A Meta-Analysis of the Facial Feedback Literature: Effects of Facial Feedback on Emotional Experience Are Small and Variable

Abstract

The facial feedback hypothesis suggests that an individual’s experience of emotion is influenced by feedback from their facial movements. To evaluate the cumulative evidence for this hypothesis, we conducted a meta-analysis on 286 effect sizes derived from 138 studies that manipulated facial feedback and collected emotion self-reports. Using random effects meta-regression with robust variance estimates, we found that the overall effect of facial feedback was significant, but small. Results also indicated that feedback effects are stronger in some circumstances than others. We examined 12 potential moderators, and three were associated with differences in effect sizes. 1. Type of emotional outcome: Facial feedback influenced emotional experience (e.g., reported amusement) and, to a greater degree, affective judgments of a stimulus (e.g., the objective funniness of a cartoon). Three publication bias detection methods did not reveal evidence of publication bias in studies examining the effects of facial feedback on emotional experience, but all three methods revealed evidence of publication bias in studies examining affective judgments. 2. Presence of emotional stimuli: Facial feedback effects on emotional experience were larger in the absence of emotionally evocative stimuli (e.g., cartoons). 3. Type of stimuli: When participants were presented with emotionally evocative stimuli, facial feedback effects were larger in the presence of some types of stimuli (e.g., emotional sentences) than others (e.g., pictures). The available evidence supports the facial feedback hypothesis’ central claim that facial feedback influences emotional experience, although these effects tend to be small and heterogeneous.

Full text

COLES, LARSEN, AND LENCH 1
A Meta-Analysis of the Facial Feedback Literature: Effects of Facial
Feedback on Emotional Experience Are Small and Variable
Nicholas A. Coles* & Jeff T. Larsen
University of Tennessee
Heather C. Lench
Texas A&M University
The facial feedback hypothesis suggests that an individual’s experience of emotion is influenced by
feedback from their facial movements. To evaluate the cumulative evidence for this hypothesis, we
conducted a meta-analysis on 286 effect sizes derived from 138 studies that manipulated facial feedback
and collected emotion self-reports. Using random effects meta-regression with robust variance estimates,
we found that the overall effect of facial feedback was significant, but small. Results also indicated that
feedback effects are stronger in some circumstances than others. We examined 12 potential moderators,
and three were associated with differences in effect sizes. 1. Type of emotional outcome: Facial feedback
influenced emotional experience (e.g., reported amusement) and, to a greater degree, affective
judgments of a stimulus (e.g., the objective funniness of a cartoon). Three publication bias detection
methods did not reveal evidence of publication bias in studies examining the effects of facial feedback
on emotional experience, but all three methods revealed evidence of publication bias in studies
examining affective judgments. 2. Presence of emotional stimuli: Facial feedback effects on emotional
experience were larger in the absence of emotionally evocative stimuli (e.g., cartoons). 3. Type of
stimuli: When participants were presented with emotionally evocative stimuli, facial feedback effects
were larger in the presence of some types of stimuli (e.g., emotional sentences) than others (e.g.,
pictures). The available evidence supports the facial feedback hypothesis’ central claim that facial
feedback influences emotional experience, although these effects tend to be small and heterogeneous.
Public significance statement
This meta-analysis suggests that posed emotional facial expressions influence self-reported emotional
experience. However, the size of these effects varies and tends to be small.
Keywords
emotion, facial feedback hypothesis, meta-analysis, replication
Forthcoming in Psychological Bulletin.
This manuscript is not the copy of record and may not exactly replicate the final, published
version. The version of record will available upon publication via its DOI: 10.1037/bul0000194

COLES, LARSEN, AND LENCH 2
“Sometimes your joy is the source of your smile, but sometimes your smile can be the source of
your joy." - Thích Nh t H nhấ ạ
Buddhist monk Thích Nh t H nh’s deep spiritual reflection on human nature has led him to an ấ ạ
idea deeply rooted both in our lay and scientific theories of emotion: feedback from our facial
movements can influence our experience of emotion. In society, people often articulate this idea through
sayings such as “grin and bear it’, “fake it ’til you make it”, and “smile your way to happiness”
(Ansfield, 2007; Kraft & Pressman, 2012; Lyubomirsky, 2008). In psychology, we simply refer to this
idea as the facial feedback hypothesis.
The facial feedback hypothesis suggests that facial movements provide sensorimotor feedback
that (a) contributes to the sensation of an emotion (Izard, 1971; Ekman, 1979; Tomkins, 1962, 1981), (b)
primes emotion-related concepts, facilitating emotion reports (Berkowitz, 1990; Guenther, 1981), or (c)
serves as a cue that individuals use to make sense of ongoing emotional feelings (Allport, 1922, 1924;
Laird & Bresler, 1992; Laird & Crosby, 1974). Unfortunately, more than a century’s worth of research
has not yet clarified whether facial feedback effects are reliable. For example, researchers have produced
a variety of theoretical disagreements about when facial feedback effects should emerge, but it remains
unclear which, if any, of these theories are correct. Furthermore, seventeen labs recently found that even
the most seminal demonstration of facial feedback effects is not clearly replicable (Wagenmakers et al.,
2016). Amid this uncertainty, we provide a narrative review of research on the facial feedback
hypothesis and a meta-analysis of all available experimental evidence. Through narrative review, our
goal is to provide a more full historical account of the facial feedback hypothesis, although one that is
certainly not exhaustive. Through meta-analysis, our goal is to assess the reliability of these facial
feedback effects, including the potential extent and impact of publication bias, and weigh-in on
theoretical disagreements in the facial feedback hypothesis literature. Last, in our discussion, we will
consider how the facial feedback hypothesis broadly fits—or does not fit—into basic, appraisal, and
constructionist theories of emotion.
The term “facial feedback” is often used to denote the effects of facial movements on any
outcome of interest, such as emotion perception (Neal & Chartrand, 2011) or implicit racial bias (Ito,
Chiao, Devine, Lorig, & Cacioppo, 2006). However, the term “facial feedback hypothesis” is usually
reserved to refer to the effects of facial feedback on emotional experience. This review will focus almost
exclusively on the facial feedback hypothesis. Consequently, for our purposes we will use the following
definition of “facial feedback” throughout this review: the effects of facial movements1 prototypically
associated with the expression of emotion on emotional experience.
1 Some researchers have opted to define facial feedback in terms of “facial expressions” instead
of “facial movements”. However, others have argued this terminology is inappropriate because an
individual’s face can resemble an emotional expression even when they are not experiencing that
emotion (e.g., polite smiles; Zajonc, 1985).

COLES, LARSEN, AND LENCH 3
The Origins of Research on the Facial Feedback Hypothesis
Research related to the facial feedback hypothesis was catalyzed by the writings of William
James (1884, 1890, 1894) and Carl Lange (1885), who both proposed that our conscious experience of
emotion is built from sensed changes in our bodily states2. However, although these theorists provided
the theoretical foundation that the facial feedback hypothesis would later be built upon, neither
emphasized the role of the face. For Lange, the face was irrelevant, as he contended that emotional
experience was produced solely by sensed changes in the autonomic nervous system. James, on the
other hand, allowed for the possibility that facial feedback could play some role in the experience of
emotion. However, acknowledging the parallels between his and Lange’s theories, James’ later writings
tended to emphasize the importance of the autonomic nervous system (1890, 1894). Indeed, James
contended that any emotional experience elicited solely by voluntary muscular movements “is apt to be
rather ‘hollow’” (1884, p. 192). Given that James and Lange focused primarily on sensed changes in the
autonomic nervous system, it is perhaps surprising that researchers would eventually narrow their focus
to the role of facial feedback. To speculate why, it helps to consider the historical debates that
surrounded the James-Lange theories.
James and Lange’s theories proved to be one of the most controversial set of ideas in early
psychological research on emotion, meeting strong opposition from the likes of Wundt (1886),
Worcester (1893), Irons (1894), Sherrington (1900) and Cannon (1915). One enduring concern, perhaps
first raised by Worcester (1893), was that autonomic nervous system activity was too undifferentiated to
distinguish among various discrete emotional experiences, such as anger or sadness. Indeed, Cannon
(1915, 1927) noted that (a) different emotional states evoked similar changes in the autonomic nervous
system and (b) non-emotional states shared similar autonomic nervous system patterns as emotional
states. Consequently, Lange’s sole emphasis on autonomic nervous system activity could not explain
how people experienced discrete emotions. James’ theory, on the other hand, suggested that
differentiation was determined not only by the autonomic nervous system but also by skeletomuscular
activity. Although James never specified what these patterns of activity were, Angell (1916) suggested
that emotion differentiation may be determined by facial feedback. Several years later, Allport (1922,
1924) elaborated upon this idea in one of the first formal theories of facial feedback. According to
Allport, autonomic activity created undifferentiated feelings of positivity and negativity that were
subsequently differentiated into discrete emotional categories based on patterns of facial feedback.
Surprisingly, Allport’s key prediction that facial feedback guides the categorization of underlying
positivity/negativity seems to have never been experimentally tested. Nevertheless, Allport’s theory
highlights the historical link between the James-Lange theory of emotion and what would later be
known as the facial feedback hypothesis.
2 What we now refer to as the James-Lange theory of emotion was historically often called the
“James-Lange-Sergi” theory of emotion. The Italian anthropologist Giuseppe Sergi (1894) proposed
similar ideas as James and Lange but his contributions have become less recognized, perhaps because
his work has never been translated to English. Alexander Sutherland (1898) also independently
formulated a similar theory.

COLES, LARSEN, AND LENCH 4
The Varieties of Facial Feedback Hypotheses
Fifty-five years after Allport published his theory of facial feedback, the term facial feedback
hypothesis first appeared in print (Izard, 1977). However, by this point, researchers interested in these
effects had already spent over half a century producing theoretical disagreements about these facial
feedback effects (Allport, 1922, 1924; Bull, 1945, 1946; Gellhorn, 1958, 1964; Tomkins, 1962).
Consequently, the idea quickly splintered into various facial feedback hypotheses (Adelmann & Zajonc,
1989; McIntosh, 1996; Tourangeau & Ellsworth, 1979). Next, we review the four most prominent
theoretical disagreements in the facial feedback hypothesis literature, each of which will be addressed in
some form by our meta-analysis.
Modulation vs. initiation of emotional experience. One of the most active debates surrounding
the James-Lange theories was whether bodily activity—autonomic for Lange, autonomic and
skeletomuscular for James—initiated emotional experiences or only modified ongoing experiences of
emotion. James and Lange believed that bodily activity could do both. For example, Lange stated that
“emotions may be induced by a variety of causes which are utterly independent of disturbances of the
mind” and that they may also “be suppressed and modified by pure physical means” (1922, p. 66).
Similarly, James stated, “If our theory be true…any voluntary arousal of the so-called [bodily]
manifestations of a special emotion ought to give us the emotion itself” and, in a more well-known
quote, “Refuse to express a passion, and it dies” (1884, p. 197). Skeptics, however, were especially
critical of the proposed initiation function. In fact, many of the most well-known critics of the James-
Lange view conceded that it was possible for bodily states to modulate, but not initiate, emotional
experiences (Cannon, 1927; Irons, 1894; Sherrington, 1900; Worcester, 1893). For example, Cannon
believed that the perception of an emotional stimulus caused the thalamus to discharge a signal that
independently produced the experience of emotion and an accompanying set of bodily responses.
However, Cannon acknowledged that these bodily responses might generate “faint” feedback signals,
although he added that they likely played “a minor role in the affective complex” (1927, p. 114).
Consequently, the modulation vs. initiation distinction represents an important disagreement in the
James-Lange Cannon-Bard emotion debates.
Given the historical role of initiation vs. modulation debates in James and Lange’s more general
bodily feedback theories, it is not surprising that similar disagreements emerged when researchers began
developing theories of facial feedback. As noted above, Allport (1922, 1924) believed that facial
feedback could only modulate emotional experience. According to his view, facial feedback guided the
categorization of feelings of positivity and negativity, but it could not initiate emotional experiences in
the absence of these underlying feelings. Gellhorn (1958, 1964) believed that the hypothalamus was the
primary driver of emotional experience but that facial feedback could modulate ongoing hypothalamic
activity. Although Gellhorn suggested it was possible for proprioceptive feedback from the entire body
to initiate emotional experiences, he doubted whether facial feedback could initiate emotional
experiences on its own.
Although most early facial feedback theories stressed a modulating function, researchers later
proposed that facial feedback could also initiate emotional experiences. For example, Ekman (1979)
posited that each discrete emotion is activated by a biologically-innate affect program that produces a set
of bodily responses that later merge in consciousness to form emotional experience. Although these
affect programs were believed to be typically activated by stimuli in the environment, Ekman and his

COLES, LARSEN, AND LENCH 5
colleagues suggested that simply producing a facial configuration associated with an emotion could
activate its affect program, thereby initiating the corresponding emotional experience (Levenson,
Ekman, & Friesen, 1990). Similar predictions are made by some network and grounded cognition
theories of emotion (Berkowitz, 1990; Guenther, 1981; Niedenthal, Barsalou, Winkielman, Krauth-
Gruber, & Ric, 2005), although they typically posit the existence of association-based emotion networks
instead of biologically hardwired affect programs.
It is worth noting that the distinction between modulation and initiation implies that emotional
experiences are episodes with clear-cut beginnings and endings. When emotional experience is
conceptualized as a process that is constantly in flux (e.g., Russell, 2003; Wundt, 1886), the terms
modulation and initiation are less applicable (Ellsworth, 1994). Under this alternative conceptualization,
the initiation vs. modulation distinction can instead be described as the effects of facial feedback on
emotional experience in the presence [modulation] vs. absence [initiation] of external emotional stimuli.
To stay consistent with the language traditionally used in the facial feedback hypothesis literature, we
will use the terms “modulation” and “initiation” when we examine this distinction as a potential
moderator in our meta-analysis. However, we will later discuss these effects in the contexts of theories
that conceptualize emotional experience as a continuous stream.
Discrete vs. dimensional levels of emotional experience. There is ongoing debate in the
affective sciences regarding whether emotions are best conceptualized as discrete categories, such as
happiness, anger, and sadness (Ekman, 1999; Izard, 2007; Tomkins, 1962), or as phenomena that are
reducible to more primitive dimensions, such as valence (i.e., degree of positivity vs negativity) and
arousal (Russell, 1980) or positive and negative activation (Watson, Clark, & Tellegen, 1988). A similar
discrete vs. dimensional distinction exists in the facial feedback hypothesis literature. As previously
noted, facial feedback theories were initially developed to explain the role of facial feedback in the
experience of discrete emotions (Allport, 1922, 1924; Angell, 1916). For example, Tomkins (1962) and
Izard (1971) proposed that affect programs created emotional experience primarily through various
sources of facial feedback3. On the other hand, Zajonc later proposed that facial feedback could also
influence feelings of valence (Zajonc, 1985; Zajonc, Murphy, & Inglehart, 1989). According to Zajonc’s
vascular theory of emotion—which was a modernization of an earlier theory proposed by Israel
Waynbaum (1907)—subjective feelings of valence are caused by general and regional brain
temperatures. Facial movements, Zajonc suggested, regulated air flow through the nasal and cavernous
sinuses, which subsequently produced changes in brain temperature and emotional experience. By this
account, scowling might make people experience more negative affect but not necessarily anger.
Debates regarding the effects of facial feedback on discrete vs. dimensional levels of emotional
experience remain unresolved. Reviews have typically agreed that facial feedback can influence
dimensional reports of emotion. Interestingly, however, the effects of facial feedback on discrete
emotions have been described as nonexistent (Winton, 1986), preliminary (Adelmann & Zajonc, 1989),
3 Throughout the evolution of their theories, Tomkins and Izard were inconsistent in which
sources of facial feedback they discussed. For example, Tomkins (1962) initially emphasized the role of
facial movements, but later revised his theory to emphasize feedback from blood flow, temperature, and
skin on the face (1981). Izard (1971) mainly focused on the role of afferent muscular signals from the
face, but also contended that efferent signals to the facial musculature could contribute to facial
feedback effects.

COLES, LARSEN, AND LENCH 6
mixed (McIntosh, 1996), and controversial (Soussignan, 2004). Later, we will weigh-in on this issue via
moderator analyses.
Awareness of facial feedback manipulation. Another prominent debate in the facial feedback
literature concerns the role of participants’ awareness of the purpose of facial feedback manipulations.
For early facial feedback researchers, this raised the possibility that facial feedback effects are driven by
demand characteristics (Buck, 1980). To address the role of awareness, Strack, Martin, and Stepper
(1988) introduced the first incidental facial feedback manipulation: the pen-in-mouth procedure. In two
studies ostensibly about psychomotor coordination, participants held a pen in their mouth in a manner
that either forced them to smile (pen held in teeth) or prevented them from doing so (pen held by lips).
While maintaining these poses, participants viewed humorous cartoons and reported how amused they
felt. Consistent with the facial feedback hypothesis, Strack and colleagues reported that participants who
posed smiles reported feeling more amused by cartoons than those who were prevented from smiling.
In addition to reducing the role of demand characteristics, Strack and colleagues suggested that
their findings indicated that facial feedback effects occurred outside of awareness, an issue that theorists
disagreed about. For example, some researchers suggested that such effects were driven by physiological
mechanisms that occur outside of people’s awareness (Gellhorn, 1958, 1964; Zajonc, 1985). Others
contended that they are driven by consciously-accessible proprioception or self-perception mechanisms
(Izard, 1977; Laird & Bresler, 1992; Laird, 1974; Tomkins, 1962). For example, Laird suggested that
emotional experience is built from a self-perception process (e.g., people might conclude that they are
happy because they perceive themselves to be smiling). Since Strack and colleagues created a
manipulation that limited the degree to which participants were aware that their facial configurations
resembled an emotional expression, they concluded that their results were inconsistent with these latter
set of theories.
More recently, there is uncertainty regarding the reliability of Strack and colleagues’ pen-in-
mouth effect. Seventeen labs conducted pre-registered replications of one of Strack et al.’s (1988) two
studies, and none of the replications found that the pen smiling manipulation made people feel
significantly more amused while viewing cartoons (Wagenmakers et al., 2016; but see Noah, Schul, &
Mayo, 2018, Strack, 2016). This failure-to-replicate has revived the debate about the role of participant
awareness, although no one has yet considered the cumulative evidence for incidental facial feedback
manipulations. Importantly, several other researchers have tested facial feedback effects using incidental
facial feedback manipulations, some using the pen-in-mouth manipulation (e.g., Soussignan, 2002),
others creating new incidental manipulations. For example, Larsen, Kasimatis, and Frey (1992)
incidentally manipulated frowning behavior by attaching golf tees to participants’ brow regions and
asking them to touch the tees together (by pulling the brows downward). By comparing studies that used
such incidental facial feedback manipulations to studies that used manipulations more susceptible to
demand characteristics, the cumulative evidence for the role of participant awareness can be evaluated.
Effects on affective judgments. The central tenet of the facial feedback hypothesis is that facial
feedback influences emotional experience. However, many researchers in the facial feedback literature
have expanded upon the original focus of the facial feedback hypothesis by suggesting that facial
feedback also influences other emotional responses, including those we will call affective judgments. We
use this term to refer broadly to judgments about the emotional characteristics of some stimulus. For
example, a question about the objective funniness of a cartoon can be considered an affective judgment

COLES, LARSEN, AND LENCH 7
because it is a question about the stimulus, not about how an individual felt when they encountered that
stimulus.
Researchers in the facial feedback literature have disagreed about whether facial feedback can
influence affective judgments. For example, Strack and colleagues (1988) had participants report both
their affective judgments about the cartoons (i.e., how objectively funny they thought the cartoons were)
and their emotional experience (i.e., how amused they were by the cartoons). They found evidence that
facial feedback influenced emotional experience, but little evidence that it influenced affective
judgments. However, others have contended that the effects of facial feedback on emotional experience
can subsequently influence affective judgments (Dzokoto, Wallace, Peters, & Bentsi-Enchill, 2014;
Ohira & Kurono, 1993). For example, Ohira and Kurono (1993) reported that frowning participants
judged a target person to be more negative and that smiling participants judged them to be more
positive. Others have suggested that facial feedback can influence affective judgments because these
cognitive process are partially grounded in the automatic reactivation of related somatosensory and
motor systems (e.g., facial movements; Davis et al., 2015). In our meta-analysis, we will examine the
effects of facial feedback both on emotional experience and affective judgments and assess whether
these different outcomes moderate facial feedback effects.

COLES, LARSEN, AND LENCH 8
Table 1
Moderator coding criteria
Moderator (bolded) and level Criteria
Modulation vs. initiation
Modulation Emotional stimuli were presented.
Initiation Either no stimuli were presented, or non-evocative stimuli/tasks were
presented (e.g., neutral images & filler tasks).
Discrete vs. dimensional
emotion measure
Discrete Measures of discrete emotions (such as anger or happiness) were
collected.
Discrete emotion
Anger Classified according to the discrete emotions identified in Ekman and
Cordaro’s (2011) basic emotion theorya. Some studies measured emotions
that were similar, but not included in Ekman and Cordaro’s classification.
We categorized these cases into their most similar discrete emotion.
Disgust
Fear
Happiness
Sadness
Surprise
Dimensional Bipolar measures or measures of positive or negative affect, were
reported.
Dimensional emotion
Positivity If the facial feedback manipulation was positive in nature (e.g., smiling), or
the facial feedback manipulation was neither positive nor negative (e.g.,
suppression) but the stimuli were positive.
Negativity If the facial feedback manipulation was negative in nature (e.g., frowning),
or the facial feedback manipulation was neither positive nor negative (e.g.,
suppression) but the stimuli were negative.
Awareness of facial feedback
manipulation
Aware For ease of comparison, only study designs that used a control group
comparison were included: Exaggeration-control, Posing-control,
Suppression-control. Botox-control was excluded from both levels of this
moderator due to uncertainty regarding the degree to which participants
recognize the impact of botulinum toxins on facial movements.
Unaware For ease of comparison, only study designs that used a control group
comparison were included: Incidental-control. Botox-control was excluded
from both levels of this moderator due to uncertainty regarding the degree
to which participants recognize the impact of botulinum toxins on facial
movements.
Awareness of video recording
Yes Participants were told they were going to be recorded or the methodology
stated that a video camera was placed within participant view.
No Methodology stated that participants were unaware of video recording, that
the video camera was hidden, or that there was not a video camera.
Emotional experience vs.
affective judgments
Emotional experience Participants reported their emotional experience (e.g., How amused did
the photo make you feel?).
Affective judgments Participants reported their affective reaction to the stimulus (e.g., How
funny is the photo?).
Facial feedback manipulation
Botox-control All procedures were coded in a manner that captures both the procedure
used in the experimental group and the procedure used in the comparison
Exaggeration-control
Posing-control

COLES, LARSEN, AND LENCH 9
group. Incidental-control
Suppression-control
Posing-posing
Posing-suppression
Incidental-incidental
Incidental-suppression
Suppression-exaggeration
Between vs. within-subjects
design
Between Effect size estimates from between-subject comparisons.
Within Effect size estimates from within-subject comparisons.
Type of stimuli
Audio
Film
Imagined Scenarios
Pictures
Sentences
Social context
Stories
Gender (Proportion of women) Calculated using each study’s reported gender composition for their entire
sample. If studies excluded participants and reported the gender
composition of their remaining sample, we used these updated values.
Timing of measurement
During manipulation Methodology stated participants held the manipulation while providing self-
reports, or participants were instructed to hold the manipulation throughout
the experiment.
After manipulation Methodology stated participants did not hold the manipulation while giving
self-reports, there was a break between the manipulation and self-reports,
or participants were instructed to hold the manipulation at a specific
moment in the experiment.
Publication year
Publication status
Unpublished Dissertations, unpublished data, and in-prep manuscripts.
Published Peer-reviewed articles.
Note. a Ekman and Cordaro (2011) included contempt in their list of basic emotions, but no facial feedback
studies have investigated contempt.

