item” (Mather, 2007, p. 36) will receive increased pro-
cessing. This suppresses memory distortion for those details, but it
Theoretical Conceptions of Emotion-False Memory increases false memory for other details because processing atten-
tion is drained away from them. Support for this hypothesis was
Extant proposals about real-world instances of emotional false
obtained in some experiments reported by Doerksen and Shima-
memory suggest that the relation between emotion and false mem-
mura (2001).
ory may not be straightforward. A prime example is complemen-
Emotional enhancement of memory (EEM; Heuer & Reisberg,
tary hypotheses about two circumstances that are laced with highly
1990) is a broader proposal to the effect that emotional content
arousing negative emotion: flashbulb memories of major disasters,
enhances overall memory accuracy. For example, Heuer and Reis-
such as the Challenger explosion and the 9/11 attack, and memo-
berg suggested that the presence of a weapon in a crime enhances
ries of crimes during police interrogations. In the former, people
memory for peripheral details as well as for the weapon. Other
often report detailed memories of these events many years after-
evidence in line with EEM was reported by Kensinger and Corkin
ward, and it has been claimed that the events’ emotional qualities (2003) and by Ochsner (2000). At present, EEM effects are
make it very unlikely that those memories are distorted (e.g., thought to be attributable to a mix of factors that includes en-
Brown & Kulik, 1977). In police interrogations, on the other hand, hanced physiological responsiveness and enhanced semantic relat-
it has been claimed that the same qualities produce high levels of edness, as well as heightened distinctiveness and personal signif-
distortion in memory for details of crimes (e.g., Brainerd & Reyna, icance. Talmi and Moscovitch (2004), for instance, reported some
2005) and the physical appearance of perpetrators (e.g., Kassin, influential experiments on the semantic properties of emotional
2001). content.
There are also complementary hypotheses about memories of FTT’s opponent process distinctions have been used to unpack
emotionally arousing medical symptoms. Some have proposed that the effects of different components of emotional content on false
emotion makes patients’ memories for such things as past preg- memory (e.g., Bookbinder & Brainerd, 2016; Brainerd, Stein, et
nancies, children’s health care, and AIDS treatment highly resis- al., 2008; Rivers, Reyna, & Mills, 2008). This analysis focuses on
tant to distortion (D’souza-Vazirani, Minkovitz, & Strobino, 2005; the valence and arousal dimensions, making predictions about how
Githens, Glass, Sloan, & Entman, 1993; Weissman et al., 1996), each affects retrieval processes that control false memory and
but others have proposed that emotion makes memories for care providing a framework for explaining differences in the effects of
received during counseling, number of doctor visits, and hospital discrete emotions (e.g., anger and fear). Recall that in the opponent
stays quite error prone (Raina, Torrance-Rynard, Wong, & Wood- process approach, strengthening gist traces foments false memo-
ward, 2002; Tisnado et al., 2006; Wolinsky et al., 2007). The ries, whereas strengthening verbatim traces suppresses them. Emo-
denouement is that complementary hypotheses are a feature of tional content is assumed to affect both, so that whether it increases
current theoretical ideas about emotion-false memory. Beyond or decreases false memory, relative to neutral content, depends on
this, extant hypotheses are somewhat different for emotional con- whether it has more pronounced effects on the gist or verbatim
tent versus mood, and hence, we sketch them separately. side. FTT makes three specific proposals about valence and two
Content hypotheses. Early proposals were made by Easter- about arousal.
brook (1959), who argued that increased levels of emotion in The first valence proposal is that content that is either positively
events reduce the amount of information that subjects can attend valenced (words such as baby, dream, and sweet) or negatively
to, which impairs later memory. This hypothesis has been studied valenced (words such as death, mad, and spider) strengthens gist
by varying the emotional content of central and peripheral details traces, relative to neutral content, by increasing semantic connec-
of target events (Christianson & Hubinette, 1993; Christianson & tions among target events. Second, semantic connections are more
Loftus, 1990), with the general result that only memory for pe- salient for negative than for positive valence, so that other things
ripheral events supports the hypothesis. This may mean that atten- being equal, gist traces are strengthened more by negative than by
tion is optional for peripheral events but obligatory for central positive valence. Together, these two ideas seem to predict that (a)
ones. Loftus (1979) introduced a well-known forensic example of false memories will be increased by both positively and negatively
this point, weapon focus: When a dangerous weapon is present in valenced content and (b) the increase will be greater for negative
valence. However, that ignores how verbatim memory is affected The predictions of the affect-as-information hypothesis are
by emotional content. Here, third, FTT assumes that positive straightforward because the item versus relational processing dis-
valence strengthens verbatim traces, relative to neutral content, but tinction is known to predict that false memory will increase as
negative valence weakens them. This proposal falls out of a relational processing increases, but that it will decrease as item
substantial prior literature showing that subjects’ ability to retrieve processing increases (e.g., Arndt, 2006). Therefore, positive
vivid, detailed memories of events is enhanced by positive valence moods should increase false memory because focusing on relations
but impaired by negative valence (for reviews, see Gomes, Brain- among targets (“tree” for oak and pine, “animal” for cow and
erd, & Stein, 2013; Kensinger, 2004). Thus, the clear prediction horse, “color” for blue and red) supports false memory for related
about negative valence is that it will increase false memories distractors (maple, sheep, green), but negative moods should de-
because it strengthens gist traces while weakening verbatim traces, crease false memory because focusing on unique features of indi-
whereas in contrast, the prediction about positive valence is that its vidual targets suppresses false memory for related distractors (“I
effect on false memory will depend on whether a particular ma- could not have read maple because I clearly remember reading oak
nipulation strengthens verbatim or gist memory more. and pine instead.”).
Turning to arousal, FTT assumes that it chiefly affects verbatim In addition to Bower’s (1981) mood congruency principle and
memory (Brainerd et al., 2010; Rivers et al., 2008). Moderate the affect-as-information hypothesis (see Table 2 for a comparison
levels of arousal, relative to baseline, can enhance attention to of the two theories), researchers have borrowed theoretical distinc-
target events and strengthen verbatim traces, but high levels are a tions from the larger false memory literature and adapted them to
form of cognitive noise that blurs attentional focus and weakens mood differences in false memory. For example, Ruci, Tomes, and
verbatim traces (Yerkes & Dodson, 1908). Consequently, if va- Zelenski (2009) explained increased false memory in mood-
lence is controlled, increases in arousal can suppress or increase congruent conditions by combining Bower’s theory of mood con-
false memory, depending on how high the levels of arousal are. gruency with a spreading activation account that is often applied to
Last, in connection with specific emotions, emotions of the the DRM illusion (McDermott & Watson, 2001). Others have
same qualitative valence may differ in just how strong that valence attributed mood differences in false memory to increased arousal
is and also in levels of arousal. For instance, on the ANEW norms (e.g., Corson & Verrier, 2007) or to organizational differences
(e.g., Schwartz, 1975). However, there are no dominant mood
(Bradley & Lang, 1999), anger is rated as both more negative and
theories beyond the first two that we described, and hence, we rely
more arousing than fear (2.76 vs. 2.34 and 7.17 vs. 6.96). Hence,
primarily on those theories in interpreting current mood-false
FTT simply uses its proposals about gist-verbatim effects to pre-
memory research.
dict differences in false memory effects. However, specific emo-
tions may be roughly comparable in valence and arousal, and when
they are, their false memory effects may still differ. Here, FTT Review of Data I. Emotional Content and False
does not make differential predictions, but it uses opponent process Memory
distinctions to explain any observed differences.
Now, we summarize the main patterns that have been reported
Mood hypotheses. Other hypotheses about how emotion af-
in emotional content experiments. We separate these studies ac-
fects false memory have been formulated for mood. They are less
cording to their focal form of false memory and emphasize con-
extensive that those for emotional content and can be traced to
vergent findings, beginning with studies of emotional DRM ma-
Bower’s (1981) landmark studies of how mood affects true mem-
terials, continuing with studies of emotional pictures and scripted
ory. Bower proposed a network theory of emotion that featured a
material, and concluding with misinformation studies. In the next
mood congruency principle: Memory for target events is enhanced section, we do likewise for research on the relation between false
when subjects’ moods match during study and test. memory and subjects’ moods.
In recent years, hypotheses have emerged that describe memory
consequences of specific moods. The affect-as-information hy-
pothesis (Schwarz & Clore, 1983; Schwarz, 1990) is the most DRM Experiments
influential example, according to which mood affects the nature of Variants of this procedure, in which the valance and/or arousal
encoding during the study phase—in particular, subjects’ reliance of critical distractors (CDs) and list words were manipulated, have
on what Hunt and Einstein (1981) termed item versus relational been used to measure the distortive effects of emotional content. In
processing. Positive moods are assumed to promote relational the first group of experiments below, false memory for negative,
processing, which focuses on connections among target events arousing CDs was compared with false memory for neutral CDs.
(e.g., their meanings), and negative moods are assumed to promote In a second and smaller group of studies, false memory for
item processing, which focuses on targets’ distinguishing features positive, negative, and neutral CDs was compared with their
(e.g., their surface structures). In that vein, Bless et al. (1996) respective arousal levels controlled.
predicted that happy moods increase reliance on general knowl- False memory for negative, arousing content. Budson et al.
edge structures (e.g., schemata), relative to sad moods. The dis- (2006) developed the DRM lists at the top of Table 3, with
tinction between item and relational processing resembles FTT’s emotional lists being both more negative and more arousing than
verbatim-gist distinction. The key difference is that one is con- neutral ones, which are a mixture of neutral and positive lists. The
cerned with subjects’ processing and the other is concerned with basic elements of their design are shown in the left column of
the types of traces that they store, but obviously, verbatim traces Figure 1. Other experiments that are reviewed in this section
should be favored by item processing and gist traces by relational implemented variations of Budson et al.’s design, with their key
processing. features being shown in the remaining columns of Figure 1.
Immediate Immediate
Recognition
recall & Directed Final recall
only: Budson recall only: and
recognition: forgetting:
et al. (2006);
Howe (2007);
Palmer & recognition: El
Pesta et al. Howe et al. Dodson Sharkawy et
Howe et al.
(2001)
(2010) (2011) (2009) al. (2008)
List 1 List 1 List 1
Recall DF Recall
All All
… List 2 …
lists lists
List 9 Recall List 9
Recall Recall Recall
Recog Recog Recog No

Recall
nition nition nition recog
(all
(all (all (both nition
lists)
lists) lists) lists) test
Recog
nition
Figure 1. Variants of the DRM paradigm in studies of emotional content effects. See the online article for the
color version of this figure.
Budson et al. (2006) administered their materials to healthy Roediger, 2002). Although the lack of counterbalancing is a design
younger and older adults and to older adults with Alzheimer’s flaw from one perspective, it was motivated by an important
Dementia (AD). They found that emotional content did not affect real-life situation in which recall precedes recognition—namely,
false memory in any of these groups, as shown in Panel A of best-practice forensic interviewing protocols (see Brainerd &
Figure 2. Emotional content did affect true memory and response Reyna, 2005).
bias, but those effects were inconsistent across the three groups. It Howe, Toth, and Cicchetti (2011) repeated this procedure with
should be noted, however, that Budson et al.’s subject samples samples of children (ages 6 –12) and added a directed forgetting
were small (19 –20 per group), and the number of emotional CDs condition, to measure subjects’ ability to inhibit false memories.
per subject was modest (10). The effect sizes for this study and all False recall and true recall were again higher for neutral than for
subsequent ones that reported means and standard deviations can emotional lists (Panel C of Figure 2 displays Howe’s and Howe et
be found in Table 1. al.’s results), but Howe et al. did not report their recognition data.
Howe (2007) soon reported more promising results with larger In addition, they found that although children were less likely to
samples of children (ages 8 and 12). He obtained reliable differ- report emotional false memories than neutral ones, children were
ences in false memory for emotional versus neutral CDs, but the less likely to forget emotional false memories once they had been
differences interacted with the type of memory test. On recognition reported. It might appear that Howe and colleagues’ (2007) and
tests, the false alarm rate was higher, whereas on recall tests, the (2011) detection of emotional false memory in children, compared
intrusion rate was lower for emotional than for neutral CDs. Also, with Budson et al.’s (2006) lack of effects in adults, confirms the
true recognition and true recall were both better for neutral lists. commonsense idea that false memories are more common in
Obviously, these results do not square with any of the earlier children than adults. This not the case, for two reasons. First, the
hypotheses about content-false memory, but they may must be small sample size and modest number of replications in Budson et
cautiously interpreted because the order of recall and recognition al.’s study may be responsible for the lack of effects, and second,
tests was not counterbalanced. Recall always preceded recogni- we are about to see that patterns similar to those of Howe and
tion, and other DRM research has shown that recognition perfor- colleagues have been obtained with adults.
mance is affected by whether or not subjects respond to prior recall In the latter connection, Palmer and Dodson (2009) found such
tests (Brainerd, Yang, Reyna, Howe, & Mills, 2008; Gallo & patterns when they administered negative, neutral, and positive
A "Emotional" (Negative) Neutral
0.8
C 0.5 "Emotional" (Negative) Neutral
0.6
0.4
Recognition
0.3
Recall
0.4
0.2
0.2
0.1
0 0
Budson Older Budson Younger Howe 2010 Howe 2010 Pesta Young Howe 2007 Palmer & Howe 2010 Howe 2010 Howe 2011 Howe 2011
Adults Adults Young Adults Children Adults Children Dodson 2009 Young Adults Children Younger Older adults
Young Adults adults

