Nicole Kushner Thesis

advertisement
Running Header: Social Anxiety Moderates SEF and Eye Gaze
Social Anxiety as a Moderator in the Relationship between Social Emotional Fluency and Eye
Gaze
Nicole Kushner
Vanderbilt University
1
Social Anxiety Moderates SEF and Eye Gaze
2
Abstract
Emotional intelligence and interpersonal sensitivity have been identified as key
individual-difference abilities that are important for optimal social functioning. Social emotional
fluency (SEF) is proposed as a behavioral component of these two constructs. SEF is focused on
qualitative aspects and temporal dynamics of nonverbal social behavior. Social anxiety has been
studied as affecting the production of many of these nonverbal behaviors, most notably eye gaze.
It was predicted that social anxiety would moderate the effect of SEF on eye gaze behavior.
Twenty-four dyads participated in a trip-planning paradigm designed to elicit a moderate amount
of social anxiety. It was predicted that individuals with high SEF/low social anxiety would
display the highest amounts of eye contact and that those with low SEF/high social anxiety
would display the least. Limited support was found for social anxiety as a moderator of SEF on
eye gaze behavior. Future research should investigate finding more support for social anxiety as
a moderator of SEF on eye gaze behavior, as well as investigating whether the moderating
effects extend to the production of other nonverbal behaviors.
Social Anxiety Moderates SEF and Eye Gaze
3
Eye gaze is considered to be an essential component of nonverbal behavior in social
interactions. Observed in human interactions, eye gaze is an important tool that is variously
important for conveying social dominance, affective states, and behavior indicators such as
aggressive behavior. (Kleinke, 1986; Terburg, Aarts, & van Honk, 2012). Eye gaze is also
commonly associated with disorders that involve significant social impairment such as social
anxiety disorder (Terburg et al. 2012). Although many studies have examined the impairment
that social anxiety exerts on eye gaze and social interactions (Langer & Rodebaugh, 2013;
Terburg et al. 2012), the exact nature of the relations among social anxiety, eye gaze, and social
functioning remains unspecified. The current study tests the hypothesis that social anxiety
moderates the relationship between eye gaze and “social-emotional fluency” (SEF).
The purpose of the present experiment is to investigate the developing construct of SEF
and its association with eye-gaze behavior and social anxiety. Social-emotional fluency, or SEF,
is a construct concerned with the qualitative and temporal dynamics of nonverbal behavior. As a
behavioral measure, it is likely that SEF may be a contributing construct in both interpersonal
sensitivity (IS) and emotional intelligence (EI). IS concerns the accurate decoding and
interpretation of nonverbal cues and EI encompasses a wide array of emotional capabilities
(Carney & Harrigan, 2003; Lopes, Brackett, Nezlek, Schutz, Sellin, & Salovey, 2004). SEF is
concerned primarily with the production and occurrence of nonverbal behavioral cues that
enhance or detract from a social interaction, although accuracy in detection of others nonverbal
cues is required to respond appropriately. These cues include but are not limited to body posture,
head movement and orientation, smiling, laughing, eye gaze and eye contact, and vocal
regulators. It is not expected that individuals high in SEF necessarily exhibit more of these
behaviors, but rather that they produce behavioral cues of higher quality and with more effective
Social Anxiety Moderates SEF and Eye Gaze
4
social timing. For example, it would be expected that a high SEF person utilizes non-Duchenne
smiles more effectively than a low SEF person to promote ease and establish rapport in an
interaction. It is also expected that individuals high in SEF would smile at optimal moments in
the social interaction, resulting in the greatest amount of positive social feedback. This positive
social feedback would be expected to improve the establishment of rapport and overall quality of
the interaction.
A related construct that is important to considering the quality of social interactions is
that of “emotional intelligence” (EI). Characterizations of EI evolved from the idea that emotionrelated competencies are key for adaptation in various realms of life (Lopes et al., 2004). Mayer
and colleagues (2008) conceptualized EI as the ability to engage in sophisticated information
processing about one’s own and other’s emotions and the ability to use this information as a
guide to thinking and behavior. EI has been posited to consist of four abilities: identifying
emotions, using emotions, understanding emotions, and managing emotions (Kassymzhanova &
Mun, 2013). Of these four constructs, identifying emotions and managing emotions are the most
relevant to SEF. The first, identifying emotions, is described as the ability to identify and
appropriately express emotions (Kassymzhanova & Mun, 2013). The ability to distinguish
between artificial and genuine emotion also falls within this category. Managing emotions, also
known as emotion regulation, involves reducing the intensity of negative emotions, cognizance
of emotions, and being capable of solving emotionally laden problems without suppressing
negative consequences (Kassymzhanova & Mun, 