COLES, LARSEN, AND LENCH 10
Current Meta-Analysis
The last meta-analysis on the facial feedback hypothesis was performed 30 years ago and
revealed a medium effect size (r = .34) among 16 studies that included 532 participants (Matsumoto,
1987). Two more recent meta-analyses have included facial feedback effects but either did not address
the effects of facial feedback separately from other types of behavioral manipulations (e.g., changing
breathing rate; Lench, Flores, & Bench, 2011) or included a very small group of studies (s = 8;
Westermann, Spies, Stahl, & Hesse, 1996). Given (a) the large number of studies that have been
published since the last meta-analysis specifically reviewing the facial feedback hypothesis, (b) recent
controversies over the reliability of some facial feedback effects (Wagenmakers et al., 2018), (c)
laypersons’ belief in the facial feedback hypothesis (e.g., “smile your way to happiness” Lyubomirsky,
2008), and (d) unresolved theoretical disagreements, we believe that an up-to-date meta-analysis is in
order.
Moderators of Interest
In addition to coding for moderators that addressed the aforementioned theoretical disagreements
(i.e., modulation vs. initiation; discrete vs. dimensional; role of awareness; effects on affective judgment
vs. experience), we examined potential methodological moderators of facial feedback effects. All
moderator coding was completed by three coders (the lead author and two trained research assistants)
who discussed and resolved discrepancies throughout the coding process. Coding criteria for each
moderator are available in Table 1.
Facial feedback manipulation procedure. Facial feedback has been manipulated in a variety of
ways, including: tasks that incidentally produce facial postures (e.g., Strack et al., 1988), experimenter-
instructed facial posing (e.g., Tourangeau & Ellsworth, 1979), expression suppression (e.g., Gross,
1998), expression exaggeration (e.g., Demaree et al., 2006) and Botox treatments4 (e.g., Davis, Senghas,
Brandt, & Ochsner, 2010). Methodological differences are a common source of variation in effect sizes,
and Izard (1990a) speculated that some facial feedback methodologies may produce larger effect sizes
than others. We examined this possibility by including manipulation procedure in our moderator
analyses.
Effect sizes in this meta-analysis represent the magnitude of the difference between two groups.
Therefore, codes for the moderator had to convey the procedure used in both groups. Consequently, we
created a moderator variable that captures both the procedure used in the experimental group and the
procedure used in the comparison group (for a similar approach, see Webb, Miles, & Sheeran, 2012).
For example, if a study compared the effects of posing a smile to the effects of suppressing a smile, it
was coded as “posing-suppression”. If the study compared the effects of posing a smile to posing a
frown, it was coded as “posing-posing”.
4 Studies that examine the effects of Botox on emotional experience are often quasi-experimental
in that they compare people who did vs. did not opt to receive Botox injections. For ease of
communication, we will refer to both experimental and quasi-experimental approaches as
manipulations.

COLES, LARSEN, AND LENCH 11
Between vs. within-subject designs. An early criticism of the facial feedback literature was that
it focused almost exclusively on within-subject designs. Buck (1980) noted that all studies that found
evidence for the facial feedback hypothesis to that point had employed within-subject designs, which he
suggested raised concerns about demand characteristics. Since then, researchers have used more
between-subject than within-subject designs. To assess whether between- and within-subject designs
yield different effect sizes, we investigated the experimental design of an effect-size estimate as a
potential methodological moderator.
Type of stimuli. Facial feedback experiments that include emotionally evocative stimuli have
used a variety of stimuli, including emotional sounds (e.g., Vieillard, Harm, & Bigand, 2015), images
(e.g., Strack et al., 1988), films (e.g., Soussignan, 2002), imagined scenarios (e.g., McCanne &
Anderson, 1987), sentences (e.g., Lewis, 2012), stories (e.g., Paredes, Stavraki, Briñol, & Petty, 2013),
and emotional social contexts (e.g., Butler et al., 2003). We examined whether stimulus type is a
significant moderator of facial feedback effects.
Gender. There are many well-documented gender effects in the emotion literature. For example,
researchers have reported gender differences in emotion regulation (Gross & John, 2003; McRae,
Ochsner, Mauss, Gabrieli, & Gross, 2008; Nolen-Hoeksema & Aldao, 2011), emotional expressivity
(Kring, Smith, & Neale, 1994), and smiling behavior (LaFrance, Hecht, & Paluck, 2003). Some
researchers have suggested that there may also be gender differences in bodily feedback effects, like
facial feedback. For example, Pennebaker and Roberts (1992) suggested that men rely more on bodily
cues than women when making inferences about what emotions they are experiencing. If so, women
should show smaller facial feedback effects than men. To examine whether there are gender differences
in facial feedback effects, we examined the proportion of women in the sample as a moderator. If
women exhibit weaker facial feedback effects, we should find that studies with higher proportions of
women have smaller effect sizes.
Awareness of video recording. In a commentary on Wagenmakers et al.’s (2016) failure-to-
replicate, Strack (2016) suggested that one reason the results of the original experiment may not have
replicated is that cameras were directed at participants in the replication studies. Strack reasoned that
awareness of video recording may induce a subjective self-focus that disrupts the flow of experience and
suppresses emotional responses. More recently, Noah and colleagues (2018) tested this possibility by
manipulating participants’ awareness of video recording. They found marginal evidence of the effects of
video camera presence. To examine the cumulative evidence for this claim, we coded whether
participants were aware vs. unaware of video recording.
Timing of measurement. Studies differ in whether the dependent variable is measured during or
after the facial feedback manipulation. For example, Reisenzein and Studtmann (2007) had participants
maintain a facial configuration until they had completed a measure of emotional experience. In contrast,
Duncan and Laird (1980) had participants complete a measure of emotional experience after completing
the posing procedure. Research indicates that emotions can be fleeting (Verduyn, Delaveau, Rotgé,
Fossati, & Van Mechelen, 2015), so we reasoned that facial feedback effects may be stronger when the
dependent measure is assessed during the facial feedback manipulation. To test this hypothesis, we
investigated timing of measurement as a moderator.
Publication year. The decline effect refers to the observation that effect sizes sometimes get
smaller over time (Lehrer, 2010). It is unknown which mechanism produces this phenomenon, but
Schooler (2011) suggested that it may be driven by statistical self-correction or publication bias. Yet

COLES, LARSEN, AND LENCH 12
another possibility is that researchers focus on more nuanced and conceptually weaker effect sizes over
time. To test whether there is a decline effect in the facial feedback literature, we tested publication year
as a moderator.
Publication status. Publication bias is a well-documented phenomenon in science (Rothstein,
Sutton, & Borenstein, 2006). Publication bias poses a risk to meta-analyses if the unpublished literature
differs systematically from the published literature. If published studies have larger effect sizes and are
more likely to have significant findings than studies that are not published, then a meta-analysis of only
the published studies will yield inflated effect size estimates. Fortunately, we were able to gather several
unpublished records for this meta-analysis (reviewed later in Selection of studies). This moderator was
included to test whether published studies had larger effects than the unpublished studies we obtained.
Method
All materials for this meta-analysis are available on the Open Science Framework
(https://osf.io/v8kxb/), including: (a) pre-registered analysis plan, (b) detailed outline of search strategy,
(c) list of all screened articles and other reports (e.g., dissertations, unpublished manuscripts) with
explanations of exclusions, (d) quotes and rationale behind all moderator and effect size coding
decisions, (e) materials and instructions for an open-source plot extraction tool used to extract relevant
statistics (e.g., means) that were not reported but were displayed in figures (Rohatgi, 2011), and (f) R
code to replicate all analyses. After public discussion of a pre-print of this paper and feedback from
peer-reviewers, some minor modifications were made to the pre-registration plan. Materials detailing
these modifications are also available on the Open Science Framework.
Scope
For the purposes of this meta-analysis, we focused only on dependent variables that matched the
facial feedback manipulation. For example, if a researcher manipulated whether participants smiled and
collected measures of both happiness and sadness, we focused only on the happiness ratings. Although
the effects of facial feedback manipulations on non-target emotions would be theoretically interesting to
debates about whether specific facial poses have emotion-specific effects (e.g., whether posing sadness
can produce sadness, but not other discrete negative emotions), this question fell beyond our scope.5
Selection of Studies
Our literature search strategy was developed in consultation with an experienced librarian at the
University of Tennessee. Additional searches performed after reviewer feedback are denoted with
asterisks. Figure 1 is a PRISMA flowchart that outlines the overall process for selecting studies for
5 We did not examine the effects of facial movements on non-target emotions because it would
have further increased the degree to which the effect sizes in the meta-analysis are dissimilar and
complicated the analyses. Furthermore, we felt it was more important to first focus on the simpler
question of whether facial feedback influences target emotions before examining whether it can also
influence non-target emotions.

COLES, LARSEN, AND LENCH 13
inclusion in the meta-analysis (Moher, Liberati, Tetzlaff, Altman, & The PRISMA Group, 2009). To
gather reports, we first searched the following for articles published before 2017:
-PsycInfo: SU.EXACT.EXPLODE("Feedback") AND SU.EXACT("Facial Expressions")
-PsycInfo: expressive suppression AND "emotion regulation"
-*PsycInfo: ("embodiment" OR "sensorimotor simulation") AND ("emotion" OR "cognition")
AND "face"
-Pubmed: feedback[All Fields] AND "facial expressions"[All Fields] OR "facial feedback"
OR "facial feedback hypothesis"
-*Pubmed: ("embodiment" OR "sensorimotor simulation") AND ("emotion" OR "cognition")
AND "face"
-Web of Science: ("feedback" AND "facial expression*" AND emotion) OR ("facial
feedback" AND emotion) OR "facial feedback hypothesis"
-*Web of Science: ("embodiment" OR "sensorimotor simulation") AND ("emotion" OR
"cognition") AND "face"
-References of 17 reviews on the facial feedback hypothesis (Adelmann & Zajonc, 1989;
Buck, 1980; Gerrards-Hesse, Spies, & Hesse, 1994; Izard, 1990b; Laird, 1984; Lench et al.,
2011; Martin, 1990; Matsumoto, 1987; McIntosh, 1996; Price & Harmon-Jones, 2015; Price,
Peterson, & Harmon-Jones, 2012; Soussignan, 2004; Strack, 2016; Webb, Miles, & Sheeran,
2012; Wagenmakers et al., 2016; Westermann et al., 1996; Whissell, 1985)
To capture the unpublished literature, we conducted the following searches:
-ProQuest Dissertations and Theses Global: "facial feedback hypothesis" AND "emotion"
-*ProQuest Dissertations and Theses Global: ("embodiment" OR "sensorimotor simulation")
AND ("emotion" OR "cognition") AND "face")
-Calls for unpublished data: SPSP Open Forum, ResearchGate, Facebook Psychological
Methods Discussion Group
-Direct requests for unpublished data from 81 facial feedback researchers identified through
our screening process.
After removing duplicate records, there were 1,595 records to screen. The lead author screened
the titles and abstracts of these records for studies that manipulated facial movements and measured
emotional experience or affective judgments. If there was any doubt about an article’s eligibility, it was
retained for further review. During this screening, 1,158 full-text reports were excluded, leaving 437
reports to assess for eligibility.
To assess full-text reports for eligibility, the lead author used the following criteria:
(a) Facial movements were manipulated. To provide a clear assessment of the facial feedback
hypothesis, studies that simultaneously manipulated facial movements and other body postures
were excluded.
(b) Measures of emotional experience or affective judgments were collected. Studies that measured
pain were excluded because previous facial feedback and emotion researchers have argued that
pain is not a clearly emotional outcome (e.g., Lumley et al., 2011; McIntosh, 1996).
(c) Data from non-clinical samples were reported. If a study examining a clinical sample also
included data from a non-clinical sample, only the data from the non-clinical sample was
included.
(d) Information necessary to compute effect sizes was included (reviewed in Variable Coding).
(e) Article was in English.
(f) Article was a primary study whose relevant results were not reported in a previous record.

COLES, LARSEN, AND LENCH 14
Based on these criteria, 98 reports were included that contained a total of 138 studies. From these
138 studies, 286 effect sizes were extracted.

COLES, LARSEN, AND LENCH 15
Figure 1. PRISMA-style flow chart showing selection of studies for meta-analysis on facial feedback literature.

COLES, LARSEN, AND LENCH 16
Variable Coding
Moderator coding was completed by three coders (the lead author & two trained research
assistants) who discussed and resolved discrepancies throughout the coding process (see Table 1 for
coding criteria). The lead author extracted all information related to effect size (sample sizes, means and
standard deviation, t-values, F-values, or p-values). If relevant statistics were not included in the report,
but informative graphs were included, we used an open-source program to extract data from the graphs
(Rohatgi, 2011). If a report did not include additional information or graphs but did indicate whether
there was or was not a significant facial feedback effect, we assumed conservative p-values of .05 or .50,
respectively, in our effect size calculations. If the sample size for each condition was not reported in a
study with between-subject comparisons, we estimated sample size by dividing the total sample by the
number of conditions.
Meta-Analytic Approach
Effect size index.We used Cohen’s standardized d as our effect size index, which represents the
difference between two group means divided by their pooled standard deviation (Cohen, 1988). Effect
sizes were calculated in R 3.4.0 (R Core Team, 2017) using formulas provided by Borenstein (2009).
Effect sizes were calculated so that positive values always indicated an effect consistent with the facial
feedback hypothesis. For example, the facial feedback hypothesis predicts that facilitating facial
expressions leads to increased emotional intensity, whereas suppressing facial expressions leads to
decreased emotional intensity. Therefore, increased emotional intensity in a facilitative condition (e.g.,
Flack, 2006) and decreased emotional intensity in a suppression condition (e.g., Gross & Levenson,
1997) both represent predicted facial feedback effects and were coded in the positive direction.
For within-subject designs, the correlation between the pre- and post- measures is necessary for
calculating Cohen’s d. Unfortunately, this correlation is rarely reported, so it is recommended that meta-
analysts assume a correlation and perform a sensitivity analysis on the assumed value (Borenstein,
2009). We pre-registered a default correlation of .50 but performed additional analyses to determine the
impact of the assumed correlation on the overall effect size estimate (testing r = .10, .30, .50, .70, 90).
This did not affect the inferences made from the overall effect size, so we only report analyses that used
the default r = .50 value. All effect sizes are reported in Table 2.
Meta-analysis with robust variance estimates. Fifty-three percent of studies provided multiple
effect sizes of interest. For example, Flack, Laird, and Cavallaro (1999b) examined the impact of angry,
sad, fearful, and happy facial expression on emotional experience. When a study provides multiple effect
size estimates, it is best to record all effect sizes in order to be comprehensive. However, one drawback
of this approach is that it violates the statistical assumption that effect sizes are independent. There are
several ways to deal with dependency in meta-analysis. The simplest approach is to aggregate effect
sizes drawn from the same study (Borenstein, Hedges, Higgins, & Rothstein, 2009; Rosenthal & Rubin,
1986). Although this removes dependency, it results in a loss of information regarding comparisons
among multiple levels of a moderator in a single study. A second approach is to use multivariate meta-
regression (Raudenbush, Becker, & Kalaian, 1988). However, this approach requires knowledge of the
underlying covariation structure among effect sizes, which is almost always unknown. A third approach

COLES, LARSEN, AND LENCH 17
is to use meta-analysis with robust variance estimates (RVE) (Hedges, Tipton, & Johnson, 2010).
Similar to its application in general linear models, RVE can be used in meta-analysis to adjust for
dependencies among effect sizes. This approach does not result in the loss of any information, does not
require knowledge of the underlying correlation structure, and can accommodate multiple sources of
dependencies. We use this RVE approach to estimate our overall effect size, conduct moderator analyses,
and perform most of our publication bias analyses.6
Meta-analysis with RVE weighting scheme. When averaging the results of multiple studies,
meta-analyses typically give more weight to effect sizes with higher precision (i.e., smaller variance) via
a procedure termed inverse-variance weighting. Meta-analysis with robust variance estimates uses
similar weighting schemes that provide adjustments for the types of dependency among effect sizes. If
dependency primarily arises from studies providing multiple effect sizes for the same outcome of
interest, the correlated effects weighting scheme is recommended. On the other hand, if dependency
primarily arises from authors reporting multiple studies, the hierarchical effects weighting scheme is
recommended (Hedges et al., 2010). In practice, both types of dependencies often exist in a meta-
analysis, and it is recommended to choose weighting based on the predominant type of dependency
(Tanner-Smith & Tipton, 2014). Twenty-one percent of the reports in the present meta-analysis included
multiple studies and 53% of the reports included studies that provided multiple effect sizes for the
outcome of interest. Therefore, we used the correlated effects weighting scheme.
When calculating weights, meta-analysis with RVE requires an estimate of the within-study
effect-size correlation (i.e., the average correlation among the dependent effect sizes). The default
assumed value is r = .80. We pre-registered this as the default value to inform our conclusions but
performed additional sensitivity analyses to determine the impact of this assumed value on our overall
effect estimate (testing r = 0, .20, .40, .60, .80, 1.00). This did not affect inferences about effect sizes, so
we only report analyses that used the default value of r = .80.
Testing overall effects and moderators. To test the overall effect size, we fit an intercept-only
random-effects meta-regression model with RVE using the R package, robumeta (Fisher & Tipton,
2015). The intercept of this model can be interpreted as the precision-weighted overall effect size,
adjusted for correlated-effect dependencies. We used the same approach to calculate overall effect sizes
for each level of each moderator. For cases where a level of a moderator had too few observations for
the RVE approach, we calculated overall effect sizes using random-effects meta-regression model (these
exceptions are noted in Table 3).
We also used the RVE approach to perform separate hypothesis tests for the effects of each
moderator7. Continuous moderators were entered into a meta-regression equation without
6 Although we believe meta-analysis with RVE was the best approach for our data analysis, we
also calculated the overall effect size using the Borenstein et al. (2009) aggregation method for
correcting for dependencies, three-level meta-analysis (Van den Noortgate, Lopez-Lopez, Marin-
Martinez, & Sanchez-Meca, 2015), and a random-effects meta- analysis without corrections for
dependencies. We obtained results that were nearly identical to those generated by the RVE approach.
Therefore, we only report the results of the RVE approach.
7 In our pre-registration plan, we also noted that we would re-examine important theoretical
moderators with any significant methodological moderators we find included as covariates. These
analyses did not affect our conclusions, so we do not report them here.

COLES, LARSEN, AND LENCH 18
transformation, except publication year, which was centered at 2017 to ease interpretation of the
regression intercept. Categorical moderators with two levels (i.e., type of experience) were dummy
coded and entered into meta-regression equations. The significance test corresponding to the regression
coefficient for the predictor variable in these models can be interpreted as a test of whether the variable
is a significant moderator.
Examining categorical moderators with more than two levels required an additional step. Like
the former process, they were first dummy coded and entered into meta-regression equations. However,
the regression coefficients only test whether there is a difference between a single level of a moderator
and a single comparison level. To perform an omnibus test of moderators with more than two levels, we
followed the recommendations of Tanner-Smith et al. (2016) and conducted Approximate Hotelling-
Zhang with small sample correction tests using the clubSandwhich R package (Pustejovsky, 2017). This
test produces an F-value that indicates whether there is a difference among all levels of the moderator.
We forewarn the reader that the Approximate Hotelling-Zhang produces atypical degrees of freedom,
and refer the curious reader to Tanner-Smith et al. (2016) for a more detailed explanation.
Notably, moderator analyses typically need a large amount of observations to achieve high power
(Hedges & Pigott, 2004), and the power to detect moderators is reduced by higher levels of
heterogeneity and robust variance estimation procedures. Consequently, null effects in our tests of
moderation should be cautiously interpreted.
Outlier detection. Methods for identifying outliers for meta-regression models with RVE are not
yet available, so we identified outliers in a random-effects intercept-only meta-regression model using
the base R function influence.measures. After fitting an intercept-only meta-regression model, this
function calculates a variety of influential outlier diagnostics (such as covariance ratios, Cook’s
distances, and diagonal elements of the hat matrix), and identifies cases that are influential on any one of
the diagnostic criteria.
Examining publication bias. Many methods for testing the extent and impact of publication
bias in meta-analysis have been developed. Unfortunately, most of these methods were developed and
tested under the assumption that the effect sizes are independent, which is typically unrealistic in meta-
analyses of the psychology literature. Below, we outline the two approaches we used to examine
publication bias with dependent effect sizes.
Publication bias analyses on aggregated dependent effect sizes. The most common way to
assess publication bias with dependent structures is to aggregate the dependent effect sizes and perform
standard publication bias tests on the aggregated estimates. To aggregate dependent effect sizes, we used
the R package MAd (Del Re & Hoyt, 2010). Using the Borenstein et al. (2009) aggregation method, this
function calculates aggregated effect size and effect size variance estimates by taking into account a pre-
specified correlation among the clusters of dependent effect sizes (set, by default, at r = .508).
8 When the correlation among clusters of dependent effect sizes is unknown, it is recommended
that meta-analysts assume a correlation and perform additional sensitivity analyses on this assumed
value (Borenstein, 2009). In line with this recommendation, we assumed a default correlation of r = .50
and performed sensitivity analyses to determine impact of the assumed correlation on our statistical tests
of publication bias (testing r = .10, .30, .50, .70, 90). We indicate in the manuscript the one instance
where this affected our conclusions.