Negative Neutral Positive

D
B Negative Neutral Positive 0.5
0.8
0.4
0.6
0.3
Recognition
Recall
0.4
0.2
0.2
0.1
Figure 2. False recall and recognition rates in the DRM paradigm. Panel A ⫽ False recognition, arousal not
controlled. Panel B ⫽ False recognition, arousal controlled. Panel C ⫽ False recall, arousal not controlled. Panel
D ⫽ False recall, arousal controlled.
DRM lists to young adults (18 –22 years; see the middle of Table tended them to young adults, and they also replicated and extended
3). However, valence and arousal values were not reported for their earlier recall results (i.e., lower true and false recall for
these lists. Palmer and Dodson also found that both true and false negative lists; see Panels A and C of Figure 2). Qualitatively, then,
recall were higher for neutral than for negative or positive lists emotion-false memory effects were similar in adults and children.
(Panel C of Figure 2), and they explained their false memory data However, effect sizes (see Table 1) reveal that adult effects are
with a variant of the item versus relational processing hypothe- consistently large, but child effects are moderate and depend on the
sis—that emotional words, regardless of their valence, promote type and timing of memory tests.
item-specific processing and draw processing attention away from In some experiments, Howe et al. measured forgetting on rec-
semantic relations among words. That, in turn, weakens the rep- ognition tests. The well-established pattern in the larger literature
resentational basis for false memory (gist traces). Notice that this (Brainerd, Reyna, & Brandse, 1995; Payne et al., 1996) is that the
does not explain why true recall was also lower for emotional lists hit rate drops sharply while the false alarm rate for related distrac-
because the expectation would be that item-specific processing tors remains relatively stable, over intervals of a few days to a
strengthens verbatim traces and, hence, increases true memory. week. Howe et al. observed both effects with emotional and
El Sharkawy, Groth, Vetter, Beraldi, and Fast (2008) used neutral lists, and further, they observed that the emotional false
Howe’s (2007) basic methodology with a sample of young adults memory effect increased as time passed. Of course, this sleeper
but obtained different results. For recall, neither true nor false effect is consistent with the opponent process notion that negative
memory differed for emotional versus neutral lists. For recogni- content produces stronger gist traces and weaker verbatim traces,
tion, however, emotional lists increased the false alarm rate for and that the latter fade more rapidly.
CDs but decreased the hit rate for targets. This pattern, which is In a remaining study, Pesta et al. (2001) created DRM-like lists
consistent with opponent-process theories, seems to hold across on which the words resembled each other in sound rather than
child and adult subject samples. However, as in Howe’s research, meaning (e.g., fell, dwell, shell, will, cell; CD ⫽ well). Some lists
recognition tests were always preceded by recall tests. were neutral but others were negative and arousing (see Table 3).
Next, Howe et al. (2010) used the same procedure with young There were two key results. First, the false alarm rate was lower
adults and children (ages 5–11) in a series of experiments. They for negative than for neutral CDs (Panel A of Figure 2), but the hit
replicated their earlier recognition results with children and ex- rate was higher. This is not inconsistent with prior experiments
showing higher false alarm rates and lower hit rates for negative parameters separate the contributions of verbatim and gist retrieval
lists because those experiments involved semantically related ma- (Brainerd, Reyna, & Mojardin, 1999; Brainerd, Stein, & Reyna,
terials. Second, when the relative distinctiveness of negative and 1998) to true and false memory. This technique measures the
neutral lists was equated, the false alarm rate was higher for effects of emotional content at the level of memory processes,
negative CDs. Remember, here, that opponent-process theories rather than raw performance, which delivers direct tests of differ-
and the item-relational processing distinction both assume that ent theoretical hypotheses about emotional content. The results are
distinctiveness amplifies processes that work against false memory shown in Figure 3 (panels A and B), where it can be seen that with
(i.e., verbatim traces, item-specific processing). CDs, negative valence increased gist retrieval and decreased ver-
Summing up the results thus far, the most consistent pattern is batim retrieval, relative to positive and neutral valence. Also with
that negative, arousing DRM lists produce higher levels of false CDs, positive valence increased verbatim retrieval and decreased
memory and lower levels of true memory than neutral ones on gist retrieval, relative to neutral lists. It can be seen that emotional
recognition tests. Also, recall tests mostly produced a different content effects were the same at the gist level but were different at
pattern in which true and false recall were both lower for negative, the verbatim level: Negative valence increased both gist and ver-
arousing lists. Two notable points of difference between recogni- batim retrieval, and positive valence decreased both verbatim and
tion and recall may explain these different patterns: (a) Recogni- gist retrieval. This suggests the theoretically important possibility
tion tests are more sensitive to gist memory, whereas recall tests that the effects of emotional content may depend on whether
are more sensitive to verbatim memory (Seamon et al., 2003), and verbatim traces are available for test cues because that is the key
(b) recognition is more sensitive to memory for contextual details memory difference between targets and related distractors.
that accompanied target events, with emotional reactions being Dehon, Laroi, and Van der Linden (2010) also administered
examples of such context details (Brainerd, Gomes, & Moran, positive, negative, and neutral DRM lists with matched arousal
2014). In the larger false memory literature, these differences are levels, and they found higher false memory for negative than for
known to produce variability in how false memory is affected by neutral lists, on both recall and recognition tests, as displayed in
the same experimental variable on recognition versus recall tests panels B and D of Figure 2. Unlike Brainerd, Stein, et al. (2008),
(for a review, see Brainerd & Reyna, 2005), and for such variables, they also found higher false memory levels for positive than for
the exact reasons must be identified with surgical manipulations. neutral lists, on both recall and recognition tests. True memory was
Emotional content may simply be another example. not affected by valence. This finding that negative valence elevates
This is subject to the proviso that recognition tests were always false memory for recall as well as false recognition raises the
confounded with prior recall tests in the above studies (see Figure possibility that Howe and associates recall results may be arousal
1), and recall tests can alter the contents of memory. Whether effects rather than valence effects.
unconfounded recognition tests produce different emotional con- Factorial manipulation of valence and arousal. At the time
tent effects than recall tests therefore is an open question, but it of this writing, there was a single published experiment in which
should be reiterated that the procedure of administering recall tests the valence and arousal of DRM lists had been manipulated
before recognition tests was a deliberate choice that was motivated factorially. Brainerd et al. (2010) did that by administering the lists
by forensic interviewing practices. in Table 4 to samples of children (age 7–11) and young adults,
Valence comparisons with arousal controlled. In the pre- followed by recognition tests. In adults, false memory exhibited
ceding studies, one cannot tell whether emotional content effects both valence and arousal effects; it was higher for negative than for
were attributable to valence, arousal, or both. Another complexity positive valence and for high-arousal than for low-arousal. Arousal
is that Budson et al.’s (2006) neutral lists are not entirely neutral; also interacted with valence such that it only influenced false
some are clearly positive. Hence, one also cannot tell whether memory when valence was negative. There were also valence and
emotional content effects were attributable to the negative-neutral arousal effects on the true memory side, with the hit rate being
difference or the negative-positive difference. higher for positive lists and high-arousal lists. The most important
To remove those limitations, Brainerd, Stein, et al. (2008), new result is that false memory may be more strongly influenced
developed pools of negative, positive, and neutral lists in such a by valence than by arousal.
way that the three groups of lists and their CDs varied in valence As for children, their levels of false memory were lower overall,
but were equated on arousal. The lists were administered to large which is another example of the well-documented but counterin-
samples of young adults, followed by recognition tests. Three tuitive developmental reversal effect in semantic false memory
results were consistent across their experiments. First, false alarms (for a review, see Brainerd, Reyna, et al., 2008). Between 7 and 11,
to CDs were higher for negative lists than for positive or neutral age increases in false memory were greater for negative than
lists and were higher for neutral than for positive lists (Panel B of positive valence, arousal effects became more marked, and like
Figure 2). Second, the hit rate was lower for positive lists than for adults, arousal effects were confined to negative lists. In short,
neutral or negative lists. Third, valence effects were much larger emotional content effects were qualitatively similar in children and
for false than for true memory. Obviously, the first finding sug- adults but were larger in adults.
gests that the higher false recognition levels for negative, arousing Summary of DRM findings. On the overriding question of
CDs in Howe and associates’ experiments is a real valence result how false memory reacts to emotional content, DRM studies have
that is not attributable to testing or arousal confounds. In contrast, produced a consistent answer for recognition, at least: Negative
the second finding suggests that data showing lower hit rates for valence increases false memory, relative to neutral and positive
negative, arousing lists may not be real valence results. valence, regardless of lists’ arousal level or whether recognition is
A useful feature of Brainerd, Stein, et al.’s (2008) research is confounded with prior recall. The data are nearly as consistent in
that the data were analyzed with a mathematical model whose showing that negative valence lowers true memory. The combined
A C
0.5
0.8
0.45
0.7 0.4
0.6 0.35
0.5 0.3
Neutral
Neutral 0.25
0.4 Negative
Negative 0.2
0.3 Positive
Positive 0.15
0.2
0.1
0.1 0.05
0 0
Recollection Phantom Similarity Identity Similarity
Rejection Recollection
B D
0.45
0.9
0.4
0.8
0.35
0.7
0.3
0.6
0.25 Neutral
0.5 Neutral
0.2 Negative
0.4 Negative
Positive
0.3 0.15
Positive
0.2 0.1
0.1 0.05
0 0
Recollection Phantom Similarity Identity Similarity
Rejection Recollection
Figure 3. Parameter estimates of the conjoint recognition model by valence for the two experiments in
Brainerd et al. (2008). Panel A ⫽ Experiment 1, false memory. Panel B ⫽ Experiment 2, false memory. Panel
C ⫽ Experiment 1, true memory. Panel D ⫽ Experiment 2, true memory.
effect is that in recognition, negative emotional content substan- semantic relations between targets and distractors is preserved.
tially decreases overall accuracy. Notice that this has clear impli- Because it has long been known that subjects use different words
cations for the aforementioned forensic claims that witnesses’ to label the same picture (Paivio, 1971), pictures are often pre-
memories for negative-arousing events (crimes) are unusually ac- sented with accompanying verbal labels at study. From the per-
curate and resistant to distortion; the data suggest the opposite and spective of opponent-process distinctions, an important difference
also that memory distortion for negative events increases with between pictures and DRM lists is the finding that pictures en-
arousal. Other results suggest that emotional content effects may hance the processing of targets’ verbatim content, relative to word
be dependent on whether verbatim traces are available for test lists (Brainerd & Reyna, 2005; Schacter et al., 1999).
items because the effects were different for targets and distractors In an early study, Gallo, Foster, and Johnson, (2009) presented
when verbatim and gist retrieval were directly measured. older and younger adults with positive, negative, and neutral IAPS
Conclusions about recall are more uncertain. Negative-arousing pictures, accompanied by verbal labels and followed by recogni-
content has been found to decrease intrusions and increase true tion tests (see Figures 4 and 5 for an explanation of their design).
recall in multiple studies, but because arousal was always con-
Arousal was equated for positive and negative pictures (both were
founded with valence, it is unclear whether this effect is attribut-
highly arousing) but not for neutral ones (low-arousal). On the
able to valence, arousal, or some interaction between them. That
recognition test, the targets were presented words that had labeled
concern is reinforced by the fact that results were different in the
the pictures, and the distractors were unpresented words that were
only study that eliminated this confound.
labels of unpresented pictures. False recognition of related distrac-
tors was higher for both positive and negative pictures than for
Picture Experiments
neutral ones but did not differ for positive versus negative pictures.
Here, study and test materials are vivid depictions of real-world Note that this could be an arousal effect, a valence effect, or both.
events, but the basic feature of false memory being grounded in For true memory, however, there was an unambiguous valence
Figure 4. Experimental design in Gallo, Foster, & Johnson (2009).
effect: The hit rate was higher for negative than for positive or An Experiment With Scripted Material
neutral pictures, which contrasts with the true memory valence
effect in DRM studies. Consistent with the larger false memory Mirandola, Toffalini, Grassano, Cornoldi, and Melinder (2014)
literature, pictures seemed to enhance verbatim processing of investigated emotional false memories with a procedure developed
targets: With neutral pictures, when performance was compared by Hannigan and Reinitz (2003). Subjects viewed pictures of
with a condition in which only neutral words were studied and scripted events (e.g., grocery shopping, with someone taking an
tested, false alarm rates were lower and hit rates were higher for orange from the bottom of a pile), along with outcomes whose
pictures. direct causes were not presented (picking up fallen oranges). In
In a follow up study, Choi, Kensinger, and Rajaram (2013) used Mirandola et al.’s version of this procedure, the outcomes were
thematically related groups of emotional words and pictures to either negative-arousing or neutral-nonarousing ones. False mem-
determine whether valence helps subjects connect shared meaning ory was tested with pictures of unpresented causes of outcomes
across members of a thematic group. Arousal was equated for and pictures of other unpresented script-consistent events (e.g.,
positive and negative pictures (high arousal) but not neutral ones looking at milk in the dairy cooler). Mirandola et al. were also
(low arousal). They tested recognition of words that had been interested in the effects of elaboration, so they included a manip-
studied with pictures, varying the retention interval and whether ulation in which half of their subjects wrote descriptions of the
thematic labels had accompanied the pictures. On an immediate test scripted events before the recognition test and half did not.
that included thematic labels, negative valence increased true memory Interestingly, emotional content effects varied as a function of
but did not affect false memory. After 24 hours, negative valence still elaboration such that the performance of subjects who wrote
increased true memory, but now it decreased false memory. In a final descriptions was dominated by gist retrieval. Those subjects (but
experiment, Choi et al. eliminated thematic labels and administered not others) displayed more false memory for negative-arousing
immediate recognition tests, but emotional content effects did not events than for neutral ones, and they displayed more true memory
change. Because negative valence consistently elevated true memory for script-consistent targets that for script-inconsistent targets.
and sometimes lowered false memory, the indicated conclusion is that When the elaboration group completed a free-recall task for the
negative pictures enhanced verbatim processing, relative to gist pro- scripted information, they recalled more negative-arousing themes
cessing. than neutral ones.
In a slightly different paradigm, Bookbinder and Brainerd
(2016) administered positive, negative, and neutral IAPS pictures
Misinformation Experiments
with equivalent arousal levels. The pictures belonged to several
categories, to promote gist processing, but they were not accom- Porter, Spencer, and Birt (2003) investigated whether misinfor-
panied by verbal labels. On both immediate and 1-week delayed mation effects were influenced by the emotional content of target
recognition tests, negative valence elevated true and false memory, events. Their subjects studied positive, negative, or neutral IAPS
but true memory declined markedly with time, particularly for pictures (see Figure 6 for similar images), each of which contained
negative valence. The data were analyzed with the same mathe- both central and peripheral details. Similar to Gallo et al.’s (2009)
matical model that Brainerd, Stein, et al. (2008) used. As with study, arousal was equated for positive and negative pictures (both
DRM lists, negative valence decreased verbatim retrieval and were highly arousing) but not neutral ones (nonarousing). During
increased gist retrieval for related distractors, compared with pos- the misinformation phase, subjects were misled about both central
itive and neutral valence. For targets, negative valence increased and peripheral details. On subsequent recall tests, misinformation
gist retrieval and reduced a type of verbatim memory that causes effects were detected for both, and there were clear valence effects:
incorrect rejection of targets. Thus, on the false memory side, these Misinformation effects were more marked for negative central and
picture data are qualitatively similar to Brainerd, Stein, et al.’s peripheral details than for positive or neutral ones. For true mem-
DRM data when it comes to the process-level effects of valence. ory, however, valence interacted with central-peripheral details,
Figure 5. Sample images with verbal labels similar to Gallo, Foster, and
Johnson (2009). From “Close-Up of Astronaut,” “Shark,” and “Electric
Iron” by Arora (2011), O’Rockerz (2006), and Syed (2015), respectively
(http://freeimages.com). In the public domain. See the online article for the
color version of this figure.
such that more central details were recalled for positive valence
but more peripheral details were recalled for negative valence.
As misinformation is an analogue of suggestive police question- Figure 6. Sample images with central and peripheral details similar to
ing, Porter et al. examined whether valence affected what they Porter, Spencer, and Birt (2003). From “Wedding Ceremony,” “Ready to
termed “major misinformation” effects—which meant falsely re- Shoot,” and “Shopping” by Gomez (2005), Lucretious (2006), and Niewia-
membering major details that were not present in the studied domski (2015), respectively (http://freeimages.com). In the public domain.
pictures. Major misinformation effects proved to be more common See the online article for the color version of this figure.
for subjects with negative valence (80% of subjects falsely remem- one for the DRM illusion—namely, that negative valence increases
bered such events) than with positive or neutral valence, under- susceptibility to false memory, relative to positive and neutral
scoring the apparent tendency of negative valence to heighten valence, and less often, it impairs true memory. As before, such
susceptibility to suggestion. results bear on the aforementioned forensic claims that memory for
In a follow-up, Porter, Bellhouse, McDougall, Ten Brinke, and negative events is unusually accurate and resistant to distortion.
Wilson (2010) again investigated whether valence influenced the Contrary to such claims, the modal pattern is that negative valence
magnitude of major and minor misinformation effects across dif- reduces memory accuracy in two ways, by amplifying false mem-
ferent time periods. They administered arousal-matched positive ory and reducing true memory.
and negative IAPS pictures but not neutral ones. As in the earlier We noted earlier that misinformation effects are different than
study, major misinformation effects were more common with spontaneous false memories in the sense that the former can be
negative than with positive valence, but there were no valence attributable to either verbatim or gist retrieval, whereas gist re-
effects for minor misinformation. Misinformation effects de- trieval predominates with the latter. All findings considered, the
creased over time, but they were still present after a month. simplest theoretical interpretation of the tendency of negative
Porter and associates’ key finding that negative valence height- content to amplify misinformation effects and reduce true memory
ens susceptibility to suggestion was also obtained by Otgaar, relies on verbatim retrieval. This pattern would result if negative
Candel, and Merckelbach (2008) in a study with children (age 7), content makes verbatim traces of misinformation more accessible,
using quite different materials. They provided children with nar- increasing false memory and also interfering with retrieval of
ratives about events that had supposedly happened to them at verbatim traces of original targets because original events conflict
school, some true and some misinformation. Some events were with the content of misinformation.
negative, others were neutral, and arousal was not controlled. On
later recognition tests, misinformation effects were larger for neg-
The Story So Far
ative than for neutral events. As in spontaneous false memory,
then, emotion-false memory effects in adults appear to extend to Before moving on to emotional context effects, here is an
children. interim summary of key results. In DRM experiments, negative
Porter, Ten Brinke, Riley, and Baker (2014) manipulated va- valence consistently increased false memory on recognition tests
lence by presenting ambiguous pictures, which were then de- and mostly decreased true memory. With recall, emotional content
scribed to subjects using positively- or negatively-valenced termi- effects were less consistent, but at least it seems that they are
nology. For instance, a picture depicting passengers boarding an different for recall as compared with recognition. That is consistent
airplane was described as headed toward a sunny vacation spot or with the theoretical principle that recall and recognition differ in
a dismal war zone. The valenced descriptions did not control their relative verbatim versus gist reliance and their relative sen-
arousal level. These descriptions were provided at study or at test sitivity to memory for contextual details. Emotional content may
or omitted altogether. When provided at study, negative descrip- be one of a handful of variables that have contrasting effects on
tions increased misinformation effects, relative to no descriptions recognition versus recall. In experiments with pictures and scripted
but not relative to positive descriptions. The opposite was true materials, there was further evidence that negative valence—and
when descriptions were provided at test. Because there was no also positive valence— can increase false memory, but it was less
positive-negative difference, descriptions were collapsed across extensive than DRM data. Finally, data from the misinformation
valence to reveal that emotional descriptions at study caused more paradigm showed that these basic patterns extend from spontane-
major misinformation effects than no descriptions, whereas the ous to implanted false memories because negative valence also
opposite was true when descriptions occurred at test. A novel heightened susceptibility to misinformation and impaired true
feature of these results is that the tendency of negative content to memory.
amplify misinformation effects was not restricted to events that
subjects experienced, but instead, it extended to the emotional Review of Data II. Emotional Context and
content of descriptions of events.
False Memory
Van Damme and Smets (2014) modified Porter, Ten Brinke, et
al.’s (2014) design by manipulating the arousal as well as the There are different types of mood-induction procedures, which
valence of study pictures. They administered both high- and low- are discussed separately in this section, and they have been im-
arousal positive and negative pictures, and they also administered plemented almost exclusively with the DRM illusion. We begin
medium- and low-arousal neutral pictures. Once again, suscepti- with mood experiments that used different induction methods in
bility to misinformation was higher for negative pictures than for conjunction with standard DRM lists, coupled with recognition or
other types, but only for peripheral details and the effect did not recall tests. We conclude with a smaller group of experiments in
depend on arousal level. For central details, differences in valence which mood-induction was combined with emotional DRM lists to
and arousal did not affect susceptibility to misinformation. In a measure mood congruency effects.
no-misinformation control condition, spontaneous false memories
for central details were reduced by negative pictures and by
Induced Mood Experiments
positive arousing pictures. One way to think about the latter results
is as baseline effects of emotional content that are overridden when Music. Music has proved to be an effective method of induc-
misinformation is presented. ing positive and negative moods in a variety of studies (e.g.,
Taken together, misinformation experiments provide a picture Niedenthal & Setterlund, 1994), and hence, it has been used in
of the effects of emotional content that is reminiscent of the earlier false memory studies to create group differences in subjects’
moods before they are exposed to target materials and memory A 0.4
tests. In a pioneering example, Storbeck and Clore (2005) exposed
subjects to either negatively valenced music (Mahler’s Adagietto)
or positively valenced music (Mozart’s Eine Kleine Nacht Musik) 0.3
or no music, followed by study and recall of DRM lists, and a
mood manipulation check. According to the affect-as-information Control
hypothesis, the former should enhance item-specific processing Angry
Recall
0.2
(stronger verbatim traces), relative to a neutral baseline, and the Sad
latter should enhance relational processing (stronger gist traces). Happy
Consistent with that prediction, Storbeck and Clore (2005) Serene

0.1
found that the positive mood group exhibited more false recall than
the negative mood group, who also exhibited lower false recall
than the control group. These data are shown in Figure 7, which
0
plots false memory levels by valence in several mood-induction Corson & Verrier 2007 Van Damme 2013
studies. As there were no mood effects for true recall, it seems to

have affected gist but not verbatim memory. That conclusion was
supported in a follow-up experiment in which a gist-memory B 0.9
condition was added. In that condition, subjects are told to recall
(or recognize) both targets and related distractors (see Brainerd et
0.7
al., 1999), so that they can rely entirely on gist memory. Therefore,
mood effects should be the same as before, if they were gist Control
Recognition
effects. They were. 0.5 Angry
Although these findings seem to show that negative moods Sad
interfere with gist processing, Corson and Verrier (2007) offered 0.3 Happy
an alternative explanation. They noted that arousal was con- Serene
founded with valence in Storbeck and Clore’s (2005) mood ma-
nipulation (the negative musical segment was more arousing than 0.1
the positive one), and they conducted a replication in which
valence and arousal were manipulated factorially in a 2 ⫻ 2 Van Damme 2013
design; that is, different groups were exposed to music that was
Figure 8. False recall and recognition rates in induced mood studies with
A 0.6
arousal controlled. Panel A ⫽ recall. Panel B ⫽ recognition.
0.5
0.4
positive/high-arousal (happy), positive/low-arousal (serene),
negative/high-arousal (angry), or negative/low-arousal (sad). A
Recall
0.3 neutral condition was also included. Following mood induction

0.2
and a mood check, subjects studied and recalled DRM lists, and
finally, responded to a recognition test, too. The core finding was
0.1 simple: There were arousal effects but not valence effects. False
recall and false recognition of CDs were higher in high-arousal
0
Neutral Negative Positive groups than in the other groups (see Figure 8), and neither true
Storbeck & Clore 2005 Storbeck 2013 Knott et al. 2014 recall nor true recognition was affected by mood group. Although
the valence findings are just null results, these data do suggest that
B 0.8
arousal may be a more important component of mood-false mem-
ory effects than valence. However, Corson and Verrier’s results
0.6 have not been consistently replicated, and some experiments have
found mood valence effects with arousal controlled.
Recognition
0.4 Storbeck (2013) reported some experiments that focused on

whether negative moods promote item-specific processing (stron-
0.2 ger verbatim traces) by adding tests of subjects’ memory for
study-phase contextual details. Following mood induction, sub-
jects studied DRM lists, but certain contextual cues (spatial posi-
tions) were varied as lists were presented. Subjects recalled each
Figure 7. False recall and recognition rates in induced mood studies with list after it was presented, which provided the usual measures of
arousal not controlled. Panel A ⫽ false recall in Storbeck and Clore (2005); true and false memory, and finally, they received a source memory
Storbeck (2013), and Knott et al. (2014). Panel B ⫽ false recognition in test, on which the quadrant of the computer screen in which each
Storbeck and Clore (2011) when mood was induced before study. word had been presented was identified. As in Storbeck and
Clore’s (2005) study, false memory was lower in the negative Emery, Hess, and Elliot (2012) used a video mood induction
group than in the positive or neutral groups (see Panel A, Figure prior to DRM list presentation, followed by recall tests, with
7). Simultaneously, the negative group was better at remembering younger and older adults. As in prior studies, false recall was lower
contextual details, which favors the item-specific processing hy- in the negative group than in either the positive or neutral groups.
pothesis. In a second experiment, other contextual details were In a related study, Knott, Threadgold, and Howe (2014) also found
manipulated at study (font), arousal levels were equated for posi- that false recall was lower in a negative mood group than in a
tive and negative groups, and recognition tests were administered positive mood group (see Panel A of Figure 7). Interestingly, the
for true and false memory and for memory for contextual details. negative group also performed more poorly on word problems that
The mood induction used IAPS pictures rather than music, but the required processing the semantic content of DRM lists. Although
results were analogous: False recognition of CDs was lower in the we noted earlier that reported mood effects on false memory,
negative mood group and recognition of contextual details was whether they are attributable to arousal or valence, can be ex-
higher than in the other groups. More complex results were ob- plained if negative moods strengthen verbatim traces, Knott et al.’s
tained in a final experiment in which half the subjects in each results for word problems suggest that negative moods may also
mood group studied DRM lists in the usual way, and half studied impair gist processing.
such lists with accompanying pictures that depicted the objects that Other mood inductions. Yang, Yang, Ceci, and Isen (2015)
the words named (cf. Schacter et al., 1999). Again, false memory used candy to induce positive moods, conjecturing that they would
for CDs was lowest in the negative group, but that effect was only reduce false memory if subjects are explicitly warned that people
present for subjects who studied words without pictures. This falsely remember CDs in the DRM paradigm. Half of the subjects
result, too, is consistent with the notion that negative mood pro- in the positive mood and control groups received such warnings
duces stronger verbatim traces by encouraging item-specific pro- and half did not. As in prior studies, positive mood affected neither
cessing. If that is true, pictures should reduce or eliminate negative true nor false recognition when subjects were not warned, but
mood effects because, as we saw earlier, pictures produce strong when they were, positive moods reduced false recognition, as
verbatim traces, leaving limited room for negative mood to predicted. Thus, despite the lack of positive mood effects up to this
strengthen verbatim traces. point, positive moods may affect false memory with specific types
Van Damme (2013) replicated Corson and Verrier’s (2007) of retrieval cues.
procedure in three experiments and added a second control con- Most recently, Mirandola and Toffalini (2016) used IAPS pic-
dition in which neutral moods were induced prior to the study tures to induce positive, negative, and neutral moods, with arousal
phase. In an initial experiment, similar to Corson and Verrier, false in the positive and negative groups being higher than the neutral
recognition of CDs was elevated for high-arousal groups on a group. True and false memory were measured with a version of
delayed recognition test but did not vary as a function of valence, Mirandola et al.’s scripted-picture paradigm. Unlike most prior
as shown in Panel A of Figure 8. However, when immediate recall studies, mood induction occurred just before the recognition test,
tests were administered in a second experiment, neither valence so that mood affected retrieval but not encoding, and subjects’
nor arousal affected false memory. When immediate recognition rated their moods for valence and arousal following induction.
tests were administered in a final experiment, the results resembled False memory for unpresented causes was lower in the positive
those of the first experiment in that false recognition of CDs was and negative groups, compared with the neutral group, and false
elevated in the high-arousal conditions, regardless of valence, as memories were less often accompanied by erroneous remember
shown in Panel B of Figure 8. judgments. Also, true memory was elevated in the positive and
Overall, the results in the three articles considered up to this negative mood groups.
point show that mood-false memory effects are localized within However, analysis of the mood rating data provides evidence
the arousal component rather than the valence component when that these differences were in fact attributable to arousal rather
they are manipulated factorially, but that there are valence effects, than valence: Subjects’ valence ratings did not correlate with their
too, when arousal is equated for negative versus positive moods. levels of false memory in either the positive or negative group, but
Theoretically, all of these effects can be explained on the ground their arousal ratings correlated negatively with false memory in
that mood chiefly affects the strength of verbatim traces. both conditions. Likewise, true memory correlated positively with
Video. Video clips can also be used to induce mood states, arousal ratings in both groups, but it did not correlate with valence
and Storbeck and Clore (2011) reported an experiment that relied ratings in either group. Overall, subjects’ ratings of their moods
on that procedure. To determine whether mood effects are local- were better predictors of their levels of false memory than which
ized at encoding or retrieval, mood induction occurred before list mood group they were in, and importantly, this was the first study
presentation for half the subjects and after it for the other half. True to provide evidence of pure-retrieval effects of mood on false
and false memory were measured with recognition tests. As in memory.
earlier experiments, false memory was reduced in the negative
mood group when induction occurred before list presentation (see Mood Induction Experiments With Emotional DRM
Panel B of Figure 7), but there were no mood effects when it
Lists: Mood Congruence
occurred later. Thus, mood seems to affect how targets are pro-
cessed during encoding, rather than how traces are retrieved, Thus far, apart from a few experiments, mood effects have been
which is consistent with the affect-as-information hypothesis. It is studied by administering standard DRM lists, followed by recog-
not possible to tell whether this effect is attributable to arousal or nition or recall tests. As noted earlier, Bower’s (1981) mood
valence, however, as arousal was again higher in the negative congruency hypothesis specifies that beyond that, the memory
condition than in the other two conditions. effects of mood states will depend on the level of match between
mood states at encoding versus test. Bower’s general prediction of mood congruency for true memory: The hit rate in the fearful
was that being in, say, a negative mood during the study phase group was higher for fear lists than for neutral lists but not anger
leads to the storage of contextual details that are shared by nega- lists, and the hit rate in the angry group was higher for anger lists
tive events, making it easier to access studied material when those compared with fear lists but not neutral lists. Because arousal was
details are reinstated by negative moods during the test phase. This equated across the two discrete emotions, it appears that at least
is analogous to predictions in classic work on study-test match/ some of these mood congruency effects are valence effects.
mismatch of contextual details, such as Godden and Baddeley’s Summing up, mood-congruency effects were detected in all of
(1975) study of deep-sea divers’ memory for word lists. In that the false memory studies in which they were investigated, and
research, divers studied lists underwater or on land and recalled the there was limited evidence that these effects are stable over time.
lists underwater or on land. Recall was better when the two Although the theoretical and practical significance of these results
environments matched than when they did not. is considerable, they do not support the principles that Bower
In an initial test by Ruci et al. (2009), a narrative mood induc- (1981) used to predict mood congruency. Mood congruency ought
tion (positive, negative, or no induction) was followed by positive, to be more pronounced for true than for false memory because
negative, and neutral DRM lists. Arousal was not controlled in episodic traces containing the contextual details that produce en-
either the mood induction or the DRM lists. Subjects recalled each coding specificity effects are stored for targets, not CDs. Although
list after it was presented, and a final recognition test with remem- they can be retrieved by either, targets are far better retrieval cues
ber/know judgments was administered after all the lists were for those traces and, hence, should yield stronger encoding spec-
recalled. That is, when subjects recognized an item as old, they ificity effects. However, the opposite was true in the above studies,
indicated whether they specifically remembered the item occurring and indeed, mood congruency was observed for targets in only one
earlier or whether they simply knew that it had been presented study. Mood congruency studies also do not resolve the question of
(Tulving, 1985). (Again, as in many other studies, recognition was whether such effects would occur if mood were only induced at
confounded with prior recall.) For false memory, there was evidence test, which is a key question in applied contexts such as forensic
of mood congruency with both recognition and recall: False memory interviewing and sworn testimony.
was elevated for negative mood/negative content relative to negative
mood/positive content and for positive mood/positive content relative
to positive mood/negative content. Also, false remember judgments Interim Summary of Mood-Induction Experiments
(conscious memory of the “presentation” of CDs) was elevated when
Across extant studies, positive moods increased false memory
mood and content matched, which is consistent with other research
and negative moods decreased it. This is the opposite of the
showing that manipulations that increase the retrieval of study-phase
valence effect for emotional content, where negative content ele-
contextual details increase false remember judgments (Arndt, 2012;
vates false memory. Forensically, these opposite trends suggest
Brainerd et al., 2014). For true memory, the mood-congruency effect
that although witnesses’ memories for negative events can be
was limited to the positive mood group and the recall test.
distorted, negative moods may actually reduce such distortion.
Knott and Thorley (2014) investigated whether Ruci et al.’s
According to the theoretical ideas that first motivated mood-false
(2009) results were stable over a forgetting interval. They admin-
memory research, the protective effect of negative moods may be
istered negative and neutral DRM lists followed by recognition
attributable to the fact that they foment item-specific processing
tests, with arousal being higher for negative lists, and again, mood
(enhanced verbatim memory), or to the fact that positive moods
induction preceded the study phase. Ruci et al.’s mood-congruency
foment relational processing (enhanced gist processing), or both.
pattern was not replicated on immediate memory tests, but all
All things considered, however, the most parsimonious interpre-
components of that pattern were present on 1-week delayed tests,
tation of the accumulated data is that differences in the strength of
for true as well as false memory. A new finding was that across
mood conditions, there was a consistent emotional content effect verbatim memory are responsible for mood-group differences in
for recognition: As in the bulk of the content studies that were false memory. Again, note the difference relative to content stud-
reviewed earlier, the false alarm rate for CDs was higher and the ies, which favor the conclusion that differences in the strength of
hit rate for targets was lower for negative lists, and unlike many gist memory are responsible for emotional content effects.
studies, recognition was not confounded with prior recall tests. The respective contributions of valence and arousal to mood-
Bland, Howe, and Knott (2016) extended this line of research to group differences in false memory remain unclear. Valence did not
mood congruency for discrete negative emotions. They paired a have reliable effects when valence and arousal were manipulated
mood induction (fearful, angry, or no induction) with angry, fear- factorially, but on the other hand, it did have reliable effects when
ful, and neutral DRM lists. Moods were induced with videos, and arousal levels were equated for different valence groups. Finally,
the DRM lists were developed using the University of Florida although published data on mood congruency in false memory are
word association database (Nelson, McEvoy, & Schreiber, 2004). not extensive, those that are available are consistent in showing
Arousal was equated for the two types of negative lists, but was that false memory levels for positive and negative moods, includ-
higher for those lists than the neutral lists. There were mood ing discrete negative moods, are higher when the emotional con-
congruency effects because false recognition of CDs in the fearful tent of target materials matches subjects’ moods.
group was higher for fear lists compared with anger and neutral
lists, and false recognition of CDs in the angry group was higher
Mood and Misinformation
for anger lists than neutral lists but not fear lists. As in prior
studies, there was also a mood-congruency effect for false remem- Fiedler and Bless (2001) predicted mood effects for misinfor-
ber judgments. This study also produced the first reliable evidence mation, using a theoretical analysis that resembles the item-
specific versus relational processing distinction. According to their Natural Moods

analysis, negative moods focus processing on actual details of
Cause-effect relations between mood and false memory can
target events, whereas positive moods encourage more abstract,
only be established if subjects’ moods are manipulated experimen-
heuristic processing. Obviously, memories that are stored pursuant
tally, but such research is prone to the criticism that induced mood
to the first type of processing ought to be more resistant to
states are transitory and artificially created. In contrast, mood
misinformation than memories that are stored pursuant to the
states that arise during personally significant events are more
second type.
enduring and natural. Further, arousal and valence levels are more
Forgas, Laham, and Vargas (2005) tested that prediction in
intense with natural mood states, and that, in turn, may yield a
some experiments in which mood induction occurred after the different picture of the relation between mood and false memory.
study phase. Subjects were exposed to positive or negative An obvious way to investigate these possibilities is to measure
pictures (a wedding or a car crash), followed by video mood moods that occur naturally in certain groups of subjects. Below,
induction (positive, negative, or neutral), followed by misinfor- we summarize findings from research in that vein, dividing it into
mation about the pictures, followed by memory tests. Arousal
studies of nonclinical populations versus clinical populations.