COLES, LARSEN, AND LENCH 19
We then used these aggregated estimates to examine the funnel plot distribution of effect sizes
and perform three statistical tests of publication bias: trim-and-fill (Duval & Tweedie, 2000), weight-
function modeling (Vevea & Hedges, 1995), and PET-PEESE (Stanley & Doucouliagos, 2014).
PET-PEESE with robust variance estimates. Although researchers typically aggregate
dependent effect sizes before examining publication bias, it is worth noting that the PET-PEESE
approach can be conducted using RVE. Because PET-PEESE is essentially a meta-regression equation
with standard error or variance as a predictor, robust variance estimates can easily be implemented when
fitting the meta-regression model. Compared to the aggregation method, the benefit of this approach is
that it does not require us to assume a correlation among the clusters of dependent effect sizes. However,
a drawback is that the statistical properties of this approach are currently unknown.
Publication bias sensitivity analyses. Heterogeneity, which represents how much variation is
observed beyond what would be expected from sampling error alone, can pose problems for many tests
of publication bias (Stanley, 2017; Sterne et al., 2011; Terrin, Schmid, Lau, & Olkin, 2003). Therefore,
we performed pre-planned sensitivity analyses on our publication bias tests by splitting our dataset by
significant moderators.
In instances where we did not uncover any evidence of publication bias, we conducted additional
pre-planned sensitivity analyses by re-running the analyses: 1) excluding suppression studies; 2)
excluding Wagenmakers et al. (2016), and 3) excluding Wagenmakers et al. (2016) and all unpublished
data. The purpose of these sensitivity analyses was to ensure that publication bias was not masked by
subsets of studies that we might expect to skew the distribution of effect sizes. For example, the emotion
regulation literature suggests that suppression is a relatively ineffective way of managing emotional
experience (e.g., Gross, 1998). Therefore, it is feasible that publication bias could be masked by the
inclusion of relatively small effect sizes from suppression studies. By this same logic, we reasoned that
the replication and unpublished studies could have similar effects on our publication bias analyses.
These sensitivity analyses never affected our conclusions, but we report them to convey the robustness
of the publication bias results.

COLES, LARSEN, AND LENCH 20
Results
Overall analyses included 98 articles, 138 studies, and 286 effect sizes (see Table 2). Notably,
20% of these effect sizes came from unpublished sources.
Overall Effect
Using meta-regression with RVE, the overall size of the effect of facial feedback on self-reported
affective experience was d = 0.20, 95% CI [0.14, 0.26], t(137) = 6.42, p = .000000001. This indicates
that, overall, facial feedback manipulations have a small effect on emotional experience and affective
judgments.
Outlier Detection
To examine whether there were any influential outliers, we used the base R function
influence.measures. This method detected eight influential outliers9, two of which were in the negative
direction. Removing the eight outliers did not affect our overall effect size estimate (adjusted d = 0.19,
95% CI [0.13, 0.25], t(137) = 6.31, p = .000000004) or any of the overall publication bias results we
report below. Therefore, all effect size estimates were retained in all further analyses.
Moderator Analyses
There was a large amount of heterogeneity in the effect sizes (T2 = 0.11, I2 = 75.41). Such
heterogeneity suggests that there may be meaningful differences among studies that can be further
explored through moderator analyses. Table 3 contains effect size estimates for each level of each
moderator and the accompanying moderator analyses.
Modulation vs. initiation of emotional experience. Researchers have long debated whether
facial feedback can only modulate emotional experiences produced by emotional stimuli, versus initiate
emotional experiences in otherwise non-emotional situations (for reviews see Adelmann & Zajonc,
1989; McIntosh, 1996; Soussignan, 2004). Our results suggested that effect sizes are larger in the
absence of emotional stimuli (d = 0.32, 95% CI [0.15, 0.49], p = .0005) than in the presence of
emotional stimuli (d = 0.13, 95% CI [0.07, 0.18], p = .00006), β1 = 0.19, 95% CI [-0.37, -0.01], p = .04,
suggesting that facial movements have larger initiating than modulating effects.
9 Single influential outliers were detected in Flack, Laird, and Cavallaro (1999a), Kalokerinos,
Greenaway, and Denson (2015), and Kircher et al. (2012). Five influential outliers were detected in
McCanne and Anderson (1987).

COLES, LARSEN, AND LENCH 21
Discrete vs. dimensional levels of emotional experience. Facial feedback researchers have
assessed the impact of facial feedback on emotional experience using both discrete emotion measures
(Whissell, 1985) and dimensional measures of positivity/negativity (Winton, 1986). Our results
uncovered no significant evidence of differences in the magnitude of the effects of facial movements on
specific emotions (d = 0.19, 95% CI [0.09, 0.29], p = .0003) versus general positivity/negativity (d =
0.14, 95% CI [0.06, 0.21], p = .0005), β1 = 0.05, 95% CI [-0.07, 0.18], p = .42.
For studies in which discrete emotions were measured, we further assessed whether different
emotions yielded different effect sizes. We found no evidence that specific discrete emotion was a
moderator of facial feedback effects, F(5, 6.42) = 0.77, p = .6010. As shown in Table 3, an examination
of the effect sizes for each specific emotion suggested that facial movements had small-to-medium
effects on self-reports of happiness (d = 0.23, 95% CI [0.08, 0.37], p = .004), sadness (d = 0.30, 95% CI
[0.06, 0.55], p = .02), anger (d = 0.53, 95% CI [0.19, 0.87], p = .006), and disgust (d = 0.29, 95% CI
[0.03, 0.56], p = .03). The effect sizes for fear (d = 0.13, 95% CI [-0.05, 0.30], p = .13) and surprise (d =
-0.31, 95% CI [-5.57, 4.95], p = .59) did not statistically differ from zero; however, these estimates are
based on relatively few effect sizes (kfear = 15; ksurprise = 9).
The effect sizes were small for both positivity (d = 0.18, 95% CI [0.07, 0.28], p = .002) and
negativity (d = 0.12, 95% CI [0.01, 0.22], p = .03), and the magnitude of these effects did not differ, β1 =
0.04, 95% CI [-0.12, 0.19], p = .64.
Awareness of facial feedback manipulation. A prominent debate in the facial feedback
literature concerns the role of participants’ awareness of their posed movements and the emotional
concepts typically associated with these movements (Strack et al., 1988). We found no evidence of
differences in the magnitude of effects in studies that used procedures that limited participants’
awareness of the purpose of the manipulation (d = 0.13, 95% CI [-0.05, 0.31], p = .15) versus studies
that used procedures that did not limit participants’ awareness (d = 0.15, 95% CI [0.06, 0.24], p = .001),
β1 = 0.004, 95% CI [-0.19, 0.19], p = .97.
Awareness of video recording. In reply to Wagenmakers and colleagues’ (2016) failed
replication attempt, Strack (2016) suggested that one reason the results of the original experiment may
not have replicated is that there was a camera directed at participants in the replication study. Across all
studies included in our review, there was very little evidence that this methodological difference is
associated with different facial feedback effects, β1 = -0.06, 95% CI [-0.20, 0.07], p = .36. Facial
feedback effects were small both when participants were aware (d = 0.17, 95% CI [0.06, 0.28], p = .003)
and unaware of video recording (d = 0.23, 95% CI [0.15, 0.32], p = .0000007).
Effects on affective judgments vs. experience. Although the facial feedback hypothesis is
primarily concerned with the effects of facial feedback on emotional experience, many researchers have
extended this phenomenon to examine the effects of facial feedback on affective judgments. Subgroup
analyses suggested that facial movements have a significant effect on both emotional experience (d =
0.17, 95% CI [0.11, 0.23], p = .0000004) and affective judgments (d = 0.38, 95% CI [0.19, 0.57], p = .
0004) (Table 3), and a moderator analysis suggested that the facial feedback effects were larger for
affective judgments than emotional experience, β1 = 0.24, 95% CI [0.04, 0.44], p = .02.
10 We remind the reader that this F-value is based on an Approximate Hotelling-Zhang test with
small sample correction. Even though this analysis has 129 effect sizes, the degrees of freedom are low
because some of the levels of this moderator had a small number of effect sizes in it. See Tanner-Smith,
Tipton, and Polanin (2016) for more information on degrees of freedom).

COLES, LARSEN, AND LENCH 22
Facial feedback manipulation procedure. Determining whether some facial feedback
manipulations have stronger effects than others is complicated by the fact that studies vary in the types
of comparison groups included in experiments. For example, some studies include comparison groups
that receive no facial movement manipulation (e.g., Stel, van den Heuvel, & Smeets, 2008), whereas
others include comparison groups that did receive a facial feedback manipulation (Larsen et al., 1992).
To provide the cleanest test of whether there are differences in effect sizes among facial feedback
manipulations, we limited our analyses to studies featuring a comparison group that received no facial
feedback manipulation11. Effect sizes varied from d = -0.04 (exaggeration-control) to d = 0.71 (Botox-
control), but most manipulation procedures produced small effect sizes (Posing-control, d = 0.30, 95%
CI [-0.16, 0.76], p = .17; Incidental-control, d = 0.13, 95% CI [-0.05, 0.31], p = .15; Suppression-
control, d = 0.15, 95% CI [0.04, 0.25], p = .006). Nevertheless, we did not find evidence that
manipulation procedure was a significant moderator of facial feedback effects, F(4, 10.41) = 0.62, p = .
66 (see Table 3), although small numbers of effects and resulting low power also limit inferences from
these results.
Between vs. within-subjects design. There were early concerns that facial feedback effects may
not emerge in between-subject comparisons (Buck, 1980). Our results indicated that facial feedback
effects emerged both in studies using between-subject (d = 0.16, 95% CI [0.08, 0.24], p = .0001) and
within-subject designs (d = 0.25, 95% CI [0.16, 0.34], p = .000001). Although within-subject designs
tended to be associated with slightly larger effect sizes, the difference was not significant, β1 = 0.09,
95% CI [-0.03, 0.21], p = .14.
Type of stimuli. Facial feedback experiments that include the presentation of emotional stimuli
have used a variety of different stimuli. We found that there were differences in the magnitude of facial
feedback effects based on the type of stimulus used, F(6, 2.77) = 92.83, p = .003 (see Table 3). Most
stimuli produced effect sizes that were small in magnitude (Pictures, d = 0.16, 95% CI [0.08, 0.23], p = .
0002; Films, d = 0.13, 95% CI [0.03, 0.22], p = .009; Stories, d = 0.41, 95% CI [-0.29, 1.10], p = .25;
Social Contexts, d = -0.14, 95% CI [-0.74, 0.46], p = .61), but emotional audio (d = 0.72, 95% CI [-0.82,
2.27], p = .18) and imagined scenarios produced very large effect sizes (d = 1.28, 95% CI [-0.98, 3.53],
p = .27).
Gender. Given gender differences in other emotion effects (Gross & John, 2003; Kring et al.,
1994; LaFrance et al., 2003; McRae et al., 2008; Nolen-Hoeksema & Aldao, 2011) and proposed gender
differences in embodied effects (Pennebaker & Roberts, 1992), we tested whether the proportion of
women in a sample was related to the magnitude of facial feedback effects. Contrary to the proposition
that proprioceptive signals may influence women’s emotional experience less so than men’s, our results
indicated that larger proportions of women tended to have larger effect sizes, but that the association
was not significant, β1 = 0.17, 95% CI [-0.09, 0.42], p = .21.
Timing of measurement. There are inconsistencies regarding whether experimenters collect
self-reports during (d = 0.18, 95% CI [0.09, 0.26], p = .0001) or after the facial feedback manipulation
11 Although we believe that comparing cases where the experiment had a control group that
received no facial feedback manipulation provides the clearest test of whether procedure is a significant
moderator, we also re-ran the analyses including studies that did not include a control group. Similar to
results reported above, effect sizes tended to be small and type of manipulation was not a significant
moderator, F(9, 14.49) = 1.62, p = .20 (see Table 3).

COLES, LARSEN, AND LENCH 23
(d = 0.22, 95% CI [0.12, 0.33], p = .00008). Results provided no evidence that this methodological
difference influences the magnitude of facial feedback effects, β1 = -0.03, 95% CI [-0.17, 0.11], p = .65.
Publication year. Our results provided marginal evidence that effect sizes in the facial feedback
literature tend to become smaller over time of publication (i.e., that effect sizes increase with distance
from 2017), β1 = -0.006, 95% CI [-0.01, 0.001], p = .06. When controlling for publication year, the
overall effect of facial feedback is smaller, but still significant, d = 0.15, 95% CI [0.06, 0.23], t(133) =
3423, p = .0008. However, exploratory follow-up analyses suggest that the relationship between
publication year and observed effect sizes may be driven by the 17 studies included in Wagenmakers et
al.’s (2016) registered replication. When removing these studies, the relationship between publication
year and observed effect sizes is smaller, β1 = -0.002, 95% CI [-0.008, 0.003], p = .45.
Publication status. A common concern in any meta-analysis is that effect sizes in the published
literature are larger than those in the unpublished literature. Twenty-six percent of effect size estimates
in this meta-analysis came from unpublished sources, but the magnitude of effect sizes was not
significantly smaller for unpublished studies (d = 0.15, 95% CI [0.04, 0.26], p = .01) than it was for
published studies (d = 0.21, 95% CI [0.14, 0.28], p = .00000003), β1 = -0.05, 95% CI [-0.18, 0.08], p = .
45. This analysis cannot rule out the possibility that there is a large unpublished literature that is not
represented in the meta-analysis, but it does not support the proposition that uncovering a file-drawer
would change the reported overall effect size.

COLES, LARSEN, AND LENCH 24
Table 3
Moderator analyses
Moderator (bolded) and level s k d β1F95% CI p
Modulation vs. initiation of
emotion
117 246 -- .19 -- [-.37, -.01] .04
Modulation 93 179 0.13 -- -- [0.07, 0.18] .00006
Initiation 28 67 0.32 -- -- [0.15, 0.49] .0005
Discrete vs. dimensional
emotion measure
117 246 -- .05 -- [-.07, .18] .42
Discrete 57 130 0.19 -- -- [0.09, 0.29] .0003
Discrete emotion 56 129 -- -- .77 -- .60
Anger 11 18 0.53 -- -- [0.19, 0.87] .006
Disgust 14 23 0.29 -- -- [0.03, 0.56] .03
Fear 11 15 0.13 -- -- [-0.05, 0.3] .13
Happiness 36 44 0.23 -- -- [0.08, 0.37] .004
Sadness 16 20 0.30 -- -- [0.06, 0.55] .02
Surprise 2 9 -0.31 -- -- [-5.57, 4.95] .59
Dimensional 64 116 0.14 -- -- [0.06, 0.21] .0005
Dimensional emotion 59 109 -- .04 -- [-0.12, 0.19] .64
Positivity 36 57 0.18 -- -- [0.07, 0.28] .002
Negativity 37 52 0.12 -- -- [0.01, 0.22] .03
Awareness of facial
feedback manipulation
81 176 -- .004 -- [-0.19, 0.19] .97
Aware 67 145 0.15 -- -- [0.06, 0.24] .001
Unaware 14 31 0.13 -- -- [-0.05, 0.31] .15
Awareness of video
recording
127 265 -- -.06 -- [-.20, .07] .36
Yes 54 116 0.17 -- -- [0.06, 0.28] .003
No 73 149 0.23 -- -- [0.15, 0.32] .0000007
Emotional experience vs.
affective judgments
138 286 .24 -- [0.04, 0.44] .02
Emotional experience 118 247 0.17 -- -- [0.11, 0.23] .0000004
Affective judgments 24 39 0.38 -- -- [0.19, 0.57] .0004
Facial feedback
manipulation
136 284 -- -- 1.62b-- .20
Botox-control 3 6 0.71 -- -- [-1.07, 2.49] .23
Exaggeration-control 15 29 -0.04 -- -- [-0.41, 0.33] .82
Posing-control 9 20 0.30 -- -- [-0.16, 0.76] .17
Incidental-control 14 31 0.13 -- -- [-0.05, 0.31] .15
Suppression-control 57 96 0.15 -- -- [0.04, 0.25] .006
Posing-posing 14 33 0.51 -- -- [0.26, 0.76] .0007
Posing-suppression 3 5 0.26 -- -- [-0.55, 1.08] .30
Incidental-incidental 10 14 0.43 -- -- [0.22, 0.63] .001
Incidental-suppression 30 43 0.07 -- -- [-0.02, 0.16] .11
Suppression-exaggeration 4 7 0.34 -- -- [-0.68, 1.36] .36
Between vs. within-subjects
design
138 286 -- .09 -- [-.03, .21] .14
Between 80 150 0.16 -- -- [0.08, 0.24] .0001
Within 60 136 0.25 -- -- [0.16, 0.34] .000001
Stimuli 112 217 -- -- 92.83 -- .003
Audio 3 10 0.72 -- -- [-0.82, 2.27] .18
Film 42 94 0.13 -- -- [0.03, 0.22] .009
Imagined Scenariosa1 5 1.28 -- -- [-0.98, 3.53] .27
Pictures 53 84 0.16 -- -- [0.08, 0.23] .0002
Sentencesa2 4 0.70 -- -- [0.43, 0.96] .0000003

COLES, LARSEN, AND LENCH 25
Social context 10 18 -0.14 -- -- [-0.74, 0.46] .61
Storiesa2 2 0.41 -- -- [-.29, 1.10] .25
Proportion of women (0-
100)
122 261 -- .17 -- [-.09, .42] .21
Timing of measurement 113 237 -- -.03 -- [-.17, .11] .65
During manipulation 42 81 0.18 -- -- [0.09, 0.26] .0001
After manipulation 71 156 0.22 -- -- [0.12, 0.33] .00008
Publication year 135 283 -- -.01 -- [-0.01, 0.001] .06
Publication status 138 286 -- -.05 -- [-.18, .08] .45
Unpublished 20 57 0.15 -- -- [0.04, 0.26] .01
Published 118 229 0.21 -- -- [0.14, 0.28] .00000003
Note. k = number of effect size estimates; s = number of studies; d = Cohen’s standardized
difference; β1 coefficients are from separate meta-regressions with RVE where a continuous
moderator was entered in the model as a predictor or a categorical moderator with two levels was
dummy-coded and entered into the model as a predictor; F values are from Approximate Hotelling-
Zhang with small sample correction omnibus tests of the effects of moderators with more than two
levels; 95% C.I corresponds to the β1 coefficient for moderators or d values for individual levels of
moderators; p corresponds to the β1 coefficient or F value for moderators, or t value for individuals
levels of a moderator.
a For cases with too few observations for the RVE approach, we calculated their mean effect size
using a traditional random-effects meta-regression model.
b F-test is comparing all types of methodologies. F-test that compares only studies featuring a true
control condition yielded the following results, F(4, 10.4) = .62, p = .66.
The number of effect size estimates and studies often do not add up as expected because some
studies provided multiple effect size estimates and/or did not provide data for a level of a
moderator.

COLES, LARSEN, AND LENCH 26
Publication Bias. Even though publication status was not a significant moderator of facial
feedback effects, we used two methods to assess potential publication bias more directly.
Publication bias analyses with aggregated dependent effect sizes. First, we used aggregated
dependent effect sizes to examine the funnel plot distribution of effect sizes and perform three statistical
tests of publication bias: trim-and-fill, PET-PEESE, and weight-function modeling.
To visually assess the possibility of publication bias, we first used the aggregated estimates to
create a funnel plot of the effect size estimates and standard errors. In the absence of publication bias,
this pattern should resemble a funnel, where effect size estimates with smaller standard errors cluster
around the mean effect size, and effect size estimates with larger standard errors fan out in both
directions. A typical pattern suggestive of publication bias is asymmetry in the bottom of the
distribution. As can be seen in Figure 2, there was no pattern in the overall funnel plot of the aggregated
effect sizes that was clearly suggestive of publication bias.
To further assess the possibility of publication bias in our overall sample, we conducted three
statistical tests of publication bias. First, we used Duval and Tweedie’s (2000) trim-and-fill technique.
This method trims the values of extreme observations that lead to asymmetry in the funnel plot
distribution and imputes values to even out the distribution. This technique was not able to impute any
missing studies in our data (i.e., did not detect any publication bias). Second, we created PET-PEESE
models (Stanley & Doucouliagos, 2014). PET-PEESE models estimate publication bias by calculating
the relationship between effect size and variability and controlling for this relationship in a meta-
regression model. Both the PET and PEESE models failed to uncover significant evidence of publication
bias, PET β1 = 0.63, p = .16; PEESE β1 = 1.59, p = .1312. Last, we used Vevea and Hedges’ (1995)
weight-function modeling. This method creates a meta-analytic model that is adjusted for publication
bias and compares its fit to an unadjusted model. If an increase in fit is observed, publication bias is a
concern. Results indicated that the model adjusted for publication bias did not increase model fit, which
provides no evidence of publication bias, χ2(1) = 0.14, p = .71.
Publication bias analyses with robust variance estimates. Our second approach for examining
publication bias was to re-examine PET-PEESE with RVE to adjust for dependency instead of
aggregating over dependent effect sizes. Compared to the aggregation method, the benefit of this
approach is that it does not require us to assume a correlation among the clusters of dependent effect
sizes. Contrary to the results produced by the aggregation method, the results of both the PET and
PEESE models with robust variance estimates uncovered significant evidence of publication bias,
PETrve β1 = 1.11, p = .02; PEESErve β1 = 2.32, p = .01. Furthermore, after controlling for this
significant bias, the estimate of the overall effect size did not significantly differ from zero, PETrve d =
-0.03, p = .73; PEESErve d = 0.08, p = .09.
Summary. Different approaches for assessing publication bias in the facial feedback literature
led to different conclusions. When we aggregated the dependent effect sizes, we consistently found no
significant evidence of publication bias. However, when we conducted PET-PEESE analyses with RVE,
we did find evidence of publication bias. Future research will shed light on which approach is superior.
In the meantime, we cannot reject the possibility of publication bias in the overall facial feedback
literature.
12 We pre-registered r = .5 as our assumed correlation among effect sizes in the aggregation of
dependent effect sizes. When we performed sensitivity analyses on this assumed correlation, we did find
evidence of publication bias in our PET-PEESE models when r = .9.

COLES, LARSEN, AND LENCH 27
Publication bias sensitivity analyses. As noted above, there is a large degree of heterogeneity
in the overall size of facial feedback effects, T2 = 0.11, I2 = 75.6. This heterogeneity can pose problems
for many tests of publication bias (Stanley, 2017; Sterne et al., 2011; Terrin et al., 2003), and suggests
that it may be more fruitful to examine publication bias on individual levels of significant moderators.
We found three significant moderators in our meta-analysis: 1) type of affective reaction (emotional
experience or affective judgments), 2) whether facial feedback initiates or modulates emotional
experience, and 3) the type of stimuli used in the experiment. In line with our pre-registration plan, we
re-ran all publication bias analyses on individual levels of these significant moderators. We found no
evidence of publication bias when we split our analyses by the initiation vs. modulation or stimulus type
moderator but did find evidence of publication bias when we split our analyses by type of affective
reaction.
Publication bias in studies examining affective judgments. Publication bias sensitivity analyses
revealed evidence of publication bias in studies that examined the effects of facial feedback on affective
judgments. As shown in the left panel of Figure 3, the funnel plot is largely asymmetrical. The trim-and-
fill method imputed 5 missing observations but suggested that the adjusted overall effect was still
significant (adjusted d = 0.25, 95% CI [0.06, 0.44], p = .01). The PET and PEESE models both
suggested that publication bias was present (PET β1 = 2.65, p = .03; PEESE β1 = 5.05, p = .048; PETrve
β1 = 2.28, p = .01; PEESErve β1 = 3.41, p = .04) and that the bias-corrected overall effect is not
significant (PET d = -0.22, p = .36; PEESE d = 0.08, p = .52; PETrve d = -0.17, p = .49; PEESErve d =
0.16, p = .28). The weight-function model also provided marginal evidence that publication bias was a
concern, (χ2(1) = 3.17, p = .07) and suggested that the bias-corrected overall effect is not significant
(adjusted d = 0.18, p = 0.18). This suggest that, when controlling for publication bias, the cumulative
evidence does not support the notion that facial feedback influences affective judgments.
Publication bias in studies examining emotional experience. When we examined the effects of
facial feedback on emotional experience, we consistently found no evidence of publication bias. As
shown in the right panel of Figure 3, the funnel plot of effect sizes appeared symmetrical. Furthermore,
the trim-and-fill method imputed no missing studies, PET-PEESE estimates of publication bias were not
significant (PET β1 = 0.14, p = .77; PEESE β1 = 0.46, p = .69; PETrve β1 = 0.70, p = .20; PEESErve β1 =
1.75, p = .13), and weight-function modeling found that the meta-analytic model that is adjusted for
publication bias did not provide better fit than a non-adjusted model, χ2(1) = 1.14, p = .29. Since we did
not find evidence of publication bias in studies that examined the effects of facial feedback on emotional
experience, we performed additional pre-planned sensitivity analyses. More specifically, we re-ran the
publication bias tests: 1) excluding suppression studies, 2) excluding Wagenmakers et al. (2016), and 3)
excluding Wagenmakers et al. (2016) and all unpublished data. None of these sensitivity analyses
suggested the presence of publication bias in studies that examined the effects of facial feedback on
emotional experience. This suggests that the cumulative evidence supports the assertion that facial
feedback influences emotional experience, the central tenet of the facial feedback hypothesis.

COLES, LARSEN, AND LENCH 28
Figure 2
Overall funnel plot for studies examining the impact of facial expressions on emotional experience and affective
judgments.
Figure 3
Funnel plots for studies examining the effect of facial feedback on emotional experience and the effect of facial
feedback on affective judgments.

COLES, LARSEN, AND LENCH 29
Discussion
Lay people and scientists alike have long wondered if feedback from our facial movements can
influence our experience of emotion. The combined results from nearly 300 effect sizes generated from
138 studies suggest that facial feedback can indeed influence emotional experience, although these
effects tend to be small and heterogenous. Importantly, based on the results of a variety of publication
bias analyses, the effects of facial feedback on emotional experience (but not affective judgments) do not
appear to be driven by publication bias
Addressing Disagreements in the Facial Feedback Hypothesis Literature
The results of this meta-analysis support the general claim that facial feedback influences
emotional experience. However, facial feedback theorists have typically disagreed not about whether
these effects exist, but rather the specific contexts in which one can expect to observe these effects.
Next, we consider the implications of our results for the major theoretical disagreements in the facial
feedback literature.
Facial feedback can initiate and modulate emotional experience. When James and Lange
proposed that bodily perturbations both initiated and modulated emotional experiences over 100 years
ago, they were met with a great deal of incredulity. While many critics conceded that bodily changes
could perhaps modulate emotional experiences, they often rejected the notion that these bodily states
were sufficient in creating experiences of emotion (Cannon, 1927; Irons, 1894; Sherrington, 1900;
Worcester, 1893). Lange speculated that these initiating effects could actually be demonstrated quite
easily. Since Lange believed that emotional experience was built entirely upon sensed changes in the
autonomic nervous system, he suggested that any substance that influenced this system (e.g., alcohol)
had the potential to initiate an emotional experience, even in otherwise non-emotional situations. James
agreed that initiation effects were theoretically possible. However, he did not agree that producing such
effects would be easy, contending that it would require a coordinated set of responses across the entire
body. Despite their disagreements, one thing that James, Lange, and their critics would have likely
agreed upon is the prediction that facial feedback, by itself, could not initiate emotional experience.
Consequently, our finding that facial feedback can both modulate as well as initiate emotional
experiences is quite remarkable.
Although surprising from a historical perspective, most theories in the facial feedback literature
are consistent with the observation that facial feedback can initiate emotional experiences (Berkowitz,
1990; Ekman, 1979; Guenther, 1981; Izard, 1977; Laird, 1974; Laird & Bresler, 1992; Tomkins, 1962).
For example, Ekman (1979) suggests that each discrete emotion is activated by a biologically-innate
affect program that produces a set of bodily responses that merge in consciousness to form emotional
experience. While these affect programs are often activated by external stimuli, Levenson and
colleagues suggested that they can also be activated by facial movements (Levenson et al., 1990).
Nevertheless, although many facial feedback theories are consistent with the observed initiation effects,
theorists have typically speculated that such effects would be difficult to obtain. For example, Tomkins
(1981) suggested that facial movements can only initiate emotional responses if they match the intensity,
duration, and configuration of naturally occurring emotional expressions. These exact specifications are

COLES, LARSEN, AND LENCH 30
rarely adhered to in experiments on the facial feedback hypothesis (Matsumoto, 1987; Soussignan,
2002). Consequently, our results suggest that initiating facial feedback effects may be easier to obtain
than researchers have previously believed.
Although consistent with most facial feedback theories, the observed initiating facial feedback
effect is inconsistent with Allport’s pioneering theory of facial feedback. Allport (1922, 1924) believed
that the autonomic nervous system created undifferentiated feelings of positivity and negativity that
were differentiated into discrete emotional categories based on patterns of facial feedback. According to
this view, facial feedback cannot initiate emotional experiences in the absence of ongoing feelings of
positivity and negativity. Assuming that participants in facial feedback experiments are not incidentally
experiencing strong feelings of positivity or negativity, the observed initiating facial feedback effects are
inconsistent with Allport’s theory.
In addition to contending that facial feedback cannot initiate emotional experiences, Allport
suggested that facial feedback could only influence discrete, but not dimensional, levels of emotion.
Next, we review results that disconfirm this prediction.
Facial feedback can influence discrete and dimensional reports of emotion. Facial feedback
theorists like Allport have tended to emphasize the effects of facial feedback on discrete emotions
(Berkowitz, 1990; Guenther, 1981; Izard, 1977; Tomkins, 1962), although later work raised the
possibility that facial feedback may also influence dimensional reports of emotion (Zajonc, 1985; Zajonc
et al., 1989). Given facial feedback theorists’ interest in discrete emotions, it is notable that previous
reviews have described these effects as non-existent (Winton, 1986), preliminary (Adelmann & Zajonc,
1989), mixed (McIntosh, 1996), and controversial (Soussignan, 2004). Our results suggest that facial
feedback can influence both discrete and dimensional reports of emotion13, and we uncovered little
evidence that facial feedback effects are larger for one than the other.
To date, facial feedback theorists have not typically considered whether facial feedback effects
might be larger for some discrete emotions than others. However, in a recent narrative review of a
theoretical model of surprise, Reisenzein, Horstmann, and Schützwohl (2017) noted that there was
mixed evidence for the effects of facial feedback on the experience of surprise. Furthermore, they
suggested that if these facial feedback effects do exist, they “cannot play a prominent role for the
experience of surprise” (p. 16). Our results indicated that facial feedback effects do not significantly
differ based on the type of discrete emotion measured. However, consistent with Reisenzein et al.’s
(2017) assertions, we failed to observe significant facial feedback effects in the subset of studies
examining surprise. In fact, the overall effect for these studies were in the opposite direction predicted
by the facial feedback hypothesis. In addition, we did not observe a significant facial feedback effect in
studies that examined feelings of fear. Although these results may suggest that facial feedback does not
influence the experience of all discrete emotions, we currently caution against this conclusion; type of
13 We found evidence that facial feedback influences both discrete and dimensional levels of
emotion. However, it is nevertheless possible that facial feedback only directly influences one of these
levels of emotion and these effects indirectly influence reports of the other level. For example, perhaps
smiling makes people feel more happy but not more positive, but people report higher levels of
positivity because they are experiencing higher levels of happiness. Nevertheless, such a speculation
seems difficult to experimentally confirm.

COLES, LARSEN, AND LENCH 31
emotion was not a significant moderator in this meta-analysis, and there is still only a handful of studies
that have examined fear and surprise facial feedback effects.
The role of awareness. Strack et al.’s (1988) pen-in-mouth paper is the most well-known
demonstration of the facial feedback hypothesis not just because of the elegance of their manipulation,
but also because the work is cited as evidence that the effects of facial feedback on emotional experience
are not driven by demand characteristics. In addition to addressing this major methodological concern,
their work is often considered to have provided evidence that facial feedback effects can occur outside
of people’s awareness. However, a large failure-to-replicate has created uncertainty regarding the
reliability of the pen-in-mouth effect (Wagenmakers et al., 2016; but see Noah et al., 2018, Strack,
2016). Although a failure-to-replicate 2% of the experimental evidence for the facial feedback
hypothesis does not invalidate the overall claim that facial movements influence emotional experience, it
has revived concerns that these effects are driven by demand characteristics and reopened the discussion
about the mechanism that underlies these effects.
The cumulative evidence suggests that studies that use procedures that limit participants’
awareness of the purpose facial feedback manipulation produce similar effect sizes as studies that do
not. Notably, these analyses included all types of incidental facial movement manipulations (i.e., were
not limited to the pen-in-mouth manipulation). These results suggest that the effects of facial feedback
on emotional experience are not necessarily driven by demand characteristics, although this does not
preclude the possibility that they sometimes are (e.g., when experimenters do not effectively mask the
purpose of their experiment). These results would seem to be inconsistent with theories that predict that
such effects are mediated by self-perception mechanisms (Laird, 1984; Laird & Bresler, 1992).
However, Laird later argued that the self-perception process did not necessarily require awareness of the
facial movements (e.g., moving the corners of one’s mouth into a smile) or the purpose of these
movements (e.g., to smile; Laird & Bresler, 1992). Consequently, although our results fail to confirm
that awareness of the purpose of the facial feedback experiment is necessary for facial feedback effects
to emerge, it does not necessarily disconfirm Laird’s self-perception theory of emotion.
Do facial movements influence affective judgments? The central tenet of the facial feedback
hypothesis is that facial feedback influences emotional experience. However, many researchers in the
facial feedback literature have expanded upon this original scope by suggesting that facial feedback can
also influence affective judgments (Davis et al., 2015; Dzokoto et al., 2014; Ohira & Kurono, 1993), a
term we have used to broadly refer to judgments about the emotional characteristics of a stimulus.
Results initially indicated that facial feedback does influence affective judgments and that facial
feedback effects are larger for affective judgments than emotional experience. However, we
subsequently uncovered consistent evidence of publication bias in this subset of studies. Depending on
the method for generating bias-corrected overall effect size estimates, the adjusted overall effect size
was either close to zero or in the opposite direction. Regardless, the bias-corrected overall effect size
estimates did not significantly differ from zero.
Although the current balance of evidence does not support the assertion that facial feedback
influences affective judgments, we strongly caution against prematurely abandoning research on these
effects. Researchers who have examined the effects of emotional states on subsequent judgments have
often emphasized that such effects do not emerge in all contexts (Clore, Schiller, & Shaked, 2018;
Schwarz & Clore, 2007). For example, Schwarz and Clore (2007) suggest that feelings only influence
judgments when they seem relevant to the task at hand. Based on this view, facial feedback will only

COLES, LARSEN, AND LENCH 32
influence affective judgments when the elicited emotional experiences are perceived to be relevant to the
target being evaluated. Interestingly, these context-dependent effects are also predicted by Laird’s self-
perception theory of emotion (Laird, 1984; Laird & Bresler, 1992), but this prediction has gotten little
attention in the facial feedback literature.
Although emotional experience and affective judgments are considered distinct in the emotion
literature, a clear operational distinction between the two remains elusive. Consider theoretical debates
about whether people can experience simultaneously mixed emotions of happiness and sadness (Larsen,
McGraw, & Cacioppo, 2001; Russell & Carroll, 1999). Russell (2017) has pointed out all that
researchers who have ostensibly observed mixed emotions (for a meta-analytic review, see Berrios,
Totterdell, & Kellett, 2015) might have inadvertently measured affective judgments rather than
emotional experience (see also Larsen, 2017). In the facial feedback literature, Strack et al. (1988)
measured affective judgments by asking participants “How funny do you think these cartoons are?”.
This dependent measure can be considered an affective judgment since it is a question about the stimuli,
not felt experience. However, it is plausible that many participants interpreted it as a question about their
experience of amusement. Future research can more clearly assess the relationship between facial
movements and affective judgments by using measures that more clearly isolate affective judgments
from emotional experience (e.g., Hunter et al., 2010; Itkes, Kimchi, Kron, & Carmel, 2017). Such
measures may help clarify whether facial feedback influences affective judgments. In any event, our
observation that facial feedback effects can occur in otherwise non-emotional situations suggests that
effects of facial feedback on emotional experience need not be mediated by affective judgments. In
summary, the current balance of evidence does not support the assertion that facial feedback influences
affective judgments, but we caution against abandoning this line of research.
Implications for Other Emotion Theories
The primary goal of this meta-analysis was to address disagreements among emotion theorists
who have made explicit predictions about the impact of facial feedback on emotional experience.
However, most of these theories fall into two categories: 1) basic emotion theories, which postulate the
existence of a finite set of biological affect programs that elicit coordinated sets of emotion-specific
responses (Allport, 1922; Ekman, 1979; Izard, 1971; Tomkins, 1962), or 2) network theories of emotion,
which postulate association-based cognitive organizations of emotion concepts (Berkowitz, 1990;
Guenther, 1981). Interestingly, facial feedback effects are less frequently discussed in the context of
contemporary appraisal and constructionist theories of emotion despite the fact that these effects are not
generally inconsistent with these theories. Next, we will briefly consider our results in the context of
appraisal and constructionist emotion theories, focusing on broad implications as opposed to nuanced
distinctions among theories within each tradition.
Appraisal theories of emotion. A fundamental assumption of appraisal theories of emotion is
that automatic or controlled cognitive appraisals are the antecedents of emotional reactions (Moors,
Ellsworth, Scherer, & Frijda, 2013; Roseman & Smith, 2001). According to these views, cognitive
appraisals produce a set of action tendencies, physiological responses, and motor behaviors, all of which
contribute to the experience of emotion. To the degree that the effects of appraisals on emotional
experience are mediated by motor behaviors (Scherer, 2009), appraisal theories would expect facial
feedback to influence emotional experience. However, given that appraisal theories argue that cognitive

COLES, LARSEN, AND LENCH 33
appraisals are the antecedents of emotional reactions, these theories have more difficulty reconciling
their views with the observation that facial movements can initiate emotional experiences in the absence
of emotional stimuli.
From one perspective, facial feedback effects might simply represent exceptions to a rule that do
not characterize typical emotional experiences (Ellsworth & Scherer, 2003; Roseman & Smith, 2001).
On the other hand, Berkowitz and Harmon-Jones (2004) have argued that, “...a truly comprehensive
theory of affective states should attempt to deal with relatively unusual occurrences as well as the more
common ones” (p. 125). To that end, appraisal theorists Smith and Kirby (2004) have suggested that
facial feedback can initiate an emotional experience if it activates the emotion’s corresponding appraisal
pattern via associative processing. Two of our findings suggest otherwise. First, it only makes sense to
suggest that facial feedback has initiated an emotional reaction if no emotional stimulus is present. In the
event, there is nothing to engage the appraisal process. Second, our results thus far have failed to provide
evidence that facial feedback influences affective judgments, which are conceptually distinct but similar
to cognitive appraisals. Nevertheless, a more direct test of this assertion would ultimately be more
informative.
Psychological constructionist theories of emotion. Modern psychological constructionist
theories of emotion postulate that the experience of discrete emotions represent the outcome of a mental
categorization process (Barrett, Wilson-Mendenhall, & Barsalou, 2014; Lindquist, 2013; Russell, 2014).
Central to these models is the concept of core affect, which represents “the most elementary consciously
accessible affective feelings” that people can experience (Russell & Barrett, 1999, p. 806). Core affect is
thought to vary along a bipolar valence dimension and a unipolar activation or arousal dimension
ranging from states of low to high arousal. According to these models, core affect is ever-present (at
least when we are awake or dreaming) but emotions only occur occasionally. Specifically, people
experience what we typically refer to “emotions” when they categorize their core affect into a discrete
emotional category (e.g., anger, fear) based on physiological states, conceptual knowledge about
emotions, and situational cues. For example, the discrete emotion that people will experience in a high-
arousal unpleasant state will depend on whether the situational cues more closely resemble their
prototype of fear, anger, or some other emotion. Constructionist theories of emotion areoften contrasted
with the basic emotion theoretical tradition, which views emotional experience as a byproduct of a
coordinated set of responses elicited by the activation of biologically hardwired affect programs.
Although the facial feedback hypothesis has traditionally been most closely associated with basic
emotion theories, modern psychological constructionist theories of emotion provide a framework for
exploring two different facial feedback effects: (a) the effects of facial feedback on core affect, and (b)
the effects of facial feedback on the mental categorization of core affect.
Core affect has been described as a “neurophysiological barometer of the individual’s relation to
an environment at a given point in time” (Barrett, 2006, p. 31; Barrett & Bliss-Moreau, 2009; Duncan &
Barrett, 2007). Researchers have tended to focus on the effects of interoceptive feedback on core affect
(MacCormack & Lindquist, 2017, 2018), but have also noted that proprioceptive feedback can influence
core affect (Barrett & Bliss-Moreau, 2009; Lindquist, 2013). Our observation that facial feedback
influences dimensional reports of emotion suggests that facial feedback may be one of type of
proprioceptive feedback that contributes to core affect.

COLES, LARSEN, AND LENCH 34
From a constructionist perspective, a second possibility is that facial feedback can influence
whether and how core affect is categorized into discrete emotions. For instance, people who are in
unpleasant but otherwise ambiguous situations may be more likely to categorize their unpleasant core
affect as anger if they have been induced to scowl as opposed to frown. This idea echoes Allport’s
(1922, 1924) contention that facial feedback guides the categorization of underlying valanced feelings.
However, whereas Allport suggested that the patterns of facial movements that guide the categorization
process are biologically innate, psychological constructionist theories would argue that these effects are
driven by learned associations between patterns of facial movements and emotional concepts. In other
words, constructionist theories of emotion would predict that the effects of smiling on the categorization
of positive affect as happiness, for example, may be mediated by the extent to which an individual
believes smiling is a symptom of happiness. It is worth noting that even though Allport proposed that
facial feedback can influence emotion categorization nearly a century ago, this hypothesis remains
untested (McIntosh, 1996).
Other Potential Sources of Heterogeneity
In addition to examining moderators that provided insight into theoretical disagreements in the
emotion literature, this meta-analysis examined several other methodological moderators proposed by
previous facial feedback researchers, including whether effect sizes came from between- or within-
subject comparisons (Buck, 1980), the procedure used to manipulate facial poses (Izard, 1990a), gender
(Pennebaker & Roberts, 1992), and whether participants were aware of video recording (Strack, 2016).
In addition, we tested methodological moderators we thought might influence facial feedback effects,
such as the timing of self-reported affective experience. We did not uncover significant evidence that
these factors were associated with differences in the magnitude of facial feedback effects. We did,
however, find evidence that facial feedback effects were larger in the presence of some types of stimuli
(e.g., emotional sentences) than others (e.g., pictures; see Table 3). Nevertheless, there are large amounts
of heterogeneity within different stimulus types, suggesting that even within a group of studies using
similar types of stimuli (e.g., pictures), other methodological choices (e.g., different pictures; different
presentation modes) may affect the magnitude of facial feedback effects.
Although we examined moderators that figured prominently in the facial feedback literature,
given the large degree of heterogeneity in facial feedback effects, we believe that there are potential
moderators that we did not evaluate. For example, Laird and colleagues argued that individual
differences in the degree to which individuals attend to their bodily cues—including but not limited to
proprioceptive cues from the face—is a key moderator of facial feedback effects (Laird & Bresler, 1992;
Laird & Crosby, 1974; Laird & Lacasse, 2014). Unfortunately, we were not able to assess this moderator
because we cannot assess how different experimental procedures influenced the degree to which
participants attended to their bodily cues. Furthermore, Laird and colleagues’ own work sheds little light
on this question because they often used circular reasoning, classifying only participants who
demonstrated larger facial feedback effects as individuals who attend more to bodily cues.
Future research can investigate the role of individual differences in proprioceptive awareness
using both self-reports and behavioral measures. For example, Mehling and colleagues (2012) have
developed a self-report measure that assesses individual differences in the degree to which people

COLES, LARSEN, AND LENCH 35
believe they attend to interoceptive cues and use these cues to make sense of their emotions. Scales like
these could potentially be adapted for research on proprioceptive awareness. In addition, researchers can
use behavioral measures of proprioceptive awareness of facial expressions (e.g., Leplow, Schluter, &
Ferstl, 1992). Furthermore, it is likely that methods for measuring proprioceptive awareness of other
bodily regions can be adapted to the study of facial feedback (for a review, see Hillier, Immink, &
Thewlis, 2015).
Exclusion criteria may be an important source of heterogeneity in facial feedback research
because researchers use varying sets of exclusion criteria to sometimes exclude large proportions of
participants. Approximately half of the studies in our review did not report any exclusion criteria, and
those that did used a variety of criteria. For example, researchers sometimes excluded participants who
were aware of the purpose of the experiment (e.g., Baumeister, Papa, & Foroni, 2016; Duncan & Laird,
1977; Laird, 1974), failed an attention-check (e.g., Kalokerinos, Greenaway, & Denson, 2015),
experienced equipment errors (e.g., Pedder et al., 2016), produced unreadable or missing data (e.g.,
Dzokoto et al., 2014; Zajonc et al., 1989), or were outliers (e.g., Korb, Grandjean, Samson, Delplanque,
& Scherer, 2012; Marmolejo-Ramos & Dunn, 2013; Zhu, Cai, Sun, & Yang-yang, 2015). Exclusion
criteria choice might be especially important in the facial feedback literature given the large proportions
of participants that are sometimes excluded. For example, Soussignan (2002) excluded approximately
30% of participants because they did not contract the correct facial muscles. Wagenmakers et al. (2016)
used a combination of several exclusion criteria and, on average, excluded 25% of their participants.
These various exclusion criteria have the potential to both deflate effect sizes (e.g., excluding
participants who exhibit demand characteristics would presumably lower the effect size) and inflate
effect sizes (e.g., excluding participants who failed to smile would presumably increase the effect size),
which further contributes to heterogeneity in the facial feedback literature.
Limitations of the Meta-Analytic Approach
This meta-analysis provides the most comprehensive integrative review of the facial feedback
hypothesis to date. However, it would be a mistake to interpret the comprehensive nature of this work as
providing authoritative conclusions about facial feedback effects. Although meta-analysis is a valuable
tool, it possesses a variety of limitations. Next, we will discuss some of the most pressing limitations of
this meta-analytic work.
Meta-analytic conclusions can be compromised by the presence of questionable research
practices (QRPs). To date, meta-analysts have been primarily interested in the effects of publication bias,
and researchers have subsequently developed several tests of the extent and impact of this bias.
However, methods for detecting publication bias are imperfect. Publication bias detection methods have
suboptimal statistical properties in a variety of scenarios (Carter, Schönbrodt, Hilgard, & Gervais, 2017;
Macaskill, Walter, & Irwig, 2001; Stanley, 2017) and were developed and tested under the assumption
that the underlying effect sizes are independent. Over half (53%) of our studies provided multiple effect
sizes, and different approaches for dealing with such dependencies led to slightly different conclusions
regarding publication bias in the overall facial feedback literature. Fortunately, more clear patterns
emerged in our sensitivity analyses, where all approaches produced evidence of publication bias in
studies examining affective judgments and a lack of evidence of publication bias in studies examining