was not controlled in either the target materials or the mood
Nonclinical populations. Here, the mood states of normal

induction. Mood influenced the misinformation effect in the adults have been assessed, true and false memories have been
predicted manner: False recognition of misinformation was measured with the DRM procedure, and relations between the two
lower in the negative mood condition than in either the positive sets of measurements have been investigated. For instance, Emery
or neutral mood conditions. There were no emotional content et al. (2012) compared the effects of induced and natural moods in
effects, however, as neither false nor true memory varied as a older and younger adults on a standard DRM task. They measured
function of the valence of the studied pictures, and there was no so-called state mood using the PANAS (Watson et al., 1988) and
evidence of mood congruency in misinformation. In a second so-called trait mood using the NEO Five-Factor Inventory (NEO-
experiment, the study phase, in which subjects viewed a live PI; Costa & McRae, 1990), a measure of extraversion, neuroti-
negative event, was a week before mood induction (positive, cism, agreeableness, conscientiousness, and openness to experi-
negative, neutral), misinformation, and memory testing. The ence. The authors paid particular attention to NEO scores for
results were the same as before. neuroticism and extraversion on the hypothesis that they are asso-
Hess, Popham, Emery, and Elliott (2012) examined the contrast- ciated with negative and positive mood, respectively (Caspi, Rob-
ing notion that positive moods increase misinformation effects by erts, & Shiner, 2005), and are indexes of long-term mood “traits.”
studying younger versus older adults. There is a large literature In contrast, PANAS scores were used as short-term measures of
pointing to the conclusion that the frequency of positive moods, mood “states.” Then, positive and negative moods were induced in
relative to negative ones, increases during healthy aging (e.g., different groups using video clips, and a control group was also
Carstensen, Isaacowitz, & Charles, 1999). Hess et al.’s subjects’ included. Finally, a series of DRM lists was studied and recalled.
moods were measured by the PANAS, after which they studied a Both natural and induced mood effects on false memory were
sequence of positive and negative pictures, followed by misinfor- observed in older adults: False recall was higher for subjects with
mation, followed by a recognition test. As predicted, misinforma- more positive natural moods and for subjects in the positive
tion effects were larger in older than in younger adults. False induction condition. Thus, although subjects’ natural moods cor-
recognition of misinformation was negatively correlated with neg- related with false memory, the relation was qualitatively the same
as that of induced mood.
ative moods, and interestingly, a regression analysis showed that
Koo and Oishi (2009) investigated natural mood-congruency
age differences in misinformation effects were entirely attributable
effects for subjects who were in positive moods, using positive,
to age differences in mood. Consistent with the emotional content
negative, and neutral DRM lists that were not equated on arousal.
studies that we reviewed earlier, misinformation effects were
In their design, a natural mood-congruency effect would have
larger for negative than for positive pictures.
involved higher levels of false and/or true memory for positive
Van Damme and Seynaeve (2013) expanded Forgas et al.’s
DRM lists than for negative or neutral ones. No overall patterns of
(2005) design to include factorial manipulation of valence and that sort were observed, although false memory for one positive
arousal levels over mood groups. Video mood induction occurred CD (happy) was higher than false memory for all neutral CDs.
before the study phase, which was followed by misinformation and Clinical populations. Considerably more research has been
a recognition test in which subjects also rated their confidence in focused on populations that experience enduring negative moods,
each response. As in Van Damme’s (2013) work on mood with the where mood-congruency effects have been investigated by varying
DRM paradigm, there were four mood conditions, in which va- the emotional content of target materials. Multiple studies are
lence and arousal were varied, and two control conditions (neutral available with two populations: Subjects with depression diagno-
mood induction and no mood induction). Neither valence nor ses and subjects with posttraumatic stress disorder (PTSD) diag-
arousal affected false recognition of misinformation, but there noses. We selected those populations for two reasons: First, neg-
were confidence effects. False recognition of misinformation was ative affect is a concomitant of these disorders; depressed
accompanied by higher confidence in the negative mood groups individuals experience depressed moods and feelings such as guilt
than in the positive mood groups, and the same was true of correct and individuals with PTSD experience negative emotions pursuant
recognition of targets. In other words, subjects in negative moods to trauma and persistent negative beliefs. Second, there have been
were more confident in all of their responses than subjects in several studies of false memory with these two disorders, but not
positive moods. with others.
Depression. It has long been understood that depressed indi- and Hua (2009) conducted a similar study using recognition of
viduals display memory impairments (Burt, Zembar, & Niederehe, Chinese emotional DRM lists, and overall, false recognition of
1995), to the point that depression in older adults can be misdiag- CDs was higher in depressed subjects than in controls. Also, false
nosed as cognitive impairment or dementia (e.g., Brainerd, Reyna, recognition was higher for positive and negative CDs than for
& Howe, 2009). It is also well known that depressed individuals neutral CDs in depressed subjects, whereas the opposite was true
exhibit a mood-congruent processing bias in true memory (Bower, in controls. Thus, there were mood-congruency effects in both of
1981); that is, it is easier for depressed individuals to remember these latter studies such that false memory was higher for nega-
negatively valenced information than positively valenced informa- tively valenced versus neutral words.
tion. Consequently, recent research has examined whether that bias Similar to Moritz et al. (2005), Howe and Malone (2011)
extends to false memory. administered depression-related lists—as well as other negative
Watkins, Mathews, Williamson, and Fuller (1992) conducted a lists, positive lists, and neutral lists—followed by recognition tests.
preliminary investigation of this possibility, albeit not with a There was a highly specific mood-congruency effect in which false
standard false memory paradigm. They presented depressed and recognition was elevated in depressed subjects for depression-
control subjects with depression-related, positive, and neutral related CDs but not for other types of CDs. There was a broader
words, followed by two memory tasks. They administered a cued- mood-congruency effect for true memory in which hit rates were
recall test that used stems of studied words (e.g., hop___ for elevated in depressed patients for depression-related and other
hopeless) to measure true recall, and they used a word completion types of negative targets.
test with stems of unstudied as well as studied words to measure Toffalini, Mirandola, Drabik, Melinder, and Cornoldi’s (2014)
implicit memory. The second test generated a false memory-like procedure differed from studies considered so far. They imple-
effect in which more stems were completed for unstudied words mented Hannigan and Reinitz’ (2003) method of administering
that were related to presented words than for other unstudied scripted pictures and assessing false memory for unpresented
words (e.g., ___ault for assault [related] or vault [unrelated]). A causal events and other related events. Toffalini et al. included
mood-congruency effect was observed on the first test: Depressed emotional scripted material (negative vs. neutral), which was ad-
subjects recalled more depression-related words than positive ministered to young adults with depressive-anxious symptoms and
words on the explicit task, but controls did not. However, there to controls. There was a mood-congruency effect for false mem-
was no mood-congruency effect on the implicit test; both groups ory: Negative causal errors (i.e., false memories of unpresented
completed more depression-related word stems than stems of other causes of negative outcomes) were more common in the
types. depressive-anxious group than in the control group, but the inci-
Although Watkins et al. (1992) did not detect mood-congruency dence of neutral causal errors did not differ for the two groups.
effects in false memory, others have. Moritz, Glascher, and Bras- To differentiate the effects of natural negative moods that are
sen (2005) studied depressed subjects and matched controls. They pursuant to depression from natural negative moods that are pur-
administered depression-related, delusion-related, positive, and suant to other conditions, Toffalini, Mirandola, Coli, and Cornoldi
neutral DRM lists (see the lists at the bottom of Table 3), with (2015) repeated the above procedure with a different subject group
arousal not equated, to determine whether the mood-congruent true that experiences predominantly negative moods, highly anxious
memory effect holds for false memory as well. For true recogni- individuals. The scripted target material included positive as well
tion, depressed subjects displayed better memory for all of the as negative and neutral scenarios. Once again, there was a mood-
emotional lists than for the neutral lists, whereas controls’ memory congruency effect in false memory because subjects in the anxious
was best for the positive lists. Depressed subjects displayed a group made more negative causal errors than neutral causal errors
similar effect for false memory: False alarm rates were higher for and made more negative causal errors than controls. Subjects in the
the three types of emotional lists than for the neutral lists, and in anxious group also made more negative than positive noncausal
addition, they were highest for the negative lists (mood congru- errors.
ency). In contrast, controls’ false alarm rates were highest for the An important methodological point about all of the preceding
neutral lists. studies that affects their theoretical interpretation is the absence of
Moritz, Voigt, Arzola, and Otte (2008) reported some follow-up arousal controls. This feature of their design means that, currently,
research that increased the number of DRM lists. They also mea- it is unclear whether mood-congruency effects in depressed sub-
sured a variable that may be a mediator of mood congruency, jects are valence effects, arousal effects, or both.
personal salience, as indexed by the extremity of subjects’ valence PTSD. Individuals with PTSD diagnoses report predomi-
ratings, beyond valence averages. There were no group differences nantly negative moods and increased negative memories—as well
in memory as a function of valence (i.e., no mood-congruency as increased arousal, evidenced by hypervigilance and/or increased
effect) but salience predicted performance: Depressed subjects’ hit startle responses (American Psychiatric Association, 2013).
and false alarm rates were both higher for personally salient words, Such individuals experience a range of negative emotions, includ-
and the same was true for controls. Joormann, Teachman, and ing fear, anger, shame, and alienation, and they may report intru-
Gotlib (2009) then extended Moritz et al.’s (2005) design from sive memories of trauma, plus numbing and dissociation to miti-
recognition to recall of positive, negative, and neutral DRM lists. gate such memories. False memory in PTSD has been a topic of
There was a mood-congruency effect for false memory: Depressed considerable interest for some time, owing to concerns about the
subjects recalled more negative CDs than controls, but false recall accuracy of recovered memories of long-forgotten childhood
of positive and neutral CDs did not differ for the two groups. trauma (e.g., Loftus, 1993). More generally, false memory re-
Concerning true memory, depressed subjects’ recall was poorer search with PTSD populations was originally undertaken to deter-
overall, and this was especially true for positive DRM lists. Yeh mine whether they are at increased risk of erroneous memories of
traumatic events, while at the same time displaying reduced mem- related (CSA) DRM lists were administered, followed by recall
ory for the actual details of traumatic events (e.g., Foa & Riggs, and recognition tests. CSA is of special interest because it is
1993). With respect to mood, because intrusive memories of associated with depression and PTSD (Roosa, Reinholtz, &
trauma are usually accompanied by strong emotions, this aspect of Angelini, 1999; Widom, 1999), and even when its effects do not
PTSD persistently reinstates negative moods. rise to the level of diagnosis, it is associated with memory
In an early article, Bremner, Shobe, and Kihlstrom (2000) impairment (Bremner et al., 2003). PTSD patients who are CSA
measured false recall and false recognition of standard DRM lists victims may experience intrusive memories of abuse and recur-
in subjects with self-reported histories of childhood sexual abuse rent negative moods, and thus, mood-congruent memory effects
and current PTSD diagnoses (i.e., PTSD pursuant to abuse), sub- can be investigated with CSA-related lists. Overall, true recall
jects with histories of abuse but no PTSD diagnosis, and was greatest for CSA-relevant lists and false recall was greatest
nonabused controls. False recognition of CDs but not false recall for CSA-relevant and negative lists, but there was no evidence
was elevated in the PTSD group, relative to the other two groups. of mood congruency in false memory: Although the adolescent
True recall but not true recognition was lower in the PTSD group, PTSD group displayed higher levels of false recognition than all
relative to the other two groups. In short, there was a global other groups, that effect was not specific to CSA-related lists or
suppression of accuracy in PTSD patients. even to negative lists.