COLES, LARSEN, AND LENCH 36
emotional experience. Nevertheless, we believe future research should continue to develop and validate
methods for detecting publication bias and evaluate the effectiveness of these approaches when
dependent data structures exist.
Other QRPs, such as optimal stopping, p-hacking, and infrequent cases of outright fraud, also
threaten the validity of meta-analytic conclusions. John, Loewenstein, and Prelec (2012) found that a
high proportion of psychology researchers admitted to performing these practices, including: deciding
whether to exclude data after looking at the impact of doing so on the results (43%), deciding whether to
continue data collection after looking to see whether the results were significant (58%) and stopping
data collection early once significant results have been found (23%). These practices inflate meta-
analytic estimates, which can create misleading conclusions (Bierman, Spottiswoode, & Bijl, 2016;
Head, Holman, Lanfear, Kahn, & Jennions, 2015). Although some newer methods for detecting bias—
such as p-curve (Simonsohn, Nelson, & Simmons, 2014) and the incredibility index (Schimmack, 2012)
—may help identify the existence of other QRP’s, these methods also currently assume that effect sizes
are independent. Therefore, it is currently unclear to what degree QRP’s may have inflated the effect
sizes we observed in this meta-analysis.
Last, despite the large size of the facial feedback literature, it is likely that many of our
moderator analyses lack adequate statistical power. Moderator analyses typically need a large amount of
observations to achieve high power (Hedges & Pigott, 2004), and the power to detect moderators is
reduced by higher levels of heterogeneity and robust variance estimation procedures. Given the high
level of heterogeneity in our meta-analysis, it is quite possible that future researchers can devise more
powerful tests of moderation by manipulating a moderating factor in an experiment. Consequently, null
effects in our tests of moderation should be cautiously interpreted, and future research should continue
to consider the impact of these potential moderators.
Conclusion
When Thích Nh t H nh stated that “sometimes your smile can be the source of your joy,” he ấ ạ
may not have been aware that what he apparently took to be a settled fact had a long, contentious history
in psychological science. Indeed, it has been over 30 years ago since the “facial feedback hypothesis”
fragmented into a variety of “facial feedback hypotheses” (Adelmann & Zajonc, 1989; McIntosh, 1996;
Tourangeau & Ellsworth, 1979). In retrospect, such fragmentation helped clarify unresolved theoretical
disagreements and facilitated more nuanced discussions about when and why facial feedback effects
emerge. Subsequent primary research studies have gone only some way toward resolving these
disagreements, in part due to discrepant findings (e.g., Strack et al., 1988; Wagenmakers et al., 2016).
We believe our meta-analysis has resolved many of these theoretical disagreements. Based on a review
of over 100 years of research, 138 studies, and 286 effect sizes, our understanding of the effects of facial
feedback on emotional experience is becoming more clear. The cumulative evidence, to date, suggests
that facial feedback does indeed influence emotional experience. Facial feedback appears to influence
undifferentiated feelings of positivity, negativity, and a variety of discrete emotions (e.g., happiness,
anger, disgust). However, so far the evidence does not suggest that facial feedback influences all
emotions (e.g., fear and surprise). Interestingly, it appears that facial feedback effects are largest in
otherwise non-emotional situations, which not only indicates that facial feedback is sufficient for the

COLES, LARSEN, AND LENCH 37
experience of emotions but also suggests that this may be the most powerful context to examine these
effects.
The nature of scientific inference prevents us from concluding that “your smile can be the source
of your joy” with anywhere near the confidence that Thích Nh t H nh could. Besides, Thích Nh t ấ ạ ấ
H nh’s concept of ạjoy is probably a rare commodity in most psychology laboratories. Nonetheless, a
half century’s worth of experimental findings does provide considerable evidence that smiles, frowns,
scowls, and other facial movements can affect emotional experience in a variety of scenarios. At the
same time, our meta-analysis indicates that the effects are quite small and appear to vary for reasons that
our meta-analysis did not shed light on. Having demonstrated that facial feedback effects can occur, we
hope that future research sheds further light on why they do.

COLES, LARSEN, AND LENCH 38
References
References with asterisks were included in the meta-analysis synthesis.
Adelmann, P. K., & Zajonc, R. B. (1989). Facial efference and the experience of emotion. Annual
Review of Psychology, 40(1), 249–480.
Allport, F. H. (1922). A physiological-genetic theory of feeling and emotion. Psychological Review,
29(2), 132–139.
Allport, F. H. (1924). Feeling and emotion. In F. H. Allport (Ed.), Social Psychology (pp. 84–98).
Boston: Houghton Mifflin Company.
*Andréasson, P. (2010). Emotional empathy, facial reactions, and facial feedback. (Doctoral
Dissertation). Uppsala University, Sweden.
*Andréasson, P., & Dimberg, U. (2008). Emotional empathy and facial feedback. Journal of Nonverbal
Behavior, 32(4), 215–224.
Angell, J. R. (1916). A reconsideration of James’s theory of emotion in light of recent criticisms.
Psychological Review, 23(4), 251–261.
Ansfield, M. E. (2007). Smiling when distressed: When a smile is a frown turned upside down.
Personality and Social Psychology Bulletin, 33(6), 763-775.
Barrett, L. F. (2006). Solving the emotion paradox: Categorization and the experience of emotin.
Personality and Social Psychology Review, 10(1), 20-46.
Barrett, L. F., & Bliss-Moreau, E. (2009). Affect as a psychological primative. Advances in
Experimental Social Psychology, 41(8), 167–218.
Barrett, L. F., Wilson-Mendenhall, C. D., & Barsalou, L. W. (2014). The conceptual act theory: A
roadmap. In L. F. Barrett & J. A. Russell (Eds.), The Psychological Construction of Emotion (pp.
83–110). New York. Retrieved from http://www.affective-science.org/pubs/2013/conceptual-act-
precis.pdf%5Cnhttp://emr.sagepub.com/cgi/doi/10.1177/1754073914534479
*Baumeister, J. C., Papa, G., & Foroni, F. (2016). Deeper than skin deep - The effect of botulinum toxin-
A on emotion processing. Toxicon, 118, 86–90.
Berkowitz, L. (1990). On the formation and regulation of anger and aggression: A cognitive-
neoassociationistic analysis. American Psychologist, 45(4), 494–503.
Berkowitz, L., & Harmon-Jones, E. (2004). Toward an understanding of the determinants of anger.
Emotion, 4(2), 107–130.
Berrios, R., Totterdell, P., & Kellett, S. (2015). Eliciting mixed emotions: A meta-analysis comparing
models, types and measures. Frontiers in Psychology, 6(MAR), 1–15.
Bierman, D. J., Spottiswoode, J. P., & Bijl, A. (2016). Testing for questionable research practices in a
meta-analysis: An example from experimental parapsychology. PloS One, 11(5), e0153049.
Retrieved from http://journals.plos.org/plosone/article/file?
id=10.1371/journal.pone.0153049&type=printable
*Bodenhausen, G. V., Kramer, G. P., & Süsser, K. (1994). Happiness and stereotypic thinking in social
judgment. Journal of Personality and Social Psychology, 66(4), 621–632.
Borenstein, M. (2009). Effect sizes for continuous data. In H. Cooper, L. V. Hedges, & J. C. Valentine
(Eds.), The Handbook of Research Synthesis and Meta-Analysis (pp. 221–235). New York: Russel

COLES, LARSEN, AND LENCH 39
Sage Foundation.
Borenstein, M., Hedges, L. V., Higgins, J. P., & Rothstein, H. R. (2009). Introduction to Meta-Analysis.
New Jersey: John Wiley & Sons.
Buck, R. (1980). Nonverbal behavior and the theory of emotion: The facial feedback hypothesis.
Journal of Personality and Social Psychology, 38(5), 811–824.
Bull, N. (1945). Towards a clarification of the concept of emotion. Psychosomatic Medicine, 7(4), 210–
214. Retrieved from papers3://publication/uuid/838615A4-76BD-4EFE-9BB7-6EAAAAC68C6F
Bull, N. (1946). Attitudes: Conscious and Unconscious. The Journal of Nervous and Mental Disease,
103(4), 337–345. Retrieved from papers3://publication/uuid/838615A4-76BD-4EFE-9BB7-
6EAAAAC68C6F
*Bush, L. K., Barr, C. L., McHugo, G. J., & Lanzetta, J. T. (1989). The effects of facial control and
facial mimicry on subjective reactions to comedy routines. Motivation and Emotion, 13(1), 31-52.
*Butler, E. A., Egloff, B., Wilhelm, F. H., Smith, N. C., Erickson, E. A., & Gross, J. J. (2003). The social
consequences of expressive suppression. Emotion, 3(1), 48–67. Retrieved from
http://doi.apa.org/getdoi.cfm?doi=10.1037/1528-3542.3.1.48
*Butler, E. A., Wilhelm, F. H., & Gross, J. J. (2006). Respiratory sinus arrhythmia, emotion, and
emotion regulation during social interaction. Psychophysiology, 43(6), 612–622.
*Cai, A., Lou, Y., Long, Q., & Yuan, J. (2016). The sex differences in regulating unpleasant emotion by
expressive suppression: Extraversion matters. Frontiers in Psychology, 7(1011), 1–11.
Cannon, W. (1915). Bodily Changes in Pain, Hunger, Fear and Rage. New York: D Appleton and
Company.
Cannon, W. (1927). The James-Lange theory of emotions : A critical examination and an alternative
theory. American Journal of Psychology, 39(3), 106–124. Retrieved from
http://www.jstor.org/stable/1415404
Carter, E. C., Schönbrodt, F. D., Hilgard, J., & Gervais, W. M. (2017). Correcting for bias in psychology:
A comparison of meta-analytic methods. Retrieved from http://osf.io/preprints/psyarxiv/9h3nu
*Ceschi, G., & Scherer, K. (2003). Children's ability to control the facial expression of laughter and
smiling: Knowledge and behaviour. Cognition & Emotion, 17(3), 385–411.
*Clapp, J. D. (2012). Expressive inhibition following trauma: Implications for understanding
posttraumatic numbing. (Doctoral Dissertation). Retrieved from ProQuest Dissertations & Theses
Global. (3541095)
Clore, G. L., Schiller, A. J., & Shaked, A. (2018). Affect and cognition: Three principles. Current
Opinion in Behavioral Sciences, 19, 78–82. Retrieved from
http://dx.doi.org/10.1016/j.cobeha.2017.11.010
Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences (Vol. 2nd). New York:
Lawrence Erlbaum Associates.
*Davey, G. C., Sired, R., Jones, S., Meeten, F., & Dash, S. R. (2013). The role of facial feedback in the
modulation of clinically-relevant ambiguity resolution. Cognitive Therapy and Research, 37(2),
284–295.
*Davis, J. D., Winkielman, P., & Coulson, S. (2015). Facial action and emotional language: ERP
evidence that blocking facial feedback selectively impairs sentence comprehension. Journal of
Cognitive Neuroscience, 27(11), 2269–2280.

COLES, LARSEN, AND LENCH 40
*Davis, J. I. (2008). The connection between facial expression and emotional experience. (Doctoral
Dissertation). Retrieved from ProQuest Dissertations & Theses Global. (3305213)
*Davis, J. I., Senghas, A., Brandt, F., & Ochsner, K. N. (2010). The effects of BOTOX injections on
emotional experience. Emotion, 10(3), 433–440.
*Davis, J. I., Senghas, A., & Ochsner, K. N. (2009). How does facial feedback modulate emotional
experience? Journal of Research in Personality, 43(5), 822–829.
Del Re, A. C., & Hoyt, W. T. (2010). MAd: Meta-analysis with mean differences. R Package Version
0.8-2. Retrieved from http://cran.r-project.org/web/packages/MAd
*Demaree, H. A., Robinson, J. L., Everhart, D. E., & Schmeichel, B. J. (2004). Resting RSA is
associated with natural and self-regulated responses to negative emotional stimuli. Brain and
Cognition, 56(1), 14–23.
*Demaree, H. A., Schmeichel, B. J., Robinson, J. L., Pu, J., Everhart, D. E., & Berntson, G. G. (2006).
Up- and down-regulating facial disgust: Affective, vagal, sympathetic, and respiratory
consequences. Biological Psychology, 71(1), 90–99.
*Dillon, D. G., Ritchey, M., Johnson, B. D., & LaBar, K. S. (2007). Dissociable effects of conscious
emotion regulation strategies on explicit and implicit memory. Emotion, 7(2), 354–365.
*Dimberg, U., & Söderkvist, S. (2011). The voluntary facial action technique: A method to test the facial
feedback hypothesis. Journal of Nonverbal Behavior, 35(1), 17–33.
*Duncan, J. W., & Laird, J. D. (1977). Cross-modality consistencies in individual differences in self-
attribution. Journal of Personality, 45(2), 191–2016.
*Duncan, J. W., & Laird, J. D. (1980). Positive and reverse placebo effects as a function of differences
in cues used in self-perception. Journal of Personality and Social Psychology, 39(6), 1024–1036.
Duncan, S., & Barrett, L. F. (2007). Affect is a form of cognition: A neurobiological analysis. Cognition
and Emotion, 21(6), 1184–1211.
Duval, S., & Tweedie, R. (2000). Trim and fill: a simple funnel-plot-based method of testing and
adjusting for publication bias in meta-analysis. Biometrics, 56(2), 455–463. Retrieved from
http://www.jstor.org/stable/pdf/2676988.pdf?refreqid=excelsior
%3A34f13c816966804a4e25dd8df41eafaf
*Dzokoto, V., Wallace, D. S., Peters, L., & Bentsi-Enchill, E. (2014). Attention to emotion and non-
Western faces: Revisiting the facial feedback hypothesis. Journal of General Psychology, 141(2),
151–168. Retrieved from http://ovidsp.ovid.com/ovidweb.cgi?
T=JS&CSC=Y&NEWS=N&PAGE=fulltext&D=medl&AN=24846789
http://oxfordsfx.hosted.exlibrisgroup.com/oxford?
sid=OVID:medline&id=pmid:24846789&id=doi:10.1080/00221309.2014.884052&issn=0022-
1309&isbn=&volume=141&issue=2&spage=15
Ekman, P. (1979). Biological and cultural contributions to body and facial movement. In J. Blacking
(Ed.), Anthropology of the Body (pp. 34–38). London: Routledge.
Ekman, P. (1999). Basic emotions. In T. Dalgleish & M. J. Power (Eds.), Handbook of Cognition and
Emotion (pp. 45–60). New York: John Wiley & Sons.
Ekman, P., & Cordaro, D. (2011). What is meant by calling emotions basic. Emotion Review, 3(4), 364-
370.
Ellsworth, P. C. (1994). William James and emotion: Is a century of fame worth a century of

COLES, LARSEN, AND LENCH 41
misunderstanding? Psychological Review, 101(2), 222–229. Retrieved from
http://www.ncbi.nlm.nih.gov/pubmed/8022957
Ellsworth, P. C., & Scherer, K. R. (2003). Appraisal Processes in Emotion. Handbook of Affective
Sciences.
Fisher, Z., & Tipton, E. (2015). Robumeta: An R-package for robust variance estimation in meta-
analysis. Retrieved from http://arxiv.org/abs/1503.02220
*Flack, W. F. (2006). Peripheral feedback effects of facial expressions, bodily postures, and vocal
expressions on emotional feelings. Cognition & Emotion, 20(2), 177–195. Retrieved from
http://www.tandfonline.com/doi/abs/10.1080/02699930500359617
*Flack, W. F. Jr., Laird, J. D., & Cavallaro, L. A. (1999a). Emotional expression and feeling in
schizophrenia: Effects of specific expressive behaviors on emotional experiences. Journal of
Clinical Psychology, 55(1), 1–20.
*Flack, W. F. Jr., Laird, J. D., & Cavallaro, L. A. (1999b). Separate and combined effects of facial
expressions and bodily postures on emotional feelings. European Journal of Social Psychology,
29(2), 203–217.
*Gan, S., Yang, J., Chen, X., & Yang, Y. (2015). The electrocortical modulation effects of different
emotion regulation strategies. Cognitive Neurodynamics, 9(4), 399–410.
Gellhorn, E. (1958). The physiological basis of neuromuscular relaxation. AMA Archives of Internal
Medicine, 102(3), 392–399. Retrieved from http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?
cmd=Retrieve&db=PubMed&dopt=Citation&list_uids=18100421
Gellhorn, E. (1964). Motion and emotion: The role of proprioception in the physiology and pathology of
the emotions. Psychological Review, 71(6), 457–472.
Gerrards-Hesse, A., Spies, K., & Hesse, F. W. (1994). Experimental inductions of emotional states and
their effectiveness: A review. British Journal of Psychology, 1(85), 55–78.
*Goldin, P. R., McRae, K., Ramel, W., & Gross, J. J. (2008). The neural bases of emotion regulation:
Reappraisal and suppression of negative emotion. Biological Psychiatry, 63(6), 577–586.
*Gross, J. J. (1993). Emotional suppression. (Doctoral Dissertation). Retrieved from ProQuest
Dissertations & Theses Global. (9430509)
*Gross, J. J. (1998). Antecedent- and response-focused emotion regulation: Divergent consequences for
experience , expression , and physiology. Journal of Personality and Social Psychology, 74(1),
224–237.
Gross, J. J., & John, O. P. (2003). Individual differences in two emotion regulation processes:
Implications for affect, relationships, and well-being. Journal of Personality and Social
Psychology, 85(2), 348–362.
*Gross, J. J., & Levenson, R. W. (1993). Emotional suppression: Physiology, self-report, and expressive
behavior. Journal of Personality and Social Psychology, 64(6), 970–986.
*Gross, J. J., & Levenson, R. W. (1997). Hiding feelings: The acute effects of inhibiting negative and
positive emotion. Journal of Abnormal Psychology, 106(1), 95–103.
Guenther, K. (1981). Mood and Memory. American Psychologist, 36(2), 129–148.
*Harris, C. R. (2001). Cardiovascular responses of embarrassment and effects of emotional suppression
in a social setting. Journal of Personality and Social Psychology, 81(5), 886-897.
*Hawk, S. T., Fischer, A. H., & Van Kleef, G. A. (2012). Face the noise: Embodied responses to

COLES, LARSEN, AND LENCH 42
nonverbal vocalizations of discrete emotions. Journal of Personality and Social
Psychology, 102(4), 796–814.
Head, M. L., Holman, L., Lanfear, R., Kahn, A. T., & Jennions, M. D. (2015). The extent and
consequences of p-hacking in science. PLoS Biology, 13(3). Retrieved from
http://journals.plos.org/plosbiology/article/file?id=10.1371/journal.pbio.1002106&type=printable
Hedges, L. V., & Pigott, T. D. (2004). The power of statistical tests for moderators in meta-analysis.
Psychological Methods, 9(4), 426–445.
Hedges, L. V, Tipton, E., & Johnson, M. C. (2010). Robust variance estimation in meta-regression with
dependent effect size estimates. Research Synthesis Methods, 1(1), 39–65.
*Helt, M. S., & Fein, D. A. (2016). Facial feedback and social input: Effects on laughter and enjoyment
in children with autism spectrum disorders. Journal of Autism and Developmental Disorders, 46(1),
83–94.
*Hendricks, M. A. (2013). An exploration of the relationship between emotional and cognitive control.
(Doctoral Dissertation). Retrieved from ProQuest Dissertations & Theses Global. (3587340)
*Hendricks, M. A., & Buchanan, T. W. (2016). Individual differences in cognitive control processes and
their relationship to emotion regulation. Cognition and Emotion, 30(5), 912–924.
*Henry, J. D., Green, M. J., de Lucia, A., Restuccia, C., McDonald, S., & O'Donnell, M. (2007).
Emotion dysregulation in schizophrenia: Reduced amplification of emotional expression is
associated with emotional blunting. Schizophrenia Research, 95(1), 197–204.
*Henry, J. D., Green, M. J., Restuccia, C., de Lucia, A., Rendell, P. G., McDonald, S., & Grisham, J. R.
(2009). Emotion dysregulation and schizotypy. Psychiatry Research, 166(2), 116–124.
*Henry, J. D., Rendell, P. G., Scicluna, A., Jackson, M., & Phillips, L. H. (2009). Emotion experience,
expression, and regulation in Alzheimer's disease. Psychology and Aging, 24(1), 252–257.
*Hess, U., Kappas, A., McHugo, G. J., Lanzetta, J. T., & Kleck, R. E. (1992). The facilitative effect of
facial expression on the self-generation of emotion. International Journal of
Psychophysiology, 12(3), 251–265.
Hillier, S., Immink, M., & Thewlis, D. (2015). Assessing proprioception. Neurorehabilitation and
Neural Repair, 29(10), 933–949. Retrieved from
http://journals.sagepub.com/doi/10.1177/1545968315573055
*Hofmann, S. G., Heering, S., Sawyer, A. T., & Asnaani, A. (2009). How to handle anxiety: The effects
of reappraisal, acceptance, and suppression strategies on anxious arousal. Behaviour Research and
Therapy, 47(5), 389–394.
Hunter, P. G., Schellenberg, E. G., & Schimmack, U. (2010). Feelings and perceptions of happiness and
sadness induced by music: Similarities, differences, and mixed emotions. Psychology of Aesthetics,
Creativity, and the Arts, 4(1), 47–56.
Irons, D. (1894). Prof. James’ Theory of Emotion. Mind, 3(9), 77–97. Retrieved from
http://www.jstor.org/stable/pdf/2247933.pdf?
refreqid=excelsior:60120736828a2b1877e305c6e7944f2e
Itkes, O., Kimchi, R., Kron, A., & Carmel, M. (2017). Dissociating affective and semantic valence.
Journal of Experimental Psychology: General, 146(7), 924–942.
*Ito, T. A., Chiao, K. W., Devine, P. G., Lorig, T. S., & Cacioppo, J. T. (2006). The influence of facial
feedback on race bias. Psychological Science, 17(3), 256–261.