Zoellner, Foa, Brigidi, and Przeworski (2000) extended Brem- Expanding Goodman et al.’s design, Baugerud, Howe, Magnus-
ner et al.’s (2000) design by studying the same subject groups but sen, and Melinder (2016) administered negative and neutral DRM
(a) adding remember/know judgments to recognition tests and (b) lists, without controlling arousal, to maltreated and nonmaltreated
administering recall followed by recognition to half the subjects children (7- to 12-year-olds). On subsequent recall tests, true
and arithmetic problems followed by recognition to the other half, memory was lower and false recall of negative CDs was higher for
so that recognition was no longer confounded with prior recall. maltreated children, but false recall of neutral words did not differ
The false memory data were not consistent with a simple mood- between groups. This points to a mood-congruency effect in false
congruency hypothesis: False recall was higher in both abused memory, but one that fades with time: Whereas maltreated chil-
groups than in the control group, false remember judgments were dren exhibited lower true memory and higher false memory for
higher in the non-PTSD abused group than in the other two groups, negative lists, adolescents and adults who experienced maltreat-
and true recognition was also highest in the non-PTSD abused ment as children did not show these effects.
group. Conclusions about natural mood. Across nonclinical, de-
Similar to Bremner et al. (2003), Brennen, Dybdahl, and Kapidzic pressed, and PTSD subjects, valenced natural moods have usually
(2007) found some evidence of mood congruency in false memory been associated with higher levels of false memory than neutral
with PTSD patents. They measured true and false recall— using moods. However, the type of valence that elevates false memory
neutral DRM lists, negative lists, and negative trauma-related lists—in appears to be different for nonclinical versus clinical populations;
traumatized-PTSD subjects and nontraumatized-non-PTSD controls, positive in the former, negative in the latter. Note that this differ-
with PTSD subjects having experienced war trauma. There was no ence has some rather striking implications for forensic analyses of
difference in false memory for neutral lists or negative lists, but witnesses’ susceptibility to false memory because depression and
consistent with mood congruence, false memory for trauma-related PTSD diagnoses are rarely taken into account in forensic work (see
lists was elevated in the PTSD group. Toglia, Ross, Pozzulo, & Pica, 2014). At the level of specific
Jelinek, Hottenrott, Randjbar, Peters, and Moritz (2009) admin- valence, the results for nonclinical natural moods resemble the
istered Moritz et al.’s (2008) picture memory task to subjects who earlier pattern for induced moods.
had experienced trauma from traffic accidents and assault, some of With respect to mood-congruency in false memory, some evi-
whom had PTSD diagnoses and others of whom did not. Subjects dence that favors this hypothesis has been obtained with natural
viewed pictures of common positive, negative, and neutral scenes moods, but the results are different for nonclinical versus clinical
(e.g., beach, funeral) from which some typical details were re- populations. On the one hand, mood-congruency effects tend to be
moved (e.g., beach towels, coffin), followed by recognition tests. specific to negative moods in clinical populations, which is hardly
Contrary to mood congruency, there were no group differences in surprising inasmuch as negative moods are hallmarks of the pop-
false recognition for any of the three valences, and PTSD subjects’ ulations that have been studied. On the other hand, both positive
hit rates were lower for all three valences. and negative mood congruency have been observed in nonclinical
Hauschildt, Peters, Jelinek, and Moritz (2012) reported a populations and in mood induction studies.
follow-up study that focused on false memory for videos. A
PTSD group, a traumatized non-PTSD group, and a control
Discussion and Conclusions
group viewed negative trauma-related, negative-unrelated, pos-
itive, and neutral videos with central details removed, followed How emotion influences false memory has long been a topic of
by recognition tests. Contrary to mood congruence, there were lively speculation in the forensic and clinical literatures. Now, for
no group differences in false memory, and true memory was the first time, the literature that we have surveyed provides exten-
lowest in the PTSD group. sive evidence that under controlled conditions, emotion has sub-
Childhood maltreatment. Goodman et al. (2011) reported a stantial and varied influences on false memory, some of which are
further investigation of trauma-related false memory in adults paradoxical. Below, we summarize the major patterns that have
and adolescents with and without histories of child sexual abuse emerged, highlighting the content-context paradox. We then re-
(CSA) and PTSD. As in some prior studies of depressed and consider the three questions posed at the beginning of this review,
PTSD patients, neutral, positive, negative, and negative trauma- about the direction, quality, and locus of emotion-false memory
effects. Finally, we discuss practical implications and some obvi- Direction of Effects
ous lines of research that should answer key open questions.
Our first question was whether false memory varies uniformly
with changes in emotion. The answer is no, and instead, the
Major Patterns direction of valence effects varies as a function of design details.
In the first part of this survey, we focused on one of two The direction of arousal effects is somewhat more uniform, but
implementations of emotion—namely, content. The dominant that is subject to the proviso that it is has not been extensively
methodology was the DRM paradigm, owing to its efficiency in studied by itself. With that qualification in mind, there is evidence
creating false memories and the ease with which valence and from both content and context studies that increasing arousal
arousal can be manipulated. We saw that whether arousal is elevates false memory.
confounded with valence, controlled at fixed levels, or manipu- Recall versus recognition tests was the first example of differ-
lated factorially with valence influences DRM false memory, as ences in the direction of valence effects that we encountered. In
does whether memory is tested with recall or recognition. Despite DRM studies, negative content has been found to reduce false
such variability, a consistent result was that negative valence memory on recall tests but to increase it on recognition tests. That
elevates false recognition of CDs. difference could be attributable to known differences in the rela-
Beyond the DRM illusion, we saw that encoding modality can tive sensitivity of recognition and recall to gist and verbatim
influence results with other paradigms, such as pictures and memory, respectively. If valence affects the strength of the gist
scripted materials. There are fewer experiments with other proce- traces that support false memories, those effects would be more
dures, but they also show that negative content increases false pronounced on recognition than on recall tests. If valence also
memory. Although there is a fundamental difference between false affects the strength of verbatim traces that suppress false memo-
memories that occur spontaneously because they fit the gist of ries, those effects would be more pronounced on recall tests. That
experience versus those that are implanted via misinformation, the hypothesis remains tentative because recognition has so often been
general finding of false memory being elevated by negative con- confounded with prior recall in the studies that we reviewed.
tent holds for both. In misinformation studies, subjects were usu- A second example of variability in the direction of emotion-
ally more likely to falsely remember misinformation when it was false memory effects was that the direction of mood effects can
negative. depend on whether moods are induced or natural: Negative in-
In the second part of our survey, we focused on the other duced moods suppress false memory, whereas recurrent negative
implementation of emotion—namely, context, as supplied by sub- natural moods, as in depression or PTSD, are positively correlated
jects’ moods. We began with research in which moods were with false memory. It should be added, however, that these con-
experimentally induced, chiefly by music, and then considered trasting effects are for mood states that differ in extremity and
other methodologies. The dominant patterns were that positive duration. Induced moods are weaker and more transient than
moods elevated false memory relative to negative moods, and persistent natural moods, and further, natural moods are products
aroused moods elevated false memory compared with nonaroused of real-life events of personal significance.
moods. These patterns were fairly consistent across mood meth- How could these differences affect the direction of valence
odologies. Obviously, the valence effect is in stark contrast to the effects in false memory? On the natural mood side, research on
chief result of content-false memory studies. Further, when mood true memory has shown that enduring negative moods promote a
induction was paired with emotional DRM lists, false memory was processing style that favors reliance on gist traces, resulting in
only elevated for negative lists when moods were also negative. As elevated levels of false memory for subjects with depression or
in content studies, mood valence effects were similar for sponta- PTSD diagnoses. On the induced mood side, however, negative
neous and implanted false memories. moods promote item-specific processing, which favors reliance on
Turning to natural mood studies, when clinical populations were verbatim traces and suppresses false memories. Positive induced
studied the results were different than when moods were induced. moods, as predicted by the affect-as-information hypothesis, pro-
With depressed and PTSD populations (persistent negative mote relational processing, which favors reliance on gist traces and
moods), subjects were more prone to false memory than control elevates false memory.
subjects, whereas negative induced moods reduced false memory. Differences between clinical and induced mood effects could
Mood-congruency effects were detected for these populations as also be a consequence of differences in levels of valence and
well because memory was elevated for negative DRM lists that arousal, which are much higher for the former. This may also
were specific to their condition. explain why mood-congruency effects have more often been de-
In the end, our survey identified three major ways in which tected with natural moods. In that connection, subjects who are in
emotion-false memory effects vary as a consequence of major strong negative natural moods may be more prone to connect the
design features. First, negative content increased false memories gist across different events that are congruent with their moods.
on recognition tests but not on recall tests. This is an intriguing Under the usual opponent-process account of false memory, that
datum theoretically because content manipulations that elevate would result in more mood-congruent false memories, as com-
false memory in recognition normally elevate it in recall (Brainerd, pared with subjects who experience induced moods.
Reyna, et al., 2008). Second, negative content increased false Another potential source of clinical versus induced mood dif-
memories but negative induced moods reduced false memories. ferences is the many confounding factors that are present in
Third, although negative induced moods reduced false memories, clinical populations (e.g., individual differences in memory). Such
negative natural moods elevated false memories. Below, we sketch factors are eliminated via random assignment in induced mood
theoretical explanations of these patterns. studies, but with the clinical populations that figure in the present
review, memory impairment is itself a symptom of depression and sequences of goals. This is important when it comes to false
PTSD. Thus, associations between depression, PTSD, and false memory because it is known that memory distortion varies as a
memory could be consequences of uncontrolled, extramood vari- function of such differences in attentional focus (Van Damme,
ables in clinical populations. Kaplan, Levine, & Loftus, 2016). In addition, when specific emo-
Developmental variability provides a third example of princi- tions differ in valence or arousal, opponent-process theories such
pled differences in the direction of emotion-false memory effects. as FTT make predictions about differences in false memory as a
From what is known, which is confined to content studies, function of differences in specific emotions (Bookbinder & Brain-
emotion-false memory effects are qualitatively similar in children erd, 2016).
and adults, but the magnitude of those effects is smaller in chil- Arousal-driven differences in false memory may ultimately be
dren. To many, this finding has seemed counterintuitive because it attentional in origin, and more generally, attention has been men-
suggests that children are less susceptible to false memory than tioned as a likely factor in some influential theoretical proposals
adults, when conventional wisdom says the opposite. However, about emotional false memory. For instance, some have argued
this finding is quite consistent with the large literature on devel- that the item versus relational processing distinction is about
opmental reversals in semantic false memories. There is far less narrowing versus broadening attention, and that it is the broader
research on mood effects in children, far too little to hazard attentional focus of positive moods that elevates false memory
conclusions about developmental variability in those effects. (e.g., Derryberry & Tucker, 1994). Similarly, others have inter-
preted the affect-as-information hypothesis as implicating differ-
Quality of Emotion ences in attentional focus via differences in approach versus avoid-
ance (e.g., Förster, Friedman, Özelsel, & Denzler, 2006). The basic
The second question was whether emotion-false memory effects proposal is that positive moods motivate approach, which broadens
are attributable to valence, arousal, or both. In content studies, we attention, whereas negative moods motivate avoidance, which
saw that false memory varies for both positive versus negative narrows attention. Beyond predicting that false memory will vary
valence and high- versus low-arousal. When arousal was equated, as a function of attentional breadth, this account also incorporates
negative DRM lists produced more false memory than positive evolutionary ideas, such as the role of fear in avoiding danger:
ones, and so did negative thematic pictures. Turning to arousal, There is a survival advantage to focusing attention when experi-
increasing it seems to elevate false memory, at least for negative encing fear, explaining why attention increases and false memory
DRM lists. The first result is likely attributable to the increased decreases with negative mood states.
salience of semantic connections among negative targets as com-
pared with positive ones, which strengthens gist memory. The
arousal effect is more difficult to interpret because in most cases, Locus of Emotion
arousal was confounded with valence, but it follows from FTT’s
proposal that increasing arousal reduces the strength of verbatim The third question asked whether emotion-false memory effects
traces by blurring attentional focus during encoding. depend on whether content or context is the locus of emotion. The
Valence and arousal also appear to affect false memory as answer, if we have learned anything, is yes. The picture that is
components of mood, but there has been substantially less control provided by the overall arc of data is that negative valence elevates
of arousal in that literature. In mood research in which valence and and positive valence reduces false memory when the locus
arousal were manipulated factorially, there were arousal effects but is content, but positive valence elevates and negative valence
not valence effects. Further, although negative moods reduced reduces false memory when the locus is mood. Just as recall versus
false memory relative to positive moods, it was only when arousal recognition seems to switch the direction of emotional content
was not controlled. With respect to opponent-process distinctions, effects, then, locus seems to switch the direction of effects. For-
all of this suggests that mood chiefly affects verbatim memory, and tunately, the principles of opponent-process theories, which are
consistent with affect-as-information, it does so primarily at en- grounded in the larger false memory literature, are able to explain
coding. Notwithstanding that the effects of mood valence are this dependency. At the level of memory representation, the most
uncertain because it has so often been confounded with arousal, it likely explanation is that negative content and positive moods both
does appear that those effects depend on arousal because they enhance semantic processing, strengthening gist traces relative to
seem to decrease as arousal increases. verbatim traces, but positive content and negative moods both
Most mood studies in which valence and arousal were factori- enhance the processing of targets per se, strengthening verbatim
ally manipulated did so by varying specific emotions (e.g., positive traces relative to gist traces.
valence/low arousal ⫽ serene; Van Damme, 2013). Those studies Beyond this basic account, a full explanation of the context-
produced inconsistent results, but as discussed earlier, that may content paradox may need to include differences in the relative
simply be because each Valence ⫻ Arousal combination can be contributions of valence and arousal when they are manipulated in
represented by more than one specific emotion. In that connection, content versus context. There is some slim evidence from factorial
appraisal theories (Scherer, 1999) claim that specific emotions can manipulations of the two that valence effects are more pronounced
be pre- or postgoal oriented, meaning that they focus on future goal than arousal effects in the content sphere, whereas the reverse is
attainment or failure (e.g., desire, fear) or they are reactions to past true in the context sphere. If that pattern continues to hold, then
goal attainment or failure (e.g., happiness, sadness). That has valence preferentially drives processing when it is embedded in the
straightforward implications for attention: Pregoal emotions nar- target material itself, but arousal preferentially drives processing
row attention to information that is central to the goal itself, when it figures in the context in which that material is presented.
whereas postgoal emotions broaden attention to include the con- Theoretically, that difference could be explained by relying on the
distinction between target recollection and context recollection in in content studies. Such comparisons are further hampered by the
episodic memory (see Brainerd et al., 2014). fact that much mood research conceptualizes emotion as discrete
emotional states, whereas the content literature relies primarily on
the notion of continuous valence and arousal dimensions. Obvi-
Practical Importance
ously, the two literatures’ theoretical conceptions will need to be
Returning to the forensic situations in which emotion-false aligned to effect direct comparisons of valence and arousal effects
memory effects are of great concern, pragmatic recommendations in content versus context.
derived from the literature we have reviewed are complex because Another obvious target for future research, one that is probably
emotional content and context are normally confounded in legal easier to accomplish, is to disentangle emotion-false memory
cases. That is because (a) the specific events during crimes that the effects for recall and recognition by testing theoretical hypotheses
law is concerned with are inherently emotional (negative, high- about the reasons why those effects are test-dependent. As we saw,
arousal); (b) the general context of a crime is apt to generate extant data are inadequate because recognition tests have been
negative-arousing moods in witnesses; (c) witnesses may initially routinely confounded with prior recall, so that extant recognition
be in quite different moods (positive, low-arousal); (d) some data are recall-dependent in unknown ways. Thus, it is not entirely
events that occur during the postcrime investigation phase (e.g., clear whether the different emotion-false memory effects for rec-
police interviews and interrogations) and the trial phase (e.g., ognition versus recall are real or are attributable to this confound.
direct testimony and cross examination) are inherently emotional; Fortunately, mathematical models exist for both recognition and
(e) those postcrime events are apt to induce specific moods states; recall that can pinpoint the precise retrieval processes that are
and (f) research suggests that memory distortion pursuant to a and responsible for a manipulation’s effects with each type of test.
d will be different than memory distortion pursuant to b, c, and e. Thus, experiments that combine those models with counterbal-
However that may be, remember that the many design variations in anced testing procedures would not only isolate different emo-
the literature we have reviewed supply differential evidence on the tional effects for recognition versus recall, but they would deliver
influence of all of these factors, which is what is needed for process explanations of those differences as a by-product.
forensic analysis of the reliability of witnesses’ memories. For A third direction for future research is to expand research with
instance, studies in which mood was induced before versus after clinical populations, particularly as a way of studying the effects of
encoding target material show that b and c will usually be more more intense levels of emotion. With respect to PTSD especially,
important considerations than e. The main conclusion for the available data do not provide a consistent picture of false memory
present is that because the literature shows that these factors have effects, especially for traumatic information (mood-congruent
different effects, a forensic analysis that relies on just one is flawed false memory). Obviously, it is important to determine whether
and misleading. PTSD individuals are at elevated risk for false memories and
It should be added that beyond memory distortion in the law, whether such elevation is mood-congruent. As part of such work,
there are other high-stakes remembering situations to which the current explanations of established PTSD impairments in true
same considerations apply, point for point. Patients’ reports of memory should be put to the test, such as reduced autobiographical
medical symptoms and histories are prime examples in which the memory (Brewin, Kleiner, Vasterling, & Field, 2007). As for
parallels are so self-evident as to require no discussion. Interest- depression, the research goals, both empirical and theoretical,
ingly, as mentioned earlier, this is another domain that features the would be much the same.
same conflicting hypotheses about whether emotional variations
are distortive or protective that are found in the law. What we have Concluding Comment
learned in this review is that conflicting hypotheses are tenable
because emotional variations can have both effects, depending on Emotion-false memory has been widely studied, and that work
design features. has produced some instructive findings. One of them is that those
effects vary as a function of whether emotion is manipulated in
target material or in subjects’ moods. Paradoxically, the two im-
Future Directions plementations produce largely opposite effects. Although this pat-
The most obvious near-term goal for research, which we have tern is surprising, contemporary opponent-process theories can
mentioned at several points, is more consistent control of arousal explain it. According to this explanation, the locus of emotional
to identify the nature of valence effects and arousal effects inde- content effects lies primarily with gist memory, whereas the locus
pendently of each other. Notwithstanding the progress in that of emotional context effects, at least as they are exemplified in
direction in DRM content studies, comparable studies with other subjects’ moods, lies primarily with verbatim memory.
false memory paradigms will be necessary to ensure that the
effects that have been obtained are not DRM-specific. This is far References
from idle speculation, considering that the DRM illusion is a very
special paradigm. In the same vein, there has been disappointingly Alba, J. W., & Hasher, L. (1983). Is memory schematic? Psychological
Bulletin, 93, 203–231. http://dx.doi.org/10.1037/0033-2909.93.2.203
little control of arousal in mood studies, in both experimental and
American Psychiatric Association. (2013). Diagnostic and statistical man-
correlational designs. Arousal is rarely measured in natural mood ual of mental disorders (5th ed.). Arlington, VA: American Psychiatric
studies, and in most mood-induction procedures (e.g., music and Publishing.
guided imagery), it is not varied systematically. Better separation Anderson, A. K., Christoff, K., Panitz, D., De Rosa, E., & Gabrieli, J. D.
of arousal from valence in mood studies will be essential if there (2003). Neural correlates of the automatic processing of threat facial
are to be systematic comparisons to emotion-false memory effects signals. The Journal of Neuroscience, 23, 5627–5633.
Anderson, A. K., & Phelps, E. A. (2001). Lesions of the human amygdala Brainerd, C. J., & Reyna, V. F. (2002). Fuzzy-trace theory: Dual processes
impair enhanced perception of emotionally salient events. Nature, 411, in memory, reasoning, and cognitive neuroscience. Advances in Child
305–309. http://dx.doi.org/10.1038/35077083 Development and Behavior, 28, 41–100. http://dx.doi.org/10.1016/
Appelbaum, P. S., Uyehara, L. A., & Elin, M. R. (Eds.). (1997). Trauma S0065-2407(02)80062-3
and memory: Clinical and legal controversies. New York, NY: Oxford Brainerd, C. J., & Reyna, V. F. (Eds.). (2005). The science of false memory.
University Press. New York, NY: Oxford University Press. http://dx.doi.org/10.1093/
Arndt, J. (2006). Distinctive information and false recognition: The con- acprof:oso/9780195154054.001.0001
tribution of encoding and retrieval factors. Journal of Memory and Brainerd, C. J., & Reyna, V. F. (2007). Explaining developmental reversals
Language, 54, 113–130. http://dx.doi.org/10.1016/j.jml.2005.08.003 in false memory. Psychological Science, 18, 442– 448. http://dx.doi.org/
Arndt, J. (2012). False recollection: Empirical findings and their theoretical 10.1111/j.1467-9280.2007.01919.x
implications. Psychology of Learning and Motivation, 56, 81–124. Brainerd, C. J., Reyna, V. F., & Brandse, E. (1995). Are children’s false
http://dx.doi.org/10.1016/B978-0-12-394393-4.00003-0 memories more persistent than their true memories? Psychological Sci-
Arora, Y. (Photographer). (2011, July 7). Electric iron [digital image]. ence, 6, 359 –364. http://dx.doi.org/10.1111/j.1467-9280.1995
Retrieved from http://www.freeimages.com/photo/electric-iron-1419801 .tb00526.x
Bartlett, F. C. (1932). Remembering: A study in experimental and social Brainerd, C. J., Reyna, V. F., & Ceci, S. J. (2008). Developmental reversals
psychology. Cambridge, England: Cambridge University Press. in false memory: A review of data and theory. Psychological Bulletin,
Baugerud, G. A., Howe, M. L., Magnussen, S., & Melinder, A. (2016). 134, 343–382. http://dx.doi.org/10.1037/0033-2909.134.3.343
Maltreated and non-maltreated children’s true and false memories of Brainerd, C. J., Reyna, V. F., & Forrest, T. J. (2002). Are young children
neutral and emotional word lists in the Deese/Roediger-McDermott task. susceptible to the false-memory illusion? Child Development, 73, 1363–
Journal of Experimental Child Psychology, 143, 102–110. http://dx.doi 1377. http://dx.doi.org/10.1111/1467-8624.00477
.org/10.1016/j.jecp.2015.10.007 Brainerd, C. J., Reyna, V. F., Gomes, C. F. A., Kenney, A. E., Gross, C. J.,
Beauregard, M., Lévesque, J., & Bourgouin, P. (2001). Neural correlates of Taub, E. S., . . . the Alzheimer’s Disease Neuroimaging Initiative.
conscious self-regulation of emotion. The Journal of Neuroscience, 21, (2014). Dual-retrieval models and neurocognitive impairment. Journal
RC165. of Experimental Psychology: Learning, Memory, and Cognition, 40,
Bjorklund, D. F. (Ed.), (2000). False memory creation in children and 41– 65. http://dx.doi.org/10.1037/a0034057
adults (pp. 93–128). Mahwah, NJ: Erlbaum. Brainerd, C. J., Reyna, V. F., & Howe, M. L. (2009). Trichotomous
Bjorklund, D. F., Bjorklund, B. R., Brown, R. D., & Cassel, W. S. (1998). processes in early memory development, aging, and neurocognitive
Children’s susceptibility to repeated questions: How misinformation impairment: A unified theory. Psychological Review, 116, 783– 832.
changes children’s answers and their minds. Applied Developmental http://dx.doi.org/10.1037/a0016963
Science, 2, 99 –111. http://dx.doi.org/10.1207/s1532480xads0202_4 Brainerd, C. J., Reyna, V. F., & Kneer, R. (1995). False-recognition
Bjorklund, D. F., Cassel, W. S., Bjorklund, B. R., Brown, R. D., Park, reversal: When similarity is distinctive. Journal of Memory and Lan-
C. L., Ernst, K., & Owen, F. A. (2000). Social demand characteristics in guage, 34, 157–185. http://dx.doi.org/10.1006/jmla.1995.1008
children’s and adults’ eyewitness memory and suggestibility: The effect Brainerd, C. J., Reyna, V. F., & Mojardin, A. H. (1999). Conjoint recog-
of different interviewers on free recall and recognition. Applied Cogni- nition. Psychological Review, 106, 160 –179. http://dx.doi.org/10.1037/
tive Psychology, 14, 421– 433. http://dx.doi.org/10.1002/1099- 0033-295X.106.1.160
0720(200009)14:5⬍421::AID-ACP659⬎3.0.CO;2-4 Brainerd, C. J., Stein, L. M., & Reyna, V. F. (1998). On the development
Bland, C. E., Howe, M. L., & Knott, L. (2016). Discrete emotion-congruent of conscious and unconscious memory. Developmental Psychology, 34,
false memories in the DRM paradigm. Emotion, 16, 611– 619. http://dx 342–357. http://dx.doi.org/10.1037/0012-1649.34.2.342
.doi.org/10.1037/emo0000153 Brainerd, C. J., Stein, L. M., Silveira, R. A., Rohenkohl, G., & Reyna, V. F.
Bless, H., Schwarz, N., Clore, G. L., Golisano, V., Rabe, C., & Wölk, M. (2008). How does negative emotion cause false memories? Psycholog-
(1996). Mood and the use of scripts: Does a happy mood really lead to ical Science, 19, 919 –925. http://dx.doi.org/10.1111/j.1467-9280.2008
mindlessness? Journal of Personality and Social Psychology, 71, 665– .02177.x
679. http://dx.doi.org/10.1037/0022-3514.71.4.665 Brainerd, C. J., Yang, Y., Reyna, V. F., Howe, M. L., & Mills, B. A.
Bookbinder, S. H., & Brainerd, C. J. (2016). Emotionally negative pictures (2008). Semantic processing in “associative” false memory. Psycho-
enhance gist memory. Emotion. Advance online publication. http://dx nomic Bulletin & Review, 15, 1035–1053. http://dx.doi.org/10.3758/
.doi.org/10.1037/emo0000171 PBR.15.6.1035
Bower, G. H. (1981). Mood and memory. American Psychologist, 36, Bransford, J. D., & Franks, J. J. (1971). The abstraction of linguistic ideas.
129 –148. http://dx.doi.org/10.1037/0003-066X.36.2.129 Cognitive Psychology, 2, 331–350. http://dx.doi.org/10.1016/0010-
Bradley, M. M., & Lang, P. J. (1994). Measuring emotion: The Self- 0285(71)90019-3
Assessment Manikin and the Semantic Differential. Journal of Behavior Bremner, J. D., Shobe, K. K., & Kihlstrom, J. F. (2000). False memories
Therapy and Experimental Psychiatry, 25, 49 –59. http://dx.doi.org/10 in women with self-reported childhood sexual abuse: An empirical
.1016/0005-7916(94)90063-9 study. Psychological Science, 11, 333–337. http://dx.doi.org/10.1111/
Bradley, M. M., & Lang, P. J. (1999). Affective norms for English words 1467-9280.00266
(ANEW): Stimuli, instruction manual and affective ratings (Tech. Rep. Bremner, J. D., Vythilingam, M., Vermetten, E., Southwick, S. M.,
No. C-1). Gainesville, FL: The Center for Research in Psychophysiol- McGlashan, T., Staib, L. H., . . . Charney, D. S. (2003). Neural correlates
ogy, University of Florida. of declarative memory for emotionally valenced words in women with
Brainerd, C. J., Gomes, C. F. A., & Moran, R. (2014). The two recollec- posttraumatic stress disorder related to early childhood sexual abuse.
tions. Psychological Review, 121, 563–599. http://dx.doi.org/10.1037/ Biological Psychiatry, 53, 879 – 889. http://dx.doi.org/10.1016/S0006-
a0037668 3223(02)01891-7
Brainerd, C. J., Holliday, R. E., Reyna, V. F., Yang, Y., & Toglia, M. P. Brennen, T., Dybdahl, R., & Kapidzić, A. (2007). Trauma-related and
(2010). Developmental reversals in false memory: Effects of emotional neutral false memories in war-induced Posttraumatic Stress Disorder.
valence and arousal. Journal of Experimental Child Psychology, 107, Consciousness and Cognition, 16, 877– 885. http://dx.doi.org/10.1016/j
137–154. http://dx.doi.org/10.1016/j.jecp.2010.04.013 .concog.2006.06.012
Brewin, C. R., Kleiner, J. S., Vasterling, J. J., & Field, A. P. (2007). Diener, E., Emmons, R. A., Larsen, R. J., & Griffin, S. (1985). The
Memory for emotionally neutral information in posttraumatic stress satisfaction with life scale. Journal of Personality Assessment, 49, 71–
disorder: A meta-analytic investigation. Journal of Abnormal Psychol- 75. http://dx.doi.org/10.1207/s15327752jpa4901_13
ogy, 116, 448 – 463. http://dx.doi.org/10.1037/0021-843X.116.3.448 Doerksen, S., & Shimamura, A. P. (2001). Source memory enhancement
Brown, R., & Kulik, J. (1977). Flashbulb memories. Cognition, 5, 73–99. for emotional words. Emotion, 1, 5–11. http://dx.doi.org/10.1037/1528-
http://dx.doi.org/10.1016/0010-0277(77)90018-X 3542.1.1.5
Budson, A. E., Todman, R. W., Chong, H., Adams, E. H., Kensinger, E. A., Doris, J. (Ed.), (1991). The suggestibility of children’s recollections. Wash-
Krangel, T. S., & Wright, C. I. (2006). False recognition of emotional ington, DC: American Psychological Association. http://dx.doi.org/10
word lists in aging and Alzheimer disease. Cognitive and Behavioral .1037/10097-000
Neurology, 19, 71–78. http://dx.doi.org/10.1097/01.wnn.0000213905 D’Souza-Vazirani, D., Minkovitz, C. S., & Strobino, D. M. (2005). Va-
.49525.d0 lidity of maternal report of acute health care use for children younger
Burt, D. B., Zembar, M. J., & Niederehe, G. (1995). Depression and than 3 years. Archives of Pediatrics & Adolescent Medicine, 159, 167–
memory impairment: A meta-analysis of the association, its pattern, and 172.
specificity. Psychological Bulletin, 117, 285–305. http://dx.doi.org/10 Easterbrook, J. A. (1959). The effect of emotion on cue utilization and the
.1037/0033-2909.117.2.285 organization of behavior. Psychological Review, 66, 183–201. http://dx