COLES, LARSEN, AND LENCH 43
Izard, C. E. (1971). The Face of Emotion. East Norwalk: Appleton-Century-Crofts.
Izard, C. E. (1977). Human Emotions. (C. E. Izard & J. L. Singer, Eds.). New York: Plenum Publishing
Corporation.
Izard, C. E. (1990a). Facial expressions and the regulation of emotions. Journal of Personality and
Social Psychology, 58(3), 487–498.
Izard, C. E. (1990b). The substrates and functions of emotion feelings: William James and current
emotion theory. Personality and Social Psychology Bulletin, 16(4), 626–635. Retrieved from
http://psp.sagepub.com/cgi/doi/10.1177/0146167290164004
Izard, C. E. (2007). Basic emotions, natural kinds, emotion schemas, and a new paradigm. Perspectives
on Psychological Science, 2(3), 260–280. Retrieved from http://www.jstor.org/stable/40212206
James, W. (1884). What is an emotion? Mind, 9(34), 188–205.
James, W. (1890). Principles of Psychology. New York: Henry Holt and Company.
James, W. (1894). Discussion: The physical basis of emotion. Psychological Review, 1(5), 516–529.
John, L. K., Loewenstein, G., & Prelec, D. (2012). Measuring the prevalence of questionable research
practices with incentives for truth telling. Psychological Science, 23(5), 524–532.
*Kalokerinos, E. K., Greenaway, K. H., & Denson, T. F. (2015). Reappraisal but not suppression
downregulates the experience of positive and negative emotion. Emotion, 15(3), 271–275.
Retrieved from http://proxy.lib.utk.edu:90/login?
url=http://search.proquest.com/docview/1604778202?accountid=14766
*Kao, C. H., Su, J. C., Crocker, J., & Chang, J. H. (2016). The benefits of transcending self-interest:
Examining the role of self-transcendence on expressive suppression and well-being. Journal of
Happiness Studies, 1–17, 959-975.
*Kappas, A. (1989). Control of emotion. (Doctoral Dissertation). Retrieved from
http://proxy.lib.utk.edu:90/login?url=http://search.proquest.com/docview/303722249?
accountid=14766
*Kircher, T., Pohl, A., Krach, S., Thimm, M., Schulte-Rüther, M., Anders, S., & Mathiak, K. (2012).
Affect-specific activation of shared networks for perception and execution of facial expressions.
Social Cognitive and Affective Neuroscience, 8(4), 370–377. Retrieved from https://oup.silverchair-
cdn.com/oup/backfile/Content_public/Journal/scan/8/4/10.1093_scan_nss008/3/nss008.pdf?
Expires=1491833120&Signature=fHMWCu~SGaPRQXeR~cUmXid-
DZra~hEnGX2XoJdoaSHqRV8pckzzF3zKMg-
6Xjgimmw2wk2JTfJyOrP~COU6PH~AHcj9jVTqaOqbq3fQK7TH3UvqXPGV
*Korb, S., Grandjean, D., Samson, A. C., Delplanque, S., & Scherer, K. R. (2012). Stop laughing!
Humor perception with and without expressive suppression. Social Neuroscience, 7(5), 510–524.
Kraft, T. L., & Pressman, S. D. (2012). Grin and bear it: The influence of manipulated facial expression
on the stress response. Psychological Science, 23(11), 1372-1378.
*Labott, S. M., & Teleha, M. K. (1996). Weeping propensity and the effects of laboratory expression or
inhibition. Motivation and Emotion, 20(3), 273–284.
Kring, A. M., Smith, D. A., & Neale, J. M. (1994). Individual differences in dispositional
expressiveness: Development and validation of the Emotional Expressivity Scale. Journal of
Personality and Social Psychology, 66(5), 934–949.
LaFrance, M., Hecht, M. A., & Paluck, E. L. (2003). The contingent smile: A meta-analysis of sex

COLES, LARSEN, AND LENCH 44
differences in smiling. Psychological Bulletin, 129(2), 305–334.
*Laird, J. D. (1974). Self-attribution of emotion: The effects of expressive behavior on the quality of
emotional experience. Journal of Personality and Social Psychology, 29(4), 475–486.
Laird, J. D. (1984). The real role of facial response in the experience of emotion: A reply to Tourangeau
and Ellsworth, and others. Journal of Personality and Social Psychology, 47(4), 909–917.
Laird, J. D., & Bresler, C. (1992). The process of emotional experience: A self-perception theory. In M.
S. Clark (Ed.), Review of Personality and Social Psychology: Emotion (pp. 213–234). Newbury
Park.
*Laird, J. D., & Crosby, M. (1974). Individual differences in self-attribution of emotion. In H. London &
R. E. Nisbett (Eds.), Thought and Feeling: Cognitive Alteration of Feeling States (pp. 44–59). New
York: Routledge.
Laird, J. D., & Lacasse, K. (2014). Bodily influences on emotional feelings: Accumulating evidence and
extensions of William James’s theory of emotion. Emotion Review, 6(1), 27–34.
*Lalot, F., Delplanque, S., & Sander, D. (2014). Mindful regulation of positive emotions: A comparison
with reappraisal and expressive suppression. Frontiers in Psychology, 5(243), 1–9.
Lange, C. G. (1885). Om sindsbevaegelser; et psyko-fysiologisk studie. Lund.
Lange, C. G. (1922). The emotions: A psychophysiological study. In K. Dunlap (Ed.), The Emotions (pp.
33–90). Baltimore: Williams and Wilkins Company.
Larsen, J. T. (2017). Holes in the case for mixed emotions. Emotion Review, 9(2), 118–123.
Larsen, J. T., McGraw, A. P., & Cacioppo, J. T. (2001). Can people feel happy and sad at the same time?
Journal of Personality and Social Psychology, 81(4), 684–696. Retrieved from
http://doi.apa.org/getdoi.cfm?doi=10.1037/0022-3514.81.4.684
*Larsen, R. J., Kasimatis, M., & Frey, K. (1992). Facilitating the furrowed brow: An unobtrusive test of
the facial feedback hypothesis applied to unpleasant affect. Cognition & Emotion, 6(5), 321–338.
Retrieved from http://www.tandfonline.com/action/journalInformation?journalCode=pcem20
*Lee, E. A. (2011). Expressive suppression of negative emotion: A comparison of asian american and
european american norms for emotion regulation. (Doctoral Dissertation). Retrieved from
ProQuest Dissertations & Theses Global. (3483720)
Lehrer, J. (2010, December). The decline effect and the scientific method. The New Yorker, Retrieved
from https://www.newyorker.com/magazine/2010/12/13/the-truth-wears-off
Lench, H. C., Flores, S. A., & Bench, S. W. (2011). Discrete emotions predict changes in cognition,
judgment, experience, behavior, and physiology: A meta-analysis of experimental emotion
elicitations. Psychological Bulletin, 137(5), 834–855.
Leplow, B., Schluter, V., & Ferstl, R. (1992). A new procedure for assessment of proprioception.
Perceptual and Motor Skills, 74(1), 91–98. Retrieved from
http://journals.sagepub.com/doi/pdf/10.2466/pms.1992.74.1.91
Levenson, R. W., Ekman, P., & Friesen, W. V. (1990). Voluntary facial action generates emotion-specific
autonomic nervous system activity. Psychophysiology, 27(4), 363–384.
*Lewis, M. B. (2012). Exploring the positive and negative implications of facial feedback. Emotion,
12(4), 852–859.
*Lewis, M. B., & Bowler, P. J. (2009). Botulinum toxin cosmetic therapy correlates with a more positive
mood. Journal of Cosmetic Dermatology, 8(1), 24–26.

COLES, LARSEN, AND LENCH 45
Lindquist, K. A. (2013). Emotions emerge from more basic psychological ingredients: A modern
psychological constructionist model. Emotion Review, 5(4), 356–368. Retrieved from
http://emr.sagepub.com/cgi/doi/10.1177/1754073913489750
Lumley, M. A., Cohen, J. L., Borszcz, G. S., Cano, A., Radcliffe, A. M., Porter, L. S., … Keefe, F. J.
(2011). Pain and emotion: A biopsychosocial review of recent research. Journal of Clinical
Psychology, 67(9), 942–968. Retrieved from
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3152687/pdf/nihms295222.pdf
Lyubomirsky, S. (2008). The how of happiness: A scientific approach to getting the life you want.
Penguin Press: New York, NY.
*Ma, C. (2011). Culture and emotional suppression. (Doctoral Dissertation). Retrieved from ProQuest
Dissertations & Theses Global. (3495692)
Macaskill, P., Walter, S. D., & Irwig, L. (2001). A comparison of methods to detect publication bias in
meta-analysis. Statistics in Medicine, 20(4), 641–654. Retrieved from
http://dx.doi.org/10.1002/sim.698
MacCormack, J. K., & Lindquist, K. A. (2017). Bodily contributions to emotion: Schachter’s legacy for
a psychological constructionist view on emotion. Emotion Review, 9(1), 36–45.
MacCormack, J. K., & Lindquist, K. A. (2018). Feeling hangry? When hunger is conceptualized as
emotion. Emotion, 1–19. Retrieved from
https://search.proquest.com/openview/a6ee662b30e2cc5257952e157ca4a98f/1?pq-
origsite=gscholar&cbl=18750&diss=y
*Maldonado, R. C., DiLillo, D., & Hoffman, L. (2015). Can college students use emotion regulation
strategies to alter intimate partner aggression-risk behaviors? An examination using I³
theory. Psychology of Violence, 5(1), 46–55.
*Marmolejo-Ramos, F., & Dunn, J. (2013). On the activation of sensorimotor systems during the
processing of emotionally-laden stimuli. Universitas Psychologica, 12(5), 1511–1542.
*Martijn, C., Tenbült, P., Merckelbach, H., Dreezens, E., & de Vries, N. K. (2002). Getting a grip on
ourselves: Challenging expectancies about loss of energy after self-control. Social Cognition, 20(6),
441–460.
Martin, M. (1990). On the induction of mood. Clinical Psychology Review, 10(6), 669–697.
Matsumoto, D. (1987). The role of facial response in the experience of emotion: More methodological
problems and a meta-analysis. Journal of Personality and Social Psychology, 52(4), 769–774.
Retrieved from http://proxy.lib.utk.edu:90/login?
url=http://search.proquest.com/docview/617268981?accountid=14766
*McCanne, T. R., & Anderson, J. A. (1987). Emotional responding following experimental manipulation
of facial electromyographic activity. Journal of Personality and Social Psychology, 52(4), 759–
768. Retrieved from http://proxy.lib.utk.edu:90/login?
url=http://search.proquest.com/docview/617264912?accountid=14766
*McCaul, K. D., Holmes, D. S., & Solomon, S. (1982). Voluntary expressive changes and
emotion. Journal of Personality and Social Psychology, 42(1), 145–152.
McIntosh, D. N. (1996). Facial feedback hypotheses: Evidence, implications, and directions. Motivation
and Emotion, 20(2), 121–147.
*McIntosh, D.N., Zajonc, R.B., Vig, P.S., & Emerick, S.W. (1997). Facial movement, breathing,

COLES, LARSEN, AND LENCH 46
temperature, and affect: Implications of the vascular theory of emotional efference. Cognition &
Emotion, 11(2), 171–196.
McRae, K., Ochsner, K. N., Mauss, I. B., Gabrieli, J. J. D., & Gross, J. J. (2008). Gender differences in
emotion regulation: An fMRI study of cognitive reappraisal. Group Processes & Intergroup
Relations, 11(2), 143–162. Retrieved from http://dx.doi.org/10.1177/1368430207088035
*Meeten, F., Ivak, P., Dash, S., Knowles, S., Duka, T., Scott, R., ... & Davey, G. C. (2015). The effect of
facial expressions on the evaluation of ambiguous statements. Journal of Experimental
Psychopathology, 6(3), 253–263.
Mehling, W. E., Price, C., Daubenmier, J. J., Acree, M., Bartmess, E., Stewart, A., … Holloway, R.
(2012). The Multidimensional Assessment of Interoceptive Awareness (MAIA). PLoS ONE, 7(11).
Retrieved from http://journals.plos.org/plosone/article/file?
id=10.1371/journal.pone.0048230&type=printable
*Miyamoto, Y. (2006). Culturally situated cognition: The influence of communication practices,
physical environments, and intra-individual motor actions (Doctoral dissertation). Retrieved from
ProQuest Dissertations & Theses Global. (3224701)
Moher, D., Liberati, A., Tetzlaff, J., Altman, D. G., & The PRISMA Group. (2009). Preferred reporting
items for systematic reviews and meta-analyses: The PRISMA Statement. Physical Therapy, 89(9),
873–880. Retrieved from http://journals.plos.org/plosmedicine/article/file?
id=10.1371/journal.pmed.1000097&type=printable
*Moore, S. A., & Zoellner, L. A. (2012). The effects of expressive and experiential suppression on
memory accuracy and memory distortion in women with and without PTSD. Journal of
Experimental Psychopathology, 3(3), 368–392.
Moors, A., Ellsworth, P. C., Scherer, K. R., & Frijda, N. H. (2013). Appraisal theories of emotion: State
of the art and future development. Emotion Review, 5(2), 119–124.
Neal, D. T., & Chartrand, T. L. (2011). Embodied emotion perception: Amplifying and dampening facial
feedback modulates emotion perception accuracy. Social Psychological and Personality Science,
2(6), 673–678. Retrieved from http://proxy.lib.utk.edu:90/login?
url=http://search.proquest.com/docview/912100560?accountid=14766
Niedenthal, P. M., Barsalou, L. W., Winkielman, P., Krauth-Gruber, S., & Ric, F. (2005). Embodiment in
Attitudes , Social Perception , and Emotion. Personality and Social Psychology Review, 9(3), 184–
211.
Noah, T., Schul, Y., & Mayo, R. (2018). When both the original study and its failed replication are
correct: Feeling observed eliminates the facial-feedback effect. Journal of Personality and Social
Psychology, 114(5), 657–664.
Nolen-Hoeksema, S., & Aldao, A. (2011). Gender and age differences in emotion regulation strategies
and their relationship to depressive symptoms. Personality and Individual Differences, 51(6), 704–
708. Retrieved from http://www.elsevier.com/copyright
*Ohira, H., & Kurono, K. (1993). Facial feedback effects on impression formation. Perceptual and
Motor Skills, 77(3), 1251–1258.
*Paredes, B., Stavraki, M., Briñol, P., & Petty, R. E. (2013). Smiling after thinking increases reliance on
thoughts. Social Psychology, 44(5), 349–353.
*Paul, S., Simon, D., Kniesche, R., Kathmann, N., & Endrass, T. (2013). Timing effects of antecedent-

COLES, LARSEN, AND LENCH 47
and response-focused emotion regulation strategies. Biological Psychology, 94(1), 136–142.
*Pedder, D. J., Terrett, G., Bailey, P. E., Henry, J. D., Ruffman, T., & Rendell, P. G. (2016). Reduced
facial reactivity as a contributor to preserved emotion regulation in older adults. Psychology and
Aging, 31(1), 114–125. Retrieved from http://proxy.lib.utk.edu:90/login?
url=http://search.proquest.com/docview/1751220896?accountid=14766
Pennebaker, J. W., & Roberts, T.-A. (1992). Toward a his and hers theory of emotion: Gender
differences in visceral perception. Journal of Social and Clinical Psychology, 11(3), 199–212.
Retrieved from https://www.researchgate.net/publication/273218118
*Phillips, L. H., Henry, J. D., Hosie, J. A., & Milne, A. B. (2008). Effective regulation of the experience
and expression of negative affect in old age. The Journals of Gerontology Series B: Psychological
Sciences and Social Sciences, 63(3), 138–145.
Price, T. F., & Harmon-Jones, E. (2015). Embodied emotion: The influence of manipulated facial and
bodily states on emotive responses. Wiley Interdisciplinary Reviews: Cognitive Science.
Price, T. F., Peterson, C. K., & Harmon-Jones, E. (2012). The emotive neuroscience of embodiment.
Motivation and Emotion, 36(1), 27–37.
Pustejovsky, J. E. (2017). clubSandwich: Cluster-robust (sandwich) variance estimators with small-
sample corrections. R Package Version 0.3.0.
R Core Team. (2017). R: A language and environment for statistical computing. R Foundation for
Statistical Computing, Vienna, Austria. URL https://www.R-project.org/.
Raudenbush, S. W., Becker, B. J., & Kalaian, H. (1988). Modeling multivariate effect sizes.
Psychological Bulletin, 103(1), 111.
Reisenzein, R., Horstmann, G., & Schützwohl, A. (2017). The cognitive evolutionary model of surprise:
A review of the evidence. Topics in Cognitive Science, 1–25. Retrieved from
http://doi.wiley.com/10.1111/tops.12292
*Reisenzein, R., & Studtmann, M. (2007). On the expression and experience of surprise: No evidence
for facial feedback, but evidence for a reverse self-inference effect. Emotion, 7(3), 612–627.
*Richards, J. M., Butler, E. A., & Gross, J. J. (2003). Emotion regulation in romantic relationships: The
cognitive consequences of concealing feelings. Journal of Social and Personal
Relationships, 20(5), 599–620.
*Richards, J. M., & Gross, J. J. (1999). Composure at any cost? The cognitive consequences of emotion
suppression. Personality and Social Psychology Bulletin, 25(8), 1033-1044.
*Richards, J. M., & Gross, J. J. (2000). Emotion regulation and memory: The cognitive costs of keeping
one's cool. Journal of Personality and Social Psychology, 79(3), 410–424.
*Richards, J. M., & Gross, J. J. (2006). Personality and emotional memory: How regulating emotion
impairs memory for emotional events. Journal of Research in Personality, 40(5), 631–651.
*Roberts, N. A., Levenson, R. W., & Gross, J. J. (2008). Cardiovascular costs of emotion suppression
cross ethnic lines. International Journal of Psychophysiology, 70(1), 82–87.
*Robinson, J. L., & Demaree, H. A. (2009). Experiencing and regulating sadness: Physiological and
cognitive effects. Brain and Cognition, 70(1), 13–20.
*Roemer, R. J. (2014). Smiling into happiness: Conditioning positive affect. (Doctoral Dissertation).
Available from ProQuest Dissertations & Theses Global. (3639733)
Rohatgi, A. (2011). WebPlotDigitizer. Retrieved from

COLES, LARSEN, AND LENCH 48
https://automeris.io/WebPlotDigitizer/userManual.pdf
*Rohrmann, S., Hopp, H., Schienle, A., & Hodapp, V. (2009). Emotion regulation, disgust sensitivity,
and psychophysiological responses to a disgust-inducing film. Anxiety, Stress, & Coping, 22(2),
215–236.
Roseman, I. J., & Smith, C. A. (2001). Appraisal theory: Overview, assumptions, varieties,
controversies. In K. R. Scherer, A. Schorr, & T. Johnstone (Eds.), Appraisal Processes in Emotion:
Theory, Methods, Research (pp. 3–19). New York: Oxford University Press.
Rosenthal, R., & Rubin, D. B. (1986). Meta-analytic procedures for combining studies with multiple
effect sizes. Psychological Bulletin, 99(3), 400–406.
Rothstein, H. R., Sutton, A. J., & Borenstein, M. (2006). Publication bias in meta-analysis: Prevention,
assessment and adjustments. West Sussex: John Wiley & Sons. Retrieved from
http://www.springerlink.com/index/L8N2877843431200.pdf
*Rummer, R., Schweppe, J., Schlegelmilch, R., & Grice, M. (2014). Mood is linked to vowel type: The
role of articulatory movements. Emotion, 14(2), 246–250
Russell, J. A. (1980). A circumplex model of affect. Journal of Personality and Social Psychology,
39(6), 1161–1178. Retrieved from http://psycnet.apa.org/journals/psp/39/6/1161
Russell, J. A. (2003). Core affect and the psychological construction of emotion. Psychological Review,
110(1), 145–172. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/12529060
Russell, J. A. (2014). My psychological constructionist perspective, with a focus on conscious affective
experience. In L. F. Barrett & J. A. Russell (Eds.), The Psychological Construction of Emotion (pp.
183–208). New York: Guilford Press.
Russell, J. A. (2017). Mixed emotions viewed from the psychological constructionist perspective.
Emotion Review, 9(2), 111–117.
Russell, J. A., & Barrett, L. F. (1999). Core affect, prototypical emotional episodes, and other things
called emotion: dissecting the elephant. Journal of Personality and Social Psychology, 76(5), 805.
Russell, J. A., & Carroll, J. M. (1999). On the bipolarity of positive and negative affect. Psychological
Bulletin, 125(1), 3–30. Retrieved from http://www.ncbi.nlm.nih.gov/sites/entrez?
Db=pubmed&DbFrom=pubmed&Cmd=Link&LinkName=pubmed_pubmed&LinkReadableName
=Related
Articles&IdsFromResult=9990843&ordinalpos=3&itool=EntrezSystem2.PEntrez.Pubmed.Pubmed
_ResultsPanel.Pubmed_RVDocSum
Scherer, K. R. (2009). The dynamic architecture of emotion: Evidence for the component process model.
Cognition and Emotion, 23(7), 1307–1351.
Schimmack, U. (2012). The ironic effect of significant results on the credibility of multiple-study
articles. Psychological Methods, 17(4), 551–566.
*Schmeichel, B. J., Vohs, K. D., & Baumeister, R. F. (2003). Intellectual performance and ego depletion:
Role of the self in logical reasoning and other information processing. Journal of Personality and
Social Psychology, 85(1), 33–46.
*Schmeichel, B. J., Volokhov, R. N., & Demaree, H. A. (2008). Working memory capacity and the self-
regulation of emotional expression and experience. Journal of Personality and Social
Psychology, 95(6), 1526–1540.
Schooler, J. (2011). Unpublished results hide the decline effect: Some effects diminish when tests are

COLES, LARSEN, AND LENCH 49
repeated. Nature, 470(7335), 437–438.
Schwarz, N., & Clore, G. L. (2007). Feelings and phenomenal experiences. In A. Kruglanski & E. T.
Higgins (Eds.), Social Psychology: Handbook of Basic Principles (2nd ed., pp. 385–407). New
York: Guilford.
Sergi, G. (1894). Principi Di Psicologie: Dolore e Piacere; Storia Naturale dei Sentimenti. (F.
Dumolard, Ed.). Milan: Librai Della Real Casa.
Sherrington, C. S. (1900). Experiments on the value of vascular and visceral factors for the genesis of
emotion. Proceedings of the Royal Society of London, 66, 390–403.
Simonsohn, U., Nelson, L. D., & Simmons, J. P. (2014). P-curve: A key to the file-drawer. Journal of
Experimental Psychology: General, 143(2), 534–547. Retrieved from
http://dx.doi.org/10.1037/a0033242.supp
Smith, C. A., & Kirby, L. D. (2004). Appraisal as a pervasive determinant of anger. Emotion, 4(2), 133–
138. Retrieved from https://www.researchgate.net/publication/8485283
*Söderkvist, S., & Dimberg, U. The influence of facial skin and facial muscles on facial feedback.
Manuscript in preparation.
*Söderkvist, S., Ohlén, K., & Dimberg, U. How the experience of emotion is modulated by facial
feedback; Study 1. (submitted).
*Söderkvist, S., Ohlén, K., & Dimberg, U. How the experience of emotion is modulated by facial
feedback; Study 2. (submitted).
*Söderkvist, S., Ohlén, K., & Dimberg, U. (2018). How the experience of emotion is modulated by
facial feedback. Journal of Nonverbal Behavior, 42(1), 129-151.
*Soussignan, R. (2002). Duchenne smile, emotional experience, and autonomic reactivity: A test of the
facial feedback hypothesis. Emotion, 2(1), 52–74.
*Soussignan, R. (2004). Regulatory function of facial actions in emotion processes. In F. Columbus
(Ed.), Advances in Psychology Research (Vol. 31, pp. 171–196). Hauppauge: Nova Science
Publishers, Inc. Retrieved from
http://www.researchgate.net/publication/233734596_REGULATORY_FUNCTION_OF_FACIAL_
ACTIONS_IN_EMOTION_PROCESSES/file/9fcfd50ae666b60e4f.pdf
Stanley, T. D. (2017). Limitations of PET-PEESE and other meta-analysis methods. Social
Psychological and Personality Science, 8(5), 581–591. Retrieved from
http://journals.sagepub.com/doi/pdf/10.1177/1948550617693062
Stanley, T. D., & Doucouliagos, H. (2014). Meta-regression approximations to reduce publication
selection bias. Research Synthesis Methods, 5(1), 60–78. Retrieved from
https://core.ac.uk/download/pdf/6610507.pdf?repositoryId=153
*Stel, M., van den Heuvel, C., & Smeets, R. C. (2008). Facial feedback mechanisms in autistic spectrum
disorders. Journal of Autism and Developmental Disorders, 38(7), 1250–1258. Retrieved from
http://proxy.lib.utk.edu:90/login?url=http://search.proquest.com/docview/622115409?
accountid=14766
Sterne, J. A. C., Sutton, A. J., Ioannidis, J. P. A., Terrin, N., Jones, D. R., Lau, J., … Higgins, J. P. T.
(2011). Recommendations for examining and interpreting funnel plot asymmetry in meta-analyses
of randomised controlled trials. BMJ, 343(7818), 1–8. Retrieved from
http://researchonline.lshtm.ac.uk/18619/

COLES, LARSEN, AND LENCH 50
Strack, F. (2016). Reflection on the smiling registered replication report. Perspectives on Psychological
Science, 11(6), 929–930. Retrieved from http://pps.sagepub.com/cgi/content/short/11/6/929
*Strack, F., Martin, L. L., & Stepper, S. (1988). Inhibiting and facilitating conditions of the human
smile: A nonobtrusive test of the facial feedback hypothesis. Journal of Personality and Social
Psychology, 54(5), 768–777.
Sutherland, A. (1898). The origin and growth of the moral instinct. London: Longmans, Green, and Co.
*Tamir, M., Robinson, M. D., Clore, G. L., Martin, L. L., & Whitaker, D. J. (2004). Are we puppets on a
string? The contextual meaning of unconscious expressive cues. Personality and Social Psychology
Bulletin, 30(2), 237–249.
Tanner-Smith, E. E., & Tipton, E. (2014). Robust variance estimation with dependent effect sizes:
Practical considerations including a software tutorial in Stata and SPSS. Research Synthesis
Methods, 5(1), 13–30.
Tanner-Smith, E. E., Tipton, E., & Polanin, J. R. (2016). Handling complex meta-analytic data structures
using robust variance estimates: A tutorial in R. Journal of Developmental and Life-Course
Criminology, 2(1), 85–112. Retrieved from http://dx.doi.org/10.1007/s40865-016-0026-5
Terrin, N., Schmid, C. H., Lau, J., & Olkin, I. (2003). Adjusting for publication bias in the presence of
heterogeneity. Statistics in Medicine, 22(13), 2113–2126.
Tomkins, S. S. (1962). Affect Imagery Consciousness: The Positive Affects. Springer.
Tomkins, S. S. (1981). The role of facial response in the experience of emotion: A reply to Tourangeau
and Ellsworth. Journal of Personality and Social Psychology, 37(9), 1519–1531.
*Tourangeau, R., & Ellsworth, P. C. (1979). The role of facial response in the experience of emotion.
Journal of Personality and Social Psychology, 37(9), 1519–1531.
*Trent, J. (2010). Intuition and facial feedback (Doctoral dissertation). University of Missouri,
Columbia.
Van den Noortgate, W., Lopez-Lopez, J. A., Marin-Martinez, F., & Sanchez-Meca, J. (2015). Meta-
analysis of multiple outcomes: A multilevel approach. Behavior Research Methods, 47(4), 1274–
1294. Retrieved from https://link.springer.com/content/pdf/10.3758%2Fs13428-014-0527-2.pdf
Verduyn, P., Delaveau, P., Rotgé, J. Y., Fossati, P., & Van Mechelen, I. (2015). Determinants of emotion
duration and underlying psychological and neural mechanisms. Emotion Review, 7(4), 330–335.
Vevea, J. L., & Hedges, L. V. (1995). A general linear model for estimating effect size in the presence of
publication bias. Psychometrika, 60(3), 419–435. Retrieved from
http://download.springer.com/static/pdf/66/art%253A10.1007%252FBF02294384.pdf?
originUrl=http%3A%2F%2Flink.springer.com%2Farticle
%2F10.1007%2FBF02294384&token2=exp=1496334135~acl=%2Fstatic%2Fpdf%2F66%2Fart
%25253A10.1007%25252FBF02294384.pdf%3ForiginUrl%3Dh
*Vieillard, S., Harm, J., & Bigand, E. (2015). Expressive suppression and enhancement during music-
elicited emotions in younger and older adults. Frontiers in Aging Neuroscience, 7, 14. Retrieved
from http://proxy.lib.utk.edu:90/login?url=http://search.proquest.com/docview/1692335162?
accountid=14766
*Wagenmakers, E. J., Beek, T., Dijkhoff, L., Gronau, Q. F., Acosta, A., Adams, R. B., … Zwaan, R. A.
(2016). Registered replication report: Strack, Martin, & Stepper (1988). Perspectives on
Psychological Science, 11(6), 917–928.

COLES, LARSEN, AND LENCH 51
Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures of
positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology,
54(6), 1063–1070.
Waynbaum, I. (1907). La Physionomie Humaine: Son Mecanisme Et Son Roˆle Social. (F. Alcan, Ed.).
Paris: Librairies Felix Alcan Et Guillaumin Reunies.
Webb, T. L., Miles, E., & Sheeran, P. (2012). Dealing with feeling: A meta-analysis of the effectiveness
of strategies derived from the process model of emotion regulation. Psychological Bulletin, 138(4),
775–808. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/22582737
Westermann, R., Spies, K., Stahl, G., & Hesse, F. W. (1996). Relative effectiveness and validity of mood
induction procedures: A meta-analysis. European Journal of Social Psychology, 26(4), 557–580.
Whissell, C. M. (1985). The role of the face in human emotion: First system or one of many? Perceptual
and Motor Skills.
Winton, W. M. (1986). The role of facial response in self-reports of emotion: A critique of Laird.
Journal of Personality and Social Psychology, 50(4), 808–812. Retrieved from
http://doi.apa.org/getdoi.cfm?doi=10.1037/0022-3514.50.4.808
*Wittmer, V. T. (1985). Defensive style: Manipulation of facial expressiveness and its impact on
psychophysiological reactivity and self-reported anxiety (repression-sensitization). (Doctoral
Dissertation). Retrieved from ProQuest Dissertations & Theses Global. (8601507)
Worcester, W. L. (1893). Observations on some points in James’s Psychology. II. Emotion. The Monist,
3(2), 285–298.
Wundt, W. M. (1886). Philosophische studien (Vol. 3). Leipzig: Wilhelm Engelmann.
*Yartz, A. R. (2003). Individual differences in disgust sensitivity and voluntary emotion regulation:
Subjective, physiological, and behavioral responses to disgust and pleasant pictures. (Doctoral
Dissertation). Retrieved from ProQuest Dissertations & Theses Global. (3102423)
Zajonc, R. B. (1985). Emotion and facial efference: A theory reclaimed. Science, 228(4695), 15–21.
*Zajonc, R. B., Murphy, S. T., & Inglehart, M. (1989). Feeling and facial efference: Implications of the
vascular theory of emotion. Psychological Review, 96(3), 395–416.
*Zariffa, J., Hitzig, S. L., & Popovic, M. R. (2014). Neuromodulation of emotion using functional
electrical stimulation applied to facial muscles. Neuromodulation: Technology at the Neural
Interface, 17(1), 85–92.
*Zhu, N., Cai, Y., Sun, F., & Yang-yang, Y. (2015). Mapping the emotional landscape: The role of
specific emotions in conceptual categorization. Acta Psychologica, 159, 41–51. Retrieved from
http://dx.doi.org/10.1016/j.actpsy.2015.05.003

COLES, LARSEN, AND LENCH 52
Table 2
Characteristics and Effect Size Information of Study Samples Included in the Meta-Analysis (k = 286)
Study Experience
or judgment
Modulation
or initiation
Discrete or
dimensional Emotion
Awareness
of
manipulation
Awareness
of
recording
Manipulation Design Stimuli % of
women
Measurement
timing
Publication
status Nd
Andréasson & Dimberg (2008) judgment - - - NA no incidental-suppress between film 51.79 during published 112 -0.22
Andréasson (2010) Study 3 judgment - - - NA no incidental-suppress within film - during unpublished 48 -0.05
Andréasson (2010) Study 3 judgment - - - NA no incidental-suppress within film - during unpublished 48 -0.35
Andréasson (2010) Study 4 judgment - - - NA no exp.pose-exp.pose within film 51.14 during unpublished 44 0.49
Andréasson (2010) Study 4 judgment - - - NA no exp.pose-exp.pose within film 51.14 during unpublished 44 0.31
Baumeister et al. (2016) judgment - - - NA no Botox-control within sentences 100 during published 10 1.26
Baumeister et al. (2016) judgment - - - NA no Botox-control within sentences 100 during published 10 0.63
Bodenhausen et al. (1994) experience initiation - - aware no exp.pose-control between NA 72.55 - published 51 0.55
Bush et al. (1989) experience modulation discrete happiness aware no suppress-control between film 46.58 after published 69 0.16
Butler et al. (2003) Study 1 experience modulation dimensional negativity aware yes suppress-control between context 100 after published 24 -0.1
Butler et al. (2003) Study 2 experience modulation dimensional negativity aware yes suppress-control between context 100 after published 42 -0.83
Butler et al. (2006) experience modulation dimensional negativity aware no suppress-control between film 100 after published 69 -0.03
Cai et al. (2016) experience modulation dimensional negativity aware no suppress-control within pictures 51.47 after published 68 -0.08
Ceschi & Scherer (2003) judgment - - - aware no suppress-control between context 51.56 - published 64 0.74
Clapp (2012) experience modulation discrete happiness aware yes suppress-control between film 56.8 during unpublished 99 0.69
Clapp (2012) experience modulation discrete sadness aware yes suppress-control between film 56.8 during unpublished 93 0.08
Clapp (2012) experience modulation dimensional negativity aware yes suppress-control between film 56.8 during unpublished 93 0.17
Clapp (2012) experience modulation dimensional positivity aware yes suppress-control between film 56.8 during unpublished 99 0.27
Laird & Crosby (1974) Study 1 experience modulation dimensional positivity NA no exp.pose-exp.pose within pictures 50 - published 26 -0.13
Laird & Crosby (1974) Study 2 experience modulation dimensional positivity NA no exp.pose-exp.pose within pictures 50 - published 26 0.35
Davey et al. (2013) Study 1 experience initiation discrete anger unaware no incidental-control between NA 76.67 during published 28 0.41
Davey et al. (2013) Study 1 experience initiation discrete anger unaware no incidental-control within NA 76.67 during published 14 0.62
Davey et al. (2013) Study 1 experience initiation discrete fear unaware no incidental-control between NA 76.67 during published 28 0.52
Davey et al. (2013) Study 1 experience initiation discrete fear unaware no incidental-control within NA 76.67 during published 14 0.13
Davey et al. (2013) Study 1 experience initiation discrete disgust unaware no incidental-control between NA 76.67 during published 28 0.69
Davey et al. (2013) Study 1 experience initiation discrete disgust unaware no incidental-control within NA 76.67 during published 14 0.42
Davey et al. (2013) Study 1 experience initiation discrete sadness unaware no incidental-control between NA 76.67 during published 28 0.35
Davey et al. (2013) Study 1 experience initiation discrete sadness unaware no incidental-control within NA 76.67 during published 14 0.14
Davey et al. (2013) Study 2 experience initiation discrete anger unaware no incidental-control between NA 80.65 during published 29 0.73
Davey et al. (2013) Study 2 experience initiation discrete anger unaware no incidental-control within NA 80.65 during published 15 0.63
Davey et al. (2013) Study 2 experience initiation discrete fear unaware no incidental-control between NA 80.65 during published 29 0.4

COLES, LARSEN, AND LENCH 53
Davey et al. (2013) Study 2 experience initiation discrete fear unaware no incidental-control within NA 80.65 during published 15 0
Davey et al. (2013) Study 2 experience initiation discrete disgust unaware no incidental-control between NA 80.65 during published 29 0.08
Davey et al. (2013) Study 2 experience initiation discrete disgust unaware no incidental-control within NA 80.65 during published 15 -0.25
Davey et al. (2013) Study 2 experience initiation discrete sadness unaware no incidental-control between NA 80.65 during published 29 0.03
Davey et al. (2013) Study 2 experience initiation discrete sadness unaware no incidental-control within NA 80.65 during published 15 -0.06
Davis (2008) Study 1 experience modulation dimensional positivity aware yes suppress-control between film 64.29 after unpublished 28 0.99
Davis (2008) Study 1 experience modulation dimensional negativity aware yes suppress-control between film 64.29 after unpublished 28 0.87
Davis (2008) Study 2 experience modulation dimensional positivity NA yes exp.pose-suppress between film 52.17 after unpublished 31 0.26
Davis (2008) Study 2 experience modulation dimensional negativity NA yes exp.pose-suppress between film 52.17 after unpublished 30 -0.19
Davis et al. (2009) experience modulation dimensional positivity aware no suppress-control between film 63.43 after published 69 0.07
Davis et al. (2009) experience modulation dimensional negativity aware no suppress-control between film 63.43 after published 69 0.51
Davis et al. (2010) experience modulation dimensional negativity NA no Botox-control between f ilm 100 during published 68 0.1
Davis et al. (2010) experience modulation dimensional positivity NA no Botox-control between film 100 during published 68 0.05
Davis et al. (2010) experience modulation dimensional positivity NA no Botox-control between film 100 during published 68 -0.15
Davis et al. (2015) judgment - - - NA no NA within NA 55.56 - published 18 -0.16
Demaree et al. (2004) experience modulation dimensional positivity aware yes exaggerate-control between film 49.51 after published 53 0.62
Demaree et al. (2004) experience modulation dimensional negativity aware yes exaggerate-control between film 49.51 after published 50 0.16
Demaree et al. (2006) experience modulation dimensional negativity aware yes exaggerate-control between film 52.17 after published 32 -0.64
Demaree et al. (2006) experience modulation dimensional negativity aware yes suppress-control between film 52.17 after published 35 0.06
Demaree et al. (2006) experience modulation dimensional negativity NA yes suppress-
exaggerate between film 52.17 after published 37 -0.38
Dillon et al. (2007) experience modulation dimensional negativity aware no suppress-control within pictures 50 after published 36 0.11
Dimberg & Söderkvist (2011) Study 1 experience initiation dimensional - NA no incidental-incidental within NA 50 during published 48 0.51
Dimberg & Söderkvist (2011) Study 2 experience modulation dimensional positivity NA no incidental-incidental within pictures 50 after published 96 0.1
Dimberg & Söderkvist (2011) Study 2 experience modulation dimensional negativity NA no incidental-incidental within pictures 50 after published 96 0.32
Dimberg & Söderkvist (2011) Study 3 experience initiation dimensional - NA no incidental-incidental within NA 50.82 during published 61 0.06
Dimberg & Söderkvist (2011) Study 3 experience modulation dimensional positivity NA no incidental-incidental within pictures 50.82 during published 61 0.31
Dimberg & Söderkvist (2011) Study 3 experience modulation dimensional negativity NA no incidental-incidental within pictures 50.82 during published 61 0.34
Duncan & Laird (1977) experience initiation discrete anger aware yes exp.pose-control within NA 57.5 after published 31 0.44
Duncan & Laird (1977) experience initiation discrete happiness aware yes exp.pose-control within NA 57.5 after published 31 0.38
Duncan & Laird (1977) experience initiation discrete happiness aware yes exp.pose-control within NA 57.5 after published 31 0.51
Duncan & Laird (1980) experience initiation dimensional negativity aware no exp.pose-control within NA - after published 60 0.59
Duncan & Laird (1980) experience initiation dimensional positivity aware no exp.pose-control within NA - after published 60 0.44
Dzokoto et al. (2014) judgment - - - unaware no incidental-control between pictures 56.65 during published 70 1.02
Dzokoto et al. (2014) judgment - - - unaware no incidental-control between pictures 56.65 during published 59 0.07
Dzokoto et al. (2014) judgment - - - NA no incidental-suppress between pictures 56.65 during published 35 1.07

COLES, LARSEN, AND LENCH 54
Dzokoto et al. (2014) judgment - - - NA no incidental-suppress between pictures 56.65 during published 51 0.2
Flack, Laird & Cavallaro (1999b) Study 1 experience initiation discrete anger NA yes exp.pose-exp.pose within NA 73.33 after published 60 1.2
Flack, Laird & Cavallaro (1999b) Study 1 experience initiation discrete disgust NA yes exp.pose- exp.pose within NA 73.33 after published 60 0.7
Flack, Laird & Cavallaro (1999b) Study 1 experience initiation discrete fear NA yes exp.pose-exp.pose within NA 73.33 after published 60 0.31
Flack, Laird & Cavallaro (1999b) Study 1 experience initiation discrete happiness NA yes exp.pose-exp.pose within NA 73.33 after published 60 0.86
Flack, Laird & Cavallaro (1999b) Study 1 experience initiation discrete sadness NA yes exp.pose-exp.pose within NA 73.33 after published 60 1.31
Flack, Laird & Cavallaro (1999b) Study 2 experience initiation discrete anger NA yes exp.pose-exp.pose within NA 0 after published 29 0.39
Flack, Laird & Cavallaro (1999b) Study 2 experience initiation discrete disgust NA yes exp.pose- exp.pose within NA 0 after published 29 0.23
Flack, Laird & Cavallaro (1999b) Study 2 experience initiation discrete fear NA yes exp.pose-exp.pose within NA 0 after published 29 -0.16
Flack, Laird & Cavallaro (1999b) Study 2 experience initiation discrete happiness NA yes exp.pose-exp.pose within NA 0 after published 29 -0.49
Flack, Laird & Cavallaro (1999b) Study 2 experience initiation discrete sadness NA yes exp.pose-exp.pose within NA 0 after published 29 0.25
Flack, Laird & Cavallaro (1999a) experience initiation discrete anger NA yes exp.pose-exp.pose within NA 33.33 after published 54 1.41
Flack, Laird & Cavallaro (1999a) experience initiation discrete fear NA yes exp.pose-exp.pose within NA 33.33 after published 54 0.29
Flack, Laird & Cavallaro (1999a) experience initiation discrete happiness NA yes exp.pose-exp.pose within NA 33.33 after published 54 1.18
Flack, Laird & Cavallaro (1999a) experience initiation discrete sadness NA yes exp.pose-exp.pose within NA 33.33 after published 54 1.21
Flack (2006) experience initiation discrete anger NA yes exp.pose-exp.pose within NA 61.54 after published 51 0.72
Flack (2006) experience initiation discrete fear NA yes exp.pose-exp.pose within NA 61.54 after published 51 0.35
Flack (2006) experience initiation discrete happiness NA yes exp.pose-exp.pose within NA 61.54 after published 51 0.59
Flack (2006) experience initiation discrete sadness NA yes exp.pose-exp.pose within NA 61.54 after published 51 0.68
Gan et al. (2015) experience modulation dimensional negativity aware yes suppress-control with in pictures 100 after published 34 -0.11
Goldin et al. (2008) experience modulation dimensional negativity aware yes suppress-control within film 100 after published 17 0.8
Gross & Levenson (1993) experience modulation discrete disgust aware yes suppress-control between film 49.41 after published 85 0.04
Gross & Levenson (1997) experience modulation discrete happiness aware yes suppress-control between film 100 after published 180 0.37
Gross & Levenson (1997) experience modulation discrete sadness aware yes suppress-control between film 100 after published 180 0.16
Gross (1993) experience modulation discrete happiness aware yes suppress-control between film 100 after unpublished 180 0.37
Gross (1993) experience modulation discrete happiness aware yes suppress-control between film 100 after unpublished 180 0.09
Gross (1993) experience modulation dimensional positivity aware yes suppress-control between film 100 after unpublished 180 0.2
Gross (1993) experience modulation discrete sadness aware yes suppress-control between film 100 after unpublished 180 0.16
Gross (1993) experience modulation dimensional positivity aware yes suppress-control between film 100 after unpublished 180 -0.23
Gross (1998) experience modulation discrete disgust aware yes suppress-control between film 100 after published 80 0.18
Harris (2001) experience modulation discrete - aware yes suppress-control between context 58.33 after published 36 0.07
Hawk et al. (2012) experience modulation discrete disgust aware no suppress-control between audio 85.5 after published 41 0.85
Helt & Fein (2016) experience modulation dimensional positivity unaware NA incidental-control within film 16.28 - published 43 0.42
Hendricks & Buchanan (2016) experience modulation dimensional negativity aw are NA suppress-control within pictures 56.96 after published 79 -0.08
Hendricks (2013) experience modulation dimensional negativity aware NA suppress-control within pictures 56.96 after unpublished 79 0.02