Carstensen, L. L., Isaacowitz, D. M., & Charles, S. T. (1999). Taking time .doi.org/10.1037/h0047707
seriously. A theory of socioemotional selectivity. American Psycholo- Ekman, P. (1992). An argument for basic emotions. Cognition and Emo-
gist, 54, 165–181. http://dx.doi.org/10.1037/0003-066X.54.3.165 tion, 6, 169 –200. http://dx.doi.org/10.1080/02699939208411068
Caspi, A., Roberts, B. W., & Shiner, R. L. (2005). Personality develop- Ekman, P., & Friesen, W. V. (1971). Constants across cultures in the face
ment: Stability and change. Annual Review of Psychology, 56, 453– 484. and emotion. Journal of Personality and Social Psychology, 17, 124 –
http://dx.doi.org/10.1146/annurev.psych.55.090902.141913 129. http://dx.doi.org/10.1037/h0030377
Ceci, S. J., & Bruck, M. (1993). Suggestibility of the child witness: A Ekman, P., Levenson, R. W., & Friesen, W. V. (1983). Autonomic nervous
historical review and synthesis. Psychological Bulletin, 113, 403– 439. system activity distinguishes among emotions. Science, 221, 1208 –
http://dx.doi.org/10.1037/0033-2909.113.3.403 1210. http://dx.doi.org/10.1126/science.6612338
Choi, H. Y., Kensinger, E. A., & Rajaram, S. (2013). Emotional content El Sharkawy, J., Groth, K., Vetter, C., Beraldi, A., & Fast, K. (2008). False
enhances true but not false memory for categorized stimuli. Memory & memories of emotional and neutral words. Behavioural Neurology, 19,
Cognition, 41, 403– 415. http://dx.doi.org/10.3758/s13421-012-0269-2 7–11. http://dx.doi.org/10.1155/2008/587239
Christianson, S. Å., & Hübinette, B. (1993). Hands up! A study of wit- Emery, L., Hess, T. M., & Elliot, T. (2012). The illusion of the positive:
nesses’ emotional reactions and memories associated with bank robber- The impact of natural and induced mood on older adults’ false recall.
ies. Applied Cognitive Psychology, 7, 365–379. http://dx.doi.org/10 Aging, Neuropsychology, and Cognition, 19, 677– 698. http://dx.doi.org/
.1002/acp.2350070502 10.1080/13825585.2011.645012
Christianson, S. Å., & Loftus, E. F. (1990). Some characteristics of Feldman Barrett, L. R., & Russell, J. A. (1998). Independence and bipo-
people’s traumatic memories. Bulletin of the Psychonomic Society, 28, larity in the structure of current affect. Journal of Personality and Social
195–198. http://dx.doi.org/10.3758/BF03334001 Psychology, 74, 967–984. http://dx.doi.org/10.1037/0022-3514.74.4.967
Colibazzi, T., Posner, J., Wang, Z., Gorman, D., Gerber, A., Yu, S., . . . Fiedler, K., & Bless, H. (2001). The formation of beliefs in the interface of
Peterson, B. S. (2010). Neural systems subserving valence and arousal affective and cognitive processes. In N. Frijda, A. Manstead, & S. Bem
during the experience of induced emotions. Emotion, 10, 377–389. (Eds.), Emotions and beliefs: How feelings influence thoughts. New
http://dx.doi.org/10.1037/a0018484 York, NY: Cambridge University Press.
Corson, Y., & Verrier, N. (2007). Emotions and false memories: Valence Foa, E. B., & Riggs, D. S. (1993). Post-traumatic stress disorder in rape
or arousal? Psychological Science, 18, 208 –211. http://dx.doi.org/10 victims. In J. Oldham, M. B. Riba, & A. Tasman (Eds.), Annual review
.1111/j.1467-9280.2007.01874.x of psychiatry (Vol. 12, pp. 273–303). Washington, DC: American Psy-
Costa, P. T., Jr., & McCrae, R. R. (1990). Personality disorders and the chiatric Association.
five-factor model of personality. Journal of Personality Disorders, 4, Forgas, J. P., Laham, S. M., & Vargas, P. T. (2005). Mood effects on
362–371. http://dx.doi.org/10.1521/pedi.1990.4.4.362 eyewitness memory: Affective influences on susceptibility to misinfor-
Crawford, J. R., & Henry, J. D. (2004). The positive and negative affect mation. Journal of Experimental Social Psychology, 41, 574 –588. http://
schedule (PANAS): Construct validity, measurement properties and dx.doi.org/10.1016/j.jesp.2004.11.005
normative data in a large non-clinical sample. British Journal of Clinical Förster, J., Friedman, R. S., Özelsel, A., & Denzler, M. (2006). Enactment
Psychology, 43, 245–265. http://dx.doi.org/10.1348/0144665031752934 of approach and avoidance behavior influences the scope of perceptual
Dan-Glauser, E. S., & Scherer, K. R. (2011). The Geneva affective picture and conceptual attention. Journal of Experimental Social Psychology,
database (GAPED): A new 730-picture database focusing on valence 42, 133–146. http://dx.doi.org/10.1016/j.jesp.2005.02.004
and normative significance. Behavior Research Methods, 43, 468 – 477. Gallo, D. A., Foster, K. T., & Johnson, E. L. (2009). Elevated false
http://dx.doi.org/10.3758/s13428-011-0064-1 recollection of emotional pictures in young and older adults. Psychology
Deese, J. (1959). On the prediction of occurrence of particular verbal and Aging, 24, 981–988. http://dx.doi.org/10.1037/a0017545
intrusions in immediate recall. Journal of Experimental Psychology, 58, Gallo, D. A., & Roediger, H. L., III. (2002). Variability among word lists
17–22. http://dx.doi.org/10.1037/h0046671 in eliciting memory illusions: Evidence for associative activation and
Dehon, H., Larøi, F., & Van der Linden, M. (2010). Affective valence monitoring. Journal of Memory and Language, 47, 469 – 497. http://dx
influences participant’s susceptibility to false memories and illusory .doi.org/10.1016/S0749-596X(02)00013-X
recollection. Emotion, 10, 627– 639. http://dx.doi.org/10.1037/a0019595 Garry, M., Manning, C. G., Loftus, E. F., & Sherman, S. J. (1996).
Derryberry, D., & Tucker, D. M. (1994). Motivating the focus of attention. Imagination inflation: Imagining a childhood event inflates confidence
In P. M. Neidenthal & S. Kitayama (Eds.), The heart’s eye: Emotional that it occurred. Psychonomic Bulletin & Review, 3, 208 –214. http://dx
influences in perception and attention (pp. 167–196). Cambridge, UK: .doi.org/10.3758/BF03212420
Academic Press. http://dx.doi.org/10.1016/B978-0-12-410560-7 Githens, P. B., Glass, C. A., Sloan, F. A., & Entman, S. S. (1993). Maternal
.50014-4 recall and medical records: An examination of events during pregnancy,
childbirth, and early infancy. Birth, 20, 136 –141. http://dx.doi.org/10 Hunt, R. R., & Einstein, G. O. (1981). Relational and item-specific infor-
.1111/j.1523-536X.1993.tb00438.x mation in memory. Journal of Verbal Learning & Verbal Behavior, 20,
Godden, D. R., & Baddeley, A. D. (1975). Context-dependent memory in 497–514. http://dx.doi.org/10.1016/S0022-5371(81)90138-9
two natural environments: On land and underwater. British Journal of Hyman, I. E., Jr., & Pentland, J. (1996). The role of mental imagery in the
Psychology, 66, 325–331. http://dx.doi.org/10.1111/j.2044-8295.1975 creation of false childhood memories. Journal of Memory and Lan-
.tb01468.x guage, 35, 101–117. http://dx.doi.org/10.1006/jmla.1996.0006
Gomes, C. F., Brainerd, C. J., & Stein, L. M. (2013). Effects of emotional Inbau, F. E., Reid, J. E., Buckley, J. P., & Jayne, B. C. (2001). Criminal
valence and arousal on recollective and nonrecollective recall. Journal of interrogation and confessions (4th ed.). Gaithersburg, MD: Aspen.
Experimental Psychology: Learning, Memory, and Cognition, 39, 663– Israel, L., & Schacter, D. L. (1997). Pictorial encoding reduces false
677. http://dx.doi.org/10.1037/a0028578 recognition of semantic associates. Psychonomic Bulletin & Review, 4,
Gomez, E. (Photographer). (2005, November 26). Wedding ceremony 577–581. http://dx.doi.org/10.3758/BF03214352
[digital image]. Retrieved from http://www.freeimages.com/photo/ Jacoby, L. L., Begg, I. M., & Toth, J. P. (1997). In defense of functional
independence: Violations of assumptions underlying the process-
wedding-ceremony-1435376
dissociation procedure? Journal of Experimental Psychology: Learning,
Goodman, G. S. (2006). Children’s eyewitness memory: A modern history
Memory, and Cognition, 23, 484 – 495. http://dx.doi.org/10.1037/0278-

and contemporary commentary. Journal of Social Issues, 62, 811– 832.
7393.23.2.484
http://dx.doi.org/10.1111/j.1540-4560.2006.00488.x
Jelinek, L., Hottenrott, B., Randjbar, S., Peters, M. J., & Moritz, S. (2009).
Goodman, G. S., Ogle, C. M., Block, S. D., Harris, L. S., Larson, R. P.,
Visual false memories in post-traumatic stress disorder (PTSD). Journal
Augusti, E. M., . . . Urquiza, A. (2011). False memory for trauma-related
of Behavior Therapy and Experimental Psychiatry, 40, 374 –383. http://
Deese-Roediger-McDermott lists in adolescents and adults with histories
dx.doi.org/10.1016/j.jbtep.2009.02.003
of child sexual abuse. Development and Psychopathology, 23, 423– 438. Johnson, M. K., Hashtroudi, S., & Lindsay, D. S. (1993). Source monitor-
http://dx.doi.org/10.1017/S0954579411000150 ing. Psychological Bulletin, 114, 3–28. http://dx.doi.org/10.1037/0033-
Gross, J. J., & Levenson, R. W. (1995). Emotion elicitation using films. 2909.114.1.3
Cognition and Emotion, 9, 87–108. http://dx.doi.org/10.1080/ Joormann, J., Teachman, B. A., & Gotlib, I. H. (2009). Sadder and less
02699939508408966 accurate? False memory for negative material in depression. Journal of
Hamilton, M. (1960). A rating scale for depression. Journal of Neurology, Abnormal Psychology, 118, 412– 417. http://dx.doi.org/10.1037/
Neurosurgery, & Psychiatry, 23, 56 – 62. http://dx.doi.org/10.1136/jnnp a0015621
.23.1.56 Kassin, S. M. (2001). Human judges of truth, deception, and credibility:
Hannigan, S. L., & Reinitz, M. T. (2003). Migration of objects and Confident but erroneous. Cardozo Law Review, 23, 809.
inferences across episodes. Memory & Cognition, 31, 434 – 444. http:// Kassin, S. M., Drizin, S. A., Grisso, T., Gudjonsson, G. H., Leo, R. A., &
dx.doi.org/10.3758/BF03194401 Redlich, A. D. (2010). Police-induced confessions: Risk factors and
Hauschildt, M., Peters, M. J. V., Jelinek, L., & Moritz, S. (2012). Veridical recommendations. Law and Human Behavior, 34, 3–38.
and false memory for scenic material in posttraumatic stress disorder. Kassin, S. M., Ellsworth, P. C., & Smith, V. L. (1989). The “general
Consciousness and Cognition, 21, 80 – 89. http://dx.doi.org/10.1016/j acceptance” of psychological research on eyewitness testimony: A sur-
.concog.2011.10.013 vey of the experts. American Psychologist, 44, 1089 –1098. http://dx.doi
Hess, T. M., Popham, L. E., Emery, L., & Elliott, T. (2012). Mood, .org/10.1037/0003-066X.44.8.1089
motivation, and misinformation: Aging and affective state influences on Kensinger, E. A. (2004). Remembering emotional experiences: The con-
memory. Aging, Neuropsychology, and Cognition, 19, 13–34. http://dx tribution of valence and arousal. Reviews in the Neurosciences, 15,
.doi.org/10.1080/13825585.2011.622740 241–251. http://dx.doi.org/10.1515/REVNEURO.2004.15.4.241
Heuer, F., & Reisberg, D. (1990). Vivid memories of emotional events: Kensinger, E. A., & Corkin, S. (2003). Memory enhancement for emo-
The accuracy of remembered minutiae. Memory & Cognition, 18, 496 – tional words: Are emotional words more vividly remembered than
506. http://dx.doi.org/10.3758/BF03198482 neutral words? Memory & Cognition, 31, 1169 –1180. http://dx.doi.org/
Holliday, R. E., & Hayes, B. K. (2000). Dissociating automatic and 10.3758/BF03195800
intentional processes in children’s eyewitness memory. Journal of Ex- Kensinger, E. A., Garoff-Eaton, R. J., & Schacter, D. L. (2007). Effects of
emotion on memory specificity: Memory trade-offs elicited by negative
perimental Child Psychology, 75, 1– 42. http://dx.doi.org/10.1006/jecp
visually arousing stimuli. Journal of Memory and Language, 56, 575–
.1999.2521
591. http://dx.doi.org/10.1016/j.jml.2006.05.004
Howe, M. L. (2005). Children (but not adults) can inhibit false memories.
Kintsch, W., Welsch, D., Schmalhofer, F., & Zimny, S. (1990). Sentence
Psychological Science, 16, 927–931. http://dx.doi.org/10.1111/j.1467-
memory: A theoretical analysis. Journal of Memory and Language, 29,
9280.2005.01638.x
133–159. http://dx.doi.org/10.1016/0749-596X(90)90069-C
Howe, M. L. (2007). Children’s emotional false memories. Psychological
Knott, L. M., & Thorley, C. (2014). Mood-congruent false memories
Science, 18, 856 – 860. http://dx.doi.org/10.1111/j.1467-9280.2007
persist over time. Cognition and Emotion, 28, 903–912. http://dx.doi
.01991.x .org/10.1080/02699931.2013.860016
Howe, M. L., Candel, I., Otgaar, H., Malone, C., & Wimmer, M. C. (2010). Knott, L. M., Threadgold, E., & Howe, M. L. (2014). Negative mood state
Valence and the development of immediate and long-term false memory impairs false memory priming when problem-solving. Journal of Cog-
illusions. Memory, 18, 58 –75. http://dx.doi.org/10.1080/096582109 nitive Psychology, 26, 580 –587. http://dx.doi.org/10.1080/20445911
03476514 .2014.922091
Howe, M. L., & Malone, C. (2011). Mood-congruent true and false Koo, M., & Oishi, S. (2009). False memory and the associative network of
memory: Effects of depression. Memory, 19, 192–201. http://dx.doi.org/ happiness. Personality and Social Psychology Bulletin, 35, 212–220.
10.1080/09658211.2010.544073 http://dx.doi.org/10.1177/0146167208327191
Howe, M. L., Toth, S. L., & Cicchetti, D. (2011). Can maltreated children Lampinen, J. M., Copeland, S. M., & Neuschatz, J. S. (2001). Recollec-
inhibit true and false memories for emotional information? Child De- tions of things schematic: Room schemas revisited. Journal of Experi-
velopment, 82, 967–981. http://dx.doi.org/10.1111/j.1467-8624.2011 mental Psychology: Learning, Memory, and Cognition, 27, 1211–1222.
.01585.x http://dx.doi.org/10.1037/0278-7393.27.5.1211
Laney, C., & Loftus, E. F. (2010). Truth in emotional memories. In B. H. by salience. Memory, 16, 810 – 820. http://dx.doi.org/10.1080/
Bornstein & R. L. Wiener (Eds.), Emotion and the law (pp. 157–183). 09658210802282708
New York, NY: Springer. Nelson, D. L., McEvoy, C. L., & Schreiber, T. A. (2004). The University
Lang, P. J., Bradley, M. M., & Cuthbert, B. N. (1998). Emotion, motiva- of South Florida free association, rhyme, and word fragment norms.
tion, and anxiety: Brain mechanisms and psychophysiology. Biological Behavior Research Methods, Instruments, & Computers, 36, 402– 407.
Psychiatry, 44, 1248 –1263. http://dx.doi.org/10.1016/S0006- http://dx.doi.org/10.3758/BF03195588
3223(98)00275-3 Niedenthal, P. M., & Setterlund, M. B. (1994). Emotion congruence in
Lang, P. J., Bradley, M. M., & Cuthbert, B. N. (2008). International perception. Personality and Social Psychology Bulletin, 20, 401– 411.
affective picture system (IAPS): Affective ratings of pictures and instruc- http://dx.doi.org/10.1177/0146167294204007
tion manual. Tech. Rep. No. A-8. Gainesville, FL: University of Florida. Niewiadomski, G. (Photographer). (2006, February 6). Shopping [digital
Langner, O., Dotsch, R., Bijlstra, G., Wigboldus, D. H. J., Hawk, S. T., & image]. Retrieved from http://www.freeimages.com/photo/shopping-
van Knippenberg, A. (2010). Presentation and validation of the Radboud 1241024
Faces Database. Cognition and Emotion, 24, 1377–1388. http://dx.doi Ochsner, K. N. (2000). Are affective events richly recollected or simply
.org/10.1080/02699930903485076 familiar? The experience and process of recognizing feelings past.
Loftus, E. F. (1975). Leading questions and the eyewitness report. Cogni- Journal of Experimental Psychology: General, 129, 242–261. http://dx
tive Psychology, 7, 560 –572. http://dx.doi.org/10.1016/0010- .doi.org/10.1037/0096-3445.129.2.242