COLES, LARSEN, AND LENCH 55
Henry et al. (2007) experience modulation dimensional positivity aware yes exa ggerate-control within film 53.33 - published 30 -0.49
Henry et al. (2007) experience modulation dimensional positivity aware yes suppress-control within film 53.33 - pu blished 30 0.25
Henry et al. (2009)a experience modulation dimensional positivity aware yes exaggerate-control within film 66.67 - published 26 -0.05
Henry et al. (2009)a experience modulation dimensional positivity aware yes suppress-control within film 66.67 - pu blished 26 0.53
Henry et al. (2009)b experience modulation dimensional positivity aware yes exaggerate-control within film 65 - published 20 -0.05
Henry et al. (2009)b experience modulation dimensional positivity aware yes suppress-control within film 65 - published 20 0.48
Hess et al. (1992) experience initiation discrete anger aware no exaggerate-control within NA 100 after pu blished 28 -0.28
Hess et al. (1992) experience initiation discrete happiness aware no exaggerate-control within NA 100 after published 28 0.14
Hess et al. (1992) experience initiation discrete happiness aware no exaggerate-control within NA 100 after published 28 -0.26
Hess et al. (1992) experience initiation discrete sadness aware no exaggerate-control within NA 100 after published 28 -0.16
Hofmann et al. (2009) experience modulation discrete fear aware yes suppress-control between context - - published 134 -0.03
Ito et al. (2006) experience initiation dimensional positivity unaware no incidental-control within NA - after published 40 -0.39
Ito et al. (2006) experience initiation dimensional positivity unaware no incidental-control between NA - after published 33 -0.25
Kalokerinos et al. (2015) Study 1 experience modulation discrete happiness aware no suppress-control between film 50 after published 133.67b-0.06
Kalokerinos et al. (2015) Study 1 experience modulation discrete sadness aware no suppress-control between film 50 after published 133.67b-0.02
Kalokerinos et al. (2015) Study 2 experience modulation discrete happiness aware no suppress-control between film 43 after published 295 1.32
Kalokerinos et al. (2015) Study 2 experience modulation discrete sadness aware no suppress-control between film 43 after published 295 0.2
Kao et al. (2017) experience modulation discrete anger aware no exa ggerate-control between context 50.41 after published 41 0.09
Kao et al. (2017) experience modulation discrete anger aware no exaggerate-control between context 50.41 after published 41 -0.39
Kao et al. (2017) experience modulation discrete anger aware no suppress-control between context 50.41 after published 41 0.8
Kao et al. (2017) experience modulation discrete anger aware no suppress-control between context 50.41 after published 41 -0.34
Kao et al. (2017) experience modulation discrete anger NA no suppress-
exaggerate between context 50.41 after published 41 0.98
Kao et al. (2017) experience modulation discrete anger NA no suppress-
exaggerate between context 50.41 after published 41 -0.67
Kircher et al. (2012) experience modulation discrete happiness aware yes exp.pose-control within pictures 53.13 after published 27 1.89
Kircher et al. (2012) experience modulation discrete happiness aware yes exp.pose-control within pictures 53.13 after published 27 1.14
Korb et al. (2012) experience modulation discrete happiness aware no suppress-control within pictures 100 - published 22 0.21
Labott & Teleha (1996) experience modulation dimensional negativity NA no suppress-
exaggerate between film 100 after pu blished 19 0.04
Labott & Teleha (1996) experience modulation dimensional negativity NA no suppress-
exaggerate between film 100 after pu blished 16 0.91
Laird (1974) Study 1 experience modulation discrete anger NA no exp.pose-exp.pose within pictures - - published 38 0.46
Laird (1974) Study 1 experience modulation discrete happiness NA no exp.pose-exp.pose within pictures - - published 38 0.44
Laird (1974) Study 1 experience modulation discrete happiness NA no exp.pose-exp.pose within pictures - - published 38 0.39
Laird (1974) Study 2 judgment - - - NA no exp.pose-exp.pose within pictures - - published 26 0.55
Laird (1974) Study 2 experience - discrete happiness NA no exp.pose-exp.pose within pictures - - published 26 0.13
Lalot et al. (2014) judgment - - - aware yes suppress-control within film 66.67 after published 45 -0.17
Larsen et al. (1992) experience modulation discrete sadness NA no incidental-suppress within pictures 30 during published 27 0.43

COLES, LARSEN, AND LENCH 56
Lee (2011) experience modulation discrete disgust aware yes exaggerate-control within film 54.17 after unpublished 52 0.48
Lee (2011) experience modulation discrete disgust aware yes exaggerate-control within film 54.17 after unpublished 44 0.17
Lee (2011) experience modulation discrete disgust aware yes suppress-control within film 54.17 after unpublished 52 -0.27
Lee (2011) experience modulation discrete disgust aware yes suppress-control within film 54.17 after unpublished 44 -0.26
Lewis & Bowler (2009) experience modulation dimensional negativity NA no Botox-control between NA 100 during published 25 1.35
Lewis (2012) judgment - - - NA no exp.pose-exp.pose within sentences 100 during published 24 0.71
Lewis (2012) judgment - - - NA no exp.pose- suppress within sentences 100 during published 24 0.56
Ma (2011) experience modulation discrete fear aware yes suppress-control between film 23.44 - unpublished 42.67b-0.21
Ma (2011) experience modulation discrete disgust aware yes suppress-control between film 23.44 - unpublished 42.67b-0.21
Ma (2011) experience modulation discrete sadness aware yes suppress-control between film 23.44 - unpublished 42.67b-0.21
Ma (2011) experience modulation discrete happiness aware yes suppress-control between film 23.44 - unpublished 42.67b-0.21
Maldonado et al. (2015) experience modulation discrete anger aware NA suppress-control between stories 58.47 after unpublished 157.33b0.12
Marmolejo-Ramos & Dunn (2013) Study 1 experience initiation dimensional positivity unaware no incidental-control within NA 78.85 - published 100 -0.07
Marmolejo-Ramos & Dunn (2013) Study 2 experience initiation dimensional positivity unaware no incidental-control within NA 75.47 - published 106 -0.07
Marmolejo-Ramos & Dunn (2013) Study 3 experience modulation dimensional positivity NA no incidental-suppress within pictures 73.08 - published 104 -0.07
Marmolejo-Ramos & Dunn (2013) Study 4 experience initiation dimensional positivity unaware no incidental-control within NA 63 - published 100 -0.07
Marmolejo-Ramos & Dunn (2013) Study 5 experience initiation dimensional positivity unaware no incidental-control within NA 71.21 - published 66 0.27
Marmolejo-Ramos & Dunn (2013) Study 6 experience initiation dimensional positivity unaware no incidental-control within NA 61.19 - published 67 0.38
Martijn et al. (2002) experience modulation dimensional negativity aware no suppress-control between film 86.79 after published 33 -0.24
McCanne & Anderson (1987) experience modulation dimensional positivity aware yes exaggerate-control within context 100 after published 30 -2.16
McCanne & Anderson (1987) experience modulation dimensional negativity aware yes exaggerate-control within imaginedscenarios 100 after published 30 -2.07
McCanne & Anderson (1987) experience modulation dimensional positivity aware yes suppress-control within imaginedscenarios 100 after published 30 4.73
McCanne & Anderson (1987) experience modulation dimensional negativity aware yes suppress-control within imaginedscenarios 100 after published 30 1.67
McCanne & Anderson (1987) experience modulation dimensional positivity NA yes suppress-
exaggerate within imaginedscenarios 100 after published 30 2.48
McCanne & Anderson (1987) experience modulation dimensional negativity NA yes suppress-
exaggerate within imaginedscenarios 100 after published 30 -0.25
McCaul et al. (1982) experience initiation discrete f ear aware yes exp.pose-control within NA 55.56 after published 27 0.25
McIntosh et al. (1997) experience initiation dimensional negativity NA no incidental-incidental within NA 50 after published 26 0.54
Meeten et al. (2015) judgment - - - NA no exp.pose-exp.pose within pictures 76.06 after published 71 0.49
Miyamoto (2006) Study 1 judgment - - - NA no incidental-suppress between pictures 24.69 during unpublished 40 0.17
Miyamoto (2006) Study 1 judgment - - - NA no incidental-suppress between pictures 24.69 during unpublished 40 0.53
Miyamoto (2006) Study 2 judgment - - - NA no incidental-suppress between pictures 60 during unpublished 77 0.49
Moore & Zoellner(2012) experience modulation dimensional negativity aware no suppress-control between film - after published 23.33b-0.87
Kappas (1989) judgment - - - aware no exaggerate-control within film 43.75 - unpublished 32 0.08
Kappas (1989) judgment - - - aware no exaggerate-control within film 43.75 - unpublished 32 0.26
Kappas (1989) judgment - - - aware no suppress-control within film 43.75 - unpublished 32 0.27

COLES, LARSEN, AND LENCH 57
Kappas (1989) judgment - - - aware no suppress-control within film 43.75 - unpublished 32 0.1
Kappas (1989) experience modulation discrete disgust aware no exaggerate-control within film 43.75 - unpublished 32 0.17
Kappas (1989) experience modulation discrete happiness aware no exa ggerate-control within film 43.75 - unpublished 32 0.52
Kappas (1989) experience initiation discrete disgust NA no exp.pose-exp.pose within NA 43.75 - unpublished 32 0.62
Kappas (1989) experience initiation discrete happiness NA no exp.pose-exp.pose within NA 43.75 - unpublished 32 0.74
Kappas (1989) experience modulation discrete disgust aware no suppress-control within film 43.75 - unpublished 32 0.18
Kappas (1989) experience modulation discrete happiness aware no suppress-control within film 43.75 - unpublished 32 0.42
Ohira & Kurono (1993) Study 1 judgment - - - aware no exaggerate-control between context 100 after published 20 1.23
Ohira & Kurono (1993) Study 1 judgment - - - aware no suppress-control between context 100 after published 20 0.31
Ohira & Kurono (1993) Study 2 judgment - - - aware no exaggerate-control between context 100 after published 20 1.61
Ohira & Kurono (1993) Study 2 judgment - - - aware no suppress-control between context 100 after published 20 -1.38
Paredes et al. (2013) judgment - - - NA no incidental-suppress between stories - - published 31 0.85
Paul et al. (2013) judgment - - - aware no suppress-control within pictures 50 - published 20 0.91
Pedder et al. (2016) experience modulation dimensional positivity aware NA suppress-control within pictures 64.29 after published 68 0.7
Pedder et al. (2016) experience modulation dimensional negativity aware NA suppress-control within pictures 64.29 after published 68 0.22
Phillips et al. (2008) experience modulation dimensional negativity aware yes suppress-control between film 54.7 after published 32 0.18
Phillips et al. (2008) experience modulation dimensional negativity aware yes suppress-control between film 54.7 after published 32 0.08
Reisenzein & Studtmann (2007) Study 1 experience initiation discrete surprise aw are no exp.pose-control between NA 61.25 during published 53 0.18
Reisenzein & Studtmann (2007) Study 1 experience initiation discrete surprise aware no exp.pose-control between NA 61.25 during published 55 0.34
Reisenzein & Studtmann (2007) Study 1 experience modulation discrete surprise aware no exp.pose-control between pictures 61.25 during published 55 -0.08
Reisenzein & Studtmann (2007) Study 1 experience modulation discrete surprise aware no exp.pose-control between pictures 61.25 during published 55 0.3
Reisenzein & Studtmann (2007) Study 1 experience modulation discrete surprise NA no exp.pose-suppress between pictures 61.25 during published 53 -0.12
Reisenzein & Studtmann (2007) Study 1 experience modulation discrete surprise NA no exp.pose-suppress between pictures 61.25 during published 53 0.22
Reisenzein & Studtmann (2007) Study 1 experience modulation discrete surprise aware no suppress-control between pictures 61.25 during published 52 -0.04
Reisenzein & Studtmann (2007) Study 1 experience modulation discrete surprise aware no suppress-control between pictures 61.25 during published 52 -0.09
Reisenzein & Studtmann (2007) Study 3 experience modulation discrete surprise aware no exp.pose-control between pictures 50 after published 40 -0.74
Richards, Butler & Gross (2003) experience modulation dimensional positivity aware no suppress-control between context 50 after published 59 0.19
Richards, Butler & Gross (2003) experience modulation dimensional negativity aware no suppress-control between context 50 after published 59 -0.12
Richards & Gross (1999) Study 1 experience modulation dimensional negativity aware NA suppress-control between pictures 100 after published 58 -0.1
Richards & Gross (1999) Study 1 experience modulation dimensional negativity aware NA suppress-control between pictures 100 after published 58 0.25
Richards & Gross (1999) Study 1 experience modulation dimensional negativity aware NA suppress-control between pictures 100 after published 58 0.36
Richards & Gross (1999) Study 2 experience modulation dimensional negativity aware NA suppress-control between pictures 100 after published 85 0.13
Richards & Gross (1999) Study 2 experience modulation dimensional negativity aware NA suppress-control between pictures 100 after published 85 0.24
Richards & Gross (1999) Study 2 experience modulation dimensional negativity aware NA suppress-control between pictures 100 after published 85 0.06
Richards & Gross (2000) Study 1 experience modulation dimensional negativity aware no suppress-control between film 55 after published 53 -0.12

COLES, LARSEN, AND LENCH 58
Richards & Gross (2000) Study 2 experience modulation dimensional negativity aware no suppress-control between pictures 100 after published 61 0.39
Richards & Gross (2006) experience modulation dimensional negativity aware no suppress-control between film 65 after published 131 0.34
Roberts et al. (2008) experience modulation discrete disgust aware no suppress-control between film 60 after published 160 0.07
Robinson & Demaree (2009) experience modulation dimensional negativity aware NA exaggerate-control within film 50.98 after published 102 -0.04
Robinson & Demaree (2009) experience modulation discrete sadness aware NA exaggerate-control within film 50.98 after published 102 0.03
Robinson & Demaree (2009) experience modulation dimensional negativity aware NA suppress-control within film 50.98 after published 102 0
Robinson & Demaree (2009) experience modulation discrete sadness aware NA suppress-control within film 50.98 after published 102 0
Roemer (2014) experience modulation dimensional positivity unaware yes incidental-control between film 81.82 after unpublished 44 0.58
Roemer (2014) experience modulation dimensional positivity unaware yes incidental-control between film 81.82 after unpublished 44 0.29
Rohrmann et al. (2009) experience modulation discrete disgust aware NA suppress-control between film 0 after pu blished 36 0.16
Rohrmann et al. (2009) experience modulation discrete disgust aware NA suppress-control between film 0 after published 36 0.13
Rummer et al. (2014) judgment - - - NA no incidental-incidental between pictures - during published 74 0.57
Rummer et al. (2014) judgment - - - NA no incidental-suppress between pictures - during published 74 0.46
Schmeichel , Vohs, & Baumeister (2003) experience modulation dimensional negativity aware yes suppress-control between film 59.46 after published 37 -0.23
Schmeichel et al. (2008) experience modulation dimensional positivity aware NA suppress-control between film 62 during published 50 0.1
Söderkvist & Dimberg (unpublished) experience modulation dimensional - NA no incidental-incidental within pictures 50 during unpublished 32 0.36
Söderkvist et al. (2018) Study 1aexperience modulation dimensional - NA no incidental-incidental within pictures 50 during unpublished 32 0.34
Söderkvist et al. (2018) Study 2aexperience modulation dimensional - NA no incidental-incidental within pictures 50 during unpublished 64 0.17
Soussignan (2002) experience modulation dimensional positivity NA yes incidental-suppress between film 100 after published 33 -0.17
Soussignan (2002) experience modulation dimensional positivity NA yes incidental-suppress between film 100 after published 33 0.48
Soussignan (2002) experience modulation dimensional positivity NA yes incidental-suppress between film 100 after published 33 0.47
Soussignan (2002) experience modulation dimensional positivity NA yes incidental-suppress between film 100 after published 33 0.44
Soussignan (2002) experience modulation dimensional positivity NA yes incidental-suppress between film 100 after published 32 0.53
Soussignan (2002) experience modulation dimensional positivity NA yes incidental-suppress between film 100 after published 32 1.1
Soussignan (2002) experience modulation dimensional positivity NA yes incidental-suppress between film 100 after published 32 1.11
Soussignan (2002) experience modulation dimensional positivity NA yes incidental-suppress between film 100 after published 32 0.94
Stel et al. (2008) Study 2 experience initiation dimensional positivity NA no NA between NA - after published 18.67b1.11
Stel et al. (2008) Study 3 judgment - - - unaware no incidental-control between pictures - during published 24 1
Strack et al. (1988) Study 1 judgment - - - NA no incidental-suppress between pictures - during published 76.67b0.43
Strack et al. (1988) Study 2 judgment - - - NA no incidental-suppress between pictures 45.78 during published 83 -0.15
Strack et al. (1988) Study 2 experience modulation discrete happiness NA no incidental-suppress between pictures 45.78 during published 41.5 0.55
Strack et al. (1988) Study 2 experience modulation discrete happiness NA no incidental-suppress between pictures 45.78 during published 41.5 -0.51
Tamir et al. (2004) experience modulation dimensional positivity NA no exp.pose-exp.pose between pictures - after published 72 -0.16
Tourangeau & Ellsworth (1979) experience initiation discrete fear aware yes exp.pose-control between NA - after published 20.5b0.3
Tourangeau & Ellsworth (1979) experience initiation discrete sadness aware yes exp.pose-control between NA - after published 20.5b0.3

COLES, LARSEN, AND LENCH 59
Tourangeau & Ellsworth (1979) experience modulation discrete fear aware yes exp.pose-control between film - after published 20.5b0.3
Tourangeau & Ellsworth (1979) experience modulation discrete sadness aware yes exp.pose-control between film - after published 20.5b0.3
Trent (2010) judgment - - - unaware no incidental-control between pictures 74.07 after unpublished 107.33b-0.22
Trent (2010) judgment - - - NA no incidental-suppress between pictures 74.07 after unpublished 107.33b-0.22
Trent (2010) experience modulation dimensional positivity unaware no incidental-control between pictures 74.07 after unpublished 107.33b-0.06
Trent (2010) experience modulation dimensional positivity NA no incidental-suppress between pictures 74.07 after unpublished 107.33b-0.06
Vieillard et al. (2015) experience modulation dimensional positivity aware no exaggerate-control within audio 59.02 after published 31 0.25
Vieillard et al. (2015) experience modulation dimensional negativity aware no exaggerate-control within audio 59.02 after published 31 0.66
Vieillard et al. (2015) experience modulation dimensional positivity aware no exaggerate-control within audio 59.02 after published 30 0.21
Vieillard et al. (2015) experience modulation dimensional negativity aware no exaggerate-control within audio 59.02 after published 30 0.14
Vieillard et al. (2015) experience modulation dimensional positivity aware no suppress-control within audio 59.02 after published 31 -0.05
Vieillard et al. (2015) experience modulation dimensional negativity aware no suppress-control within audio 59.02 after published 31 -0.5
Vieillard et al. (2015) experience modulation dimensional positivity aware no suppress-control within audio 59.02 after published 30 0.07
Vieillard et al. (2015) experience modulation dimensional negativity aware no suppress-control within audio 59.02 after published 30 -0.12
Wagenmakers et al. (2016) Albohn site experience modulation discrete happiness NA yes incidental-suppress between pictures 55.21 during published 139 0.09
Wagenmakers et al. (2016) Allard site experience modulation discrete happiness NA yes incidental-suppress between pictures 74.85 during published 125 0.09
Wagenmakers et al. (2016) Benning site experience modulation discrete happiness NA yes incidental-suppress between pictures 58.04 during published 115 -0.01
Wagenmakers et al. (2016) Bulnes site experience modulation discrete happiness NA yes incidental-suppress between pictures 83.33 during published 101 0.09
Wagenmakers et al. (2016) Capaldi site experience modulation discrete happiness NA yes incidental-suppress between pictures 69.33 during published 117 -0.07
Wagenmakers et al. (2016) Chasten site experience modulation discrete happiness NA yes incidental-suppress between pictures 70.37 during published 94 -0.04
Wagenmakers et al. (2016) Holmes site experience modulation discrete happiness NA yes incidental-suppress between pictures 63.98 during published 99 0.15
Wagenmakers et al. (2016) Koch site experience modulation discrete happiness NA yes incidental-suppress between pictures 66.38 during published 100 -0.14
Wagenmakers et al. (2016) Korb site experience modulation discrete happiness NA yes incidental-suppress between pictures 37.93 during published 101 0.01
Wagenmakers et al. (2016) Lynott site experience modulation discrete happiness NA yes incidental-suppress between pictures 38.61 during published 126 0.23
Wagenmakers et al. (2016) Oosterwijk site experience modulation discrete happiness NA yes incidental-suppress between pictures 30.2 during published 110 -0.17
Wagenmakers et al. (2016) Ozdogru site experience modulation discrete happiness NA yes incidental-suppress between pictures 35.03 during published 87 -0.3
Wagenmakers et al. (2016) Pacheco-Unguetti
site experience modulation discrete happiness NA yes incidental-suppress between pictures 24.32 during published 120 -0.08
Wagenmakers et al. (2016) Talarico site experience modulation discrete happiness NA yes incidental-suppress between pictures 23.27 during published 112 0.02
Wagenmakers et al. (2016) Wagenmakers site experience modulation discrete happiness NA yes incidental-suppress between pictures 37.02 during published 130 0.13
Wagenmakers et al. (2016) Wayand site experience modulation discrete happiness NA yes incidental-suppress between pictures 18 during published 110 -0.14
Wagenmakers et al. (2016) Zeelenberg site experience modulation discrete happiness NA yes incidental-suppress between pictures 22.76 during published 108 0.25
Wittmer (1985) experience modulation discrete fear aware yes exaggerate-control within film 0 - unpublished 30 -0.36
Wittmer (1985) experience modulation discrete f ear aware yes suppress-control within film 0 - unpublished 30 -0.21
Yartz (2004) experience modulation discrete disgust aware yes suppress-control within pictures 41.38 - unpublished 28 -0.05
Yartz (2004) experience modulation discrete disgust aware yes suppress-control within pictures 41.38 - unpublished 30 -0.18

COLES, LARSEN, AND LENCH 60
Yartz (2004) experience modulation dimensional negativity aware yes suppress-control within pictures 41.38 - unpublished 28 -0.08
Yartz (2004) experience modulation dimensional negativity aware yes suppress-control within pictures 41.38 - unpublished 30 -0.09
Yartz (2004) experience modulation dimensional positivity aware yes suppress-control within pictures 41.38 - unpublished 28 0.04
Yartz (2004) experience modulation dimensional positivity aware yes suppress-control within pictures 41.38 - unpublished 30 0.5
Zajonc et al. (1989) Study 3 judgment - - - NA no incidental-incidental within audio - after published 37 1.27
Zajonc et al. (1989) Study 4 experience initiation dimensional - NA NA incidental-incidental within NA 0 after pu blished 26 0.47
Zajonc et al. (1989) Study 4 experience initiation dimensional - NA NA incidental-incidental within NA 0 after published 26 0.31
Zariffa et al. (2014) experience initiation dimensional positivity aware no exp.pose-control within NA 50 after published 24 -0.57
Zariffa et al. (2014) experience initiation dimensional positivity aware no exp.pose-control within NA 50 after published 24 -0.14
Zhu et al. (2015) experience initiation discrete disgust NA yes exp.pose-exp.pose between NA 74.55 after published 55 1.74
Note. A more detailed data file is available on the Open Science Framework; N = total sample size for two-group comparison; d = Cohen’s standardized difference
a Results were unpublished at time of meta-analysis but are now published in Söderkvist et al. (2018).
b Estimated sample size.

COLES, LARSEN, AND LENCH 61
Author Note
Nicholas A. Coles and Jeff T. Larsen, Department of Psychology, University of Tennessee;
Heather C. Lench, Department of Psychological and Brain Sciences, Texas A&M University.
This material is based upon work supported by the National Science Foundation Graduate
Research Fellowship #R010138018 awarded to Nicholas A. Coles.
*Correspondence concerning this article should be addressed to Nicholas A. Coles, Department
of Psychology, Austin Peay Building, University of Tennessee, Knoxville, TN 37996. Email:
colesn@vols.utk.edu
OSF link: https://osf.io/v8kxb/