0285(75)90023-7 O’Rockerz, C. (Photographer). (2006, November 26). Shark [digital im-
Loftus, E. F. (1979). The malleability of human memory. American Sci- age]. Retrieved from http://www.freeimages.com/photo/shark-1384087
entist, 67, 312–320. Otgaar, H., Candel, I., & Merckelbach, H. (2008). Children’s false mem-
Loftus, E. F. (1993). The reality of repressed memories. American Psy- ories: Easier to elicit for a negative than for a neutral event. Acta
chologist, 48, 518 –537. http://dx.doi.org/10.1037/0003-066X.48.5.518 Psychologica, 128, 350 –354. http://dx.doi.org/10.1016/j.actpsy.2008.03
Loftus, E. F., & Bernstein, D. (2004). Strong memories are made of this. .009
Trends in Cognitive Sciences, 8, 199 –201. http://dx.doi.org/10.1016/j Paivio, A. (1971). Imagery and verbal processes. New York, NY: Holt,
.tics.2004.03.005 Rinehart, and Winston.
Palmer, J. E., & Dodson, C. S. (2009). Investigating the mechanisms
Loftus, E., & Ketcham, K. (1996). The myth of repressed memory: False
fuelling reduced false recall of emotional material. Cognition and Emo-
memories and allegations of sexual abuse. New York, NY: Macmillan.
tion, 23, 238 –259. http://dx.doi.org/10.1080/02699930801976663
Loftus, E. F., & Palmer, J. C. (1974). Reconstruction of automobile
Parks, C. M., & Yonelinas, A. P. (2007). Moving beyond pure signal-
destruction: An example of the interaction between language and mem-
detection models: Comment on Wixted (2007). Psychological Review,
ory. Journal of Verbal Learning & Verbal Behavior, 13, 585–589.
114, 188 –201. http://dx.doi.org/10.1037/0033-295X.114.1.188
http://dx.doi.org/10.1016/S0022-5371(74)80011-3
Payne, D. G., Elie, C. J., Blackwell, J. M., & Neuschatz, J. S. (1996).
Lucretious. (Photographer). (2006, April 11). Ready to shoot [digital im-
Memory illusions: Recalling, recognizing, and recollecting events that
age]. Retrieved from http://www.freeimages.com/photo/ready-to-shoot-
never occurred. Journal of Memory and Language, 35, 261–285. http://
1315413
dx.doi.org/10.1006/jmla.1996.0015
Mather, M. (2007). Emotional arousal and memory binding: An object-
Pesta, B. J., Murphy, M. D., & Sanders, R. E. (2001). Are emotionally
based framework. Perspectives on Psychological Science, 2, 33–52.
charged lures immune to false memory? Journal of Experimental Psy-
http://dx.doi.org/10.1111/j.1745-6916.2007.00028.x
chology: Learning, Memory, and Cognition, 27, 328 –338. http://dx.doi
Mayer, J. D., Allen, I. P., & Beauregard, K. (1995). Mood inductions for
.org/10.1037/0278-7393.27.2.328
four specific moods: A procedure employing guided imagery. Journal of
Plutchik, R. (2001). The nature of emotions. American Scientist, 89,
Mental Imagery, 19, 133–150. 344 –350. http://dx.doi.org/10.1511/2001.4.344
Mayer, J. D., & Gaschke, Y. N. (1988). The experience and meta- Porter, S., Bellhouse, S., McDougall, A., Ten Brinke, L., & Wilson, K.
experience of mood. Journal of Personality and Social Psychology, 55, (2010). A prospective investigation of the vulnerability of memory for
102–111. http://dx.doi.org/10.1037/0022-3514.55.1.102 positive and negative emotional scenes to the misinformation effect.
McCloskey, M., & Zaragoza, M. (1985). Misleading postevent information Canadian Journal of Behavioural Science / Revue canadienne des
and memory for events: Arguments and evidence against memory im- sciences du comportement, 42, 55– 61. http://dx.doi.org/10.1037/
pairment hypotheses. Journal of Experimental Psychology: General, a0016652
114, 1–16. http://dx.doi.org/10.1037/0096-3445.114.1.1 Porter, S., Spencer, L., & Birt, A. R. (2003). Blinded by emotion? Effect
McDermott, K. B., & Watson, J. M. (2001). The rise and fall of false recall: of the emotionality of a scene on susceptibility to false memories.
The impact of presentation duration. Journal of Memory and Language, Canadian Journal of Behavioural Science/Revue canadienne des sci-
45, 160 –176. http://dx.doi.org/10.1006/jmla.2000.2771 ences du comportement, 35, 165–175. http://dx.doi.org/10.1037/
Mirandola, C., & Toffalini, E. (2016). Arousal— but not valence—reduces h0087198
false memories at retrieval. PLoS ONE, 11, e0148716. http://dx.doi.org/ Porter, S., ten Brinke, L., Riley, S. N., & Baker, A. (2014). Prime time
10.1371/journal.pone.0148716 news: The influence of primed positive and negative emotion on sus-
Mirandola, C., Toffalini, E., Grassano, M., Cornoldi, C., & Melinder, A. ceptibility to false memories. Cognition and Emotion, 28, 1422–1434.
(2014). Inferential false memories of events: Negative consequences http://dx.doi.org/10.1080/02699931.2014.887000
protect from distortions when the events are free from further elabora- Raina, P., Torrance-Rynard, V., Wong, M., & Woodward, C. (2002).
tion. Memory, 22, 451– 461. http://dx.doi.org/10.1080/09658211.2013 Agreement between self-reported and routinely collected health-care
.795976 utilization data among seniors. Health Services Research, 37, 751–774.
Moritz, S., Gläscher, J., & Brassen, S. (2005). Investigation of mood- http://dx.doi.org/10.1111/1475-6773.00047
congruent false and true memory recognition in depression. Depression Reyna, V. F., & Brainerd, C. J. (1995). Fuzzy-trace theory: An interim
and Anxiety, 21, 9 –17. http://dx.doi.org/10.1002/da.20054 synthesis. Learning and Individual Differences, 7, 1–75. http://dx.doi
Moritz, S., Voigt, K., Arzola, G. M., & Otte, C. (2008). When the half-full .org/10.1016/1041-6080(95)90031-4
glass is appraised as half empty and memorised as completely empty: Reyna, V. F., & Kiernan, B. (1994). Development of gist versus verbatim
Mood-congruent true and false recognition in depression is modulated memory in sentence recognition: Effects of lexical familiarity, semantic
content, encoding instructions, and retention interval. Developmental Spanos, N. P., Cross, P. A., Dickson, K., & DuBreuil, S. C. (1993). Close
Psychology, 30, 178 –191. http://dx.doi.org/10.1037/0012-1649.30.2 encounters: An examination of UFO experiences. Journal of Abnormal
.178 Psychology, 102, 624 – 632. http://dx.doi.org/10.1037/0021-843X.102.4
Reyna, V. F., & Kiernan, B. (1995). Analogical reasoning in cognitive .624
development. Metaphor and Symbol, 10, 37– 41. Stein, N. L., Ornstein, P. A., Tversky, B., & Brainerd, C. (1997). Memory
Reyna, V. F., & Titcomb, A. L. (1997). Constraints on the suggestibility of for everyday and emotional events. Mahwah, NJ: Lawrence Erlbaum
eyewitness testimony: A fuzzy-trace theory analysis. In D. G. Payne & Associates.
F. G. Conrad (Eds.), A synthesis of basic and applied approaches to Storbeck, J. (2013). Negative affect promotes encoding of and memory for
human memory. Hillsdale, NJ: Erlbaum. details at the expense of the gist: Affect, encoding, and false memories.
Rivers, S. E., Reyna, V. F., & Mills, B. (2008). Risk taking under the Cognition and Emotion, 27, 800 – 819. http://dx.doi.org/10.1080/
influence: A fuzzy-trace theory of emotion in adolescence. Developmen- 02699931.2012.741060
tal Review, 28, 107–144. http://dx.doi.org/10.1016/j.dr.2007.11.002 Storbeck, J., & Clore, G. L. (2005). With sadness comes accuracy; with
Roediger, H. L., & McDermott, K. B. (1995). Creating false memories: happiness, false memory: Mood and the false memory effect. Psycho-
Remembering words not presented in lists. Journal of Experimental logical Science, 16, 785–791. http://dx.doi.org/10.1111/j.1467-9280
Psychology: Learning, Memory, and Cognition, 21, 803– 814. http://dx .2005.01615.x
.doi.org/10.1037/0278-7393.21.4.803 Storbeck, J., & Clore, G. L. (2011). Affect influences false memories at
Roosa, M. W., Reinholtz, C., & Angelini, P. J. (1999). The relation of child encoding: Evidence from recognition data. Emotion, 11, 981–989. http://
sexual abuse and depression in young women: Comparisons across four dx.doi.org/10.1037/a0022754
ethnic groups. Journal of Abnormal Child Psychology, 27, 65–76. Susskind, J. M., Littlewort, G., Bartlett, M. S., Movellan, J., & Anderson,
Rotello, C. M., Macmillan, N. A., & Reeder, J. A. (2004). Sum-difference A. K. (2007). Human and computer recognition of facial expressions of
theory of remembering and knowing: A two-dimensional signal- emotion. Neuropsychologia, 45, 152–162. http://dx.doi.org/10.1016/j
detection model. Psychological Review, 111, 588 – 616. http://dx.doi .neuropsychologia.2006.05.001
.org/10.1037/0033-295X.111.3.588 Syed, A. (Photographer). (2006, February 12). Close-up of astronaut
Ruci, L., Tomes, J. L., & Zelenski, J. M. (2009). Mood-congruent false [digital image]. Retrieved from http://www.freeimages.com/photo/
memories in the DRM paradigm. Cognition and Emotion, 23, 1153– astronaut-1424901
1165. http://dx.doi.org/10.1080/02699930802355420 Talmi, D., & Moscovitch, M. (2004). Can semantic relatedness explain the
Russell, J. A. (1980). A circumplex model of affect. Journal of Personality enhancement of memory for emotional words? Memory & Cognition,
and Social Psychology, 39, 1161–1178. http://dx.doi.org/10.1037/ 32, 742–751. http://dx.doi.org/10.3758/BF03195864
h0077714 Tellegen, A., Watson, D., & Clark, L. A. (1999). On the dimensional and
Russell, J. A. (1991). Culture and the categorization of emotions. Psycho- hierarchical structure of affect. Psychological Science, 10, 297–303.
logical Bulletin, 110, 426 – 450. http://dx.doi.org/10.1037/0033-2909 http://dx.doi.org/10.1111/1467-9280.00157
.110.3.426 Tisnado, D. M., Adams, J. L., Liu, H., Damberg, C. L., Chen, W. P., Hu,
Schacter, D. L., Israel, L., & Racine, C. (1999). Suppressing false recog- F. A., . . . Kahn, K. L. (2006). What is the concordance between the
nition in younger and older adults: The distinctiveness heuristic. Journal medical record and patient self-report as data sources for ambulatory
of Memory and Language, 40, 1–24. http://dx.doi.org/10.1006/jmla care? Medical Care, 44, 132–140. http://dx.doi.org/10.1097/01.mlr
.1998.2611 .0000196952.15921.bf
Scherer, K. R. (1999). Appraisal theories. In T. Dalgleish & M. Power Toffalini, E., Mirandola, C., Coli, T., & Cornoldi, C. (2015). High trait
(Eds.), Handbook of cognition and emotion (pp. 637– 663). Chichester, anxiety increases inferential false memories for negative (but not posi-
UK: Wiley. tive) emotional events. Personality and Individual Differences, 75, 201–
Schwartz, S. (1975). Individual differences in cognition: Some relation- 204. http://dx.doi.org/10.1016/j.paid.2014.11.029
ships between personality and memory. Journal of Research in Person- Toffalini, E., Mirandola, C., Drabik, M. J., Melinder, A., & Cornoldi, C.
ality, 9, 217–225. http://dx.doi.org/10.1016/0092-6566(75)90018-5 (2014). Emotional negative events do not protect against false memories
Schwarz, N. (1990). Feelings as information: Informational and motiva- in young adults with depressive–anxious personality traits. Personality
tional functions of affective states. New York, NY: Guilford Press. and Individual Differences, 66, 14 –18. http://dx.doi.org/10.1016/j.paid
Schwarz, N., & Clore, G. L. (1983). Mood, misattribution, and judgments .2014.02.042
of well-being: Informative and directive functions of affective states. Toglia, M. P., Ross, D., Pozzulo, J., & Pica, E. (Eds.). (2014). The elderly
Journal of Personality and Social Psychology, 45, 513–523. http://dx eyewitness in court. New York, NY: Taylor & Francis.
.doi.org/10.1037/0022-3514.45.3.513 Tulving, E. (1985). Memory and consciousness. Canadian Psychology/
Seamon, J. G., Goodkind, M. S., Dumey, A. D., Dick, E., Aufseeser, M. S., Psychologie canadienne, 26, 1–12. http://dx.doi.org/10.1037/h0080017
Strickland, S. E., . . . Fung, N. S. (2003). “If I didn’t write it, why would Tulving, E., & Thomson, D. M. (1971). Retrieval processes in recognition
I remember it?” Effects of encoding, attention, and practice on accurate memory: Effects of associative context. Journal of Experimental Psy-
and false memory. Memory & Cognition, 31, 445– 457. http://dx.doi.org/ chology, 87, 116 –124. http://dx.doi.org/10.1037/h0030186
10.3758/BF03194402 Tussing, A. A., & Greene, R. L. (1999). Differential effects of repetition on
Shaw, J., & Porter, S. (2015). Constructing rich false memories of com- true and false recognition. Journal of Memory and Language, 40, 520 –
mitting crime. Psychological Science, 26, 291–301. http://dx.doi.org/10 533. http://dx.doi.org/10.1006/jmla.1999.2636
.1177/0956797614562862 van Damme, I. (2013). Mood and the DRM paradigm: An investigation of
Snodgrass, J. G., & Corwin, J. (1988). Pragmatics of measuring recognition the effects of valence and arousal on false memory. Quarterly Journal of
memory: Applications to dementia and amnesia. Journal of Experimen- Experimental Psychology: Human Experimental Psychology, 66, 1060 –
tal Psychology: General, 117, 34 –50. http://dx.doi.org/10.1037/0096- 1081. http://dx.doi.org/10.1080/17470218.2012.727837
3445.117.1.34 Van Damme, I., Kaplan, R. L., Levine, L. J., & Loftus, E. F. (2016).
Spanos, N. P. (1996). Multiple identities & false memories: A sociocog- Emotion and false memory: How goal-irrelevance can be relevant for
nitive perspective. Washington, DC: American Psychological Associa- what people remember. Memory. Advance online publication. http://dx
tion. http://dx.doi.org/10.1037/10216-000 .doi.org/10.1080/09658211.2016.1150489
Van Damme, I., & Seynaeve, L. (2013). The effect of mood on confidence Wolinsky, F. D., Miller, T. R., An, H., Geweke, J. F., Wallace, R. B.,
in false memories. Journal of Cognitive Psychology, 25, 309 –318. Wright, K. B., . . . Rosenthal, G. E. (2007). Hospital episodes and
http://dx.doi.org/10.1080/20445911.2013.769440 physician visits: The concordance between self-reports and Medicare
Van Damme, I., & Smets, K. (2014). The power of emotion versus the claims. Medical Care, 45, 300 –307. http://dx.doi.org/10.1097/01.mlr
power of suggestion: Memory for emotional events in the misinforma- .0000254576.26353.09
tion paradigm. Emotion, 14, 310 –320. http://dx.doi.org/10.1037/ Yang, H., Yang, S., Ceci, S. J., & Isen, A. M. (2015). Positive affect
a0034629 facilitates the effect of a warning on false memory in the DRM para-
Watkins, P. C., Mathews, A., Williamson, D. A., & Fuller, R. D. (1992). digm. The Journal of Positive Psychology, 10, 196 –206. http://dx.doi
Mood-congruent memory in depression: Emotional priming or elabora- .org/10.1080/17439760.2014.950177
tion? Journal of Abnormal Psychology, 101, 581–586. http://dx.doi.org/ Yeh, Z. T., & Hua, M. S. (2009). Effects of depressive disorder on false
10.1037/0021-843X.101.3.581
memory for emotional information. Depression and Anxiety, 26, 456 –
Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and vali-
463. http://dx.doi.org/10.1002/da.20453
dation of brief measures of positive and negative affect: The PANAS
Yerkes, R. M., & Dodson, J. D. (1908). The relationship of strength of
scales. Journal of Personality and Social Psychology, 54, 1063–1070.
stimulus to rapidity of habit formation. Journal of Comparative Neurol-
http://dx.doi.org/10.1037/0022-3514.54.6.1063
Weissman, J. S., Levin, K., Chasan-Taber, S., Massagli, M. P., Seage, ogy & Psychology, 18, 459 – 482. http://dx.doi.org/10.1002/cne
G. R., III, & Scampini, L. (1996). The validity of self-reported health- .920180503
care utilization by AIDS patients. AIDS, 10, 775–784. http://dx.doi.org/ Zoellner, L. A., Foa, E. B., Brigidi, B. D., & Przeworski, A. (2000). Are
10.1097/00002030-199606001-00013 trauma victims susceptible to “false memories”? Journal of Abnormal
Wells, G. L., Small, M., Penrod, S., Malpass, R. S., Fulero, S. M., & Psychology, 109, 517–524. http://dx.doi.org/10.1037/0021-843X.109.3
Brimacombe, C. E. (1998). Eyewitness identification procedures: Rec- .517
ommendations for lineups and photospreads. Law and Human Behavior,
22, 603– 647. http://dx.doi.org/10.1023/A:1025750605807
Widom, C. S. (1999). Posttraumatic stress disorder in abused and neglected Received November 18, 2015
children grown up. The American Journal of Psychiatry, 156, 1223– Revision received May 16, 2016
1229. Accepted July 26, 2016 䡲

Bookbinder, S. H., & Brainerd, C. J. (2016) .

Uploaded by

Copyright:

Available Formats

Bookbinder, S. H., & Brainerd, C. J. (2016) .

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Bookbinder, S. H., & Brainerd, C. J. (2016) .

Uploaded by

Copyright:

Available Formats

Psychological Bulletin

Emotion and False Memory: The Context–Content Paradox

Emotion and False Memory: The Context–Content Paradox

Keywords: emotion, false memory, fuzzy-trace theory, mood, opponent processes

particular, whether its valence is positive or negative or how memory experiments.

arousing it is? Third, does the manner in which false memory

Study n Paradigm Population Type of test Groups Mean Comparison Cohen’s d

Delay Neu vs. pos 3.00

Study n Paradigm Population Type of test Groups Mean Comparison Cohen’s d

Word ⫹ picture Negative .43 Word ⫹ picture

Study n Paradigm Population Type of test Groups Mean Comparison Cohen’s d

monitoring framework (Johnson, Hashtroudi, & Lindsay, 1993).

Increases when true memory

Increases for positive moods

Reduced for central details

via relational processing

Increased for negative

Increased for central emotional details;

Increases for negative moods via item

reduced for peripheral or neutral

Increased for positive compared to

Increases when mood is same at

True memory predictions

Supported by recollection and

encoding and retrieval

Emotional valence strengthens gist and

congruence between study and test

Emotion enhances memory accuracy

Mood affects item versus relational

Recollection and familiarity work

opposition for false memory

Memory is enhanced by mood

Emotion reduces attention to

than positive valence

theories in the mainstream memory literature (e.g., Jacoby, Begg,

physical ones (the visual appearance of bagpipe on a computer

screen). The other, which is most often called familiarity, produces

couch, sparrow) produce strong feelings that they were presented,

but recollection suppresses false memories because it consciously

reinstates verbatim content, eliminating a variety of similar-but-

not-verbatim possibilities. Fuzzy-trace theory (FTT; Reyna &

Brainerd, 1995) is the original example of such a theory and the

one that has most often figured in emotion-false memory research.

In FTT, subjects are assumed to store two distinct traces of target

items in parallel: (a) verbatim traces of targets’ surface form (e.g.,

Budson et al. (2006)

Boring Feather Bitter Student Pane

Stay Satin Sour Learn Shade

Palmer & Dodson (2009)

Pretty Pleased Kiss

Pesta et al. (2001)

Peak Rank Shade Divvy Took Nark

Perch Risk Shake Dimwit Hock Pack

Moritz et al. (2005)

Negative/High arousal Negative/Low arousal Positive/High arousal Positive/Low arousal

Anger Cold Board Dirt Baby Beach Bird Fish

Hatred Freeze Two-by-four Bike Boy Wave Flight Fry

Hostile Air Black Brown Adult Blanket Song Shark

List 1 List 1 List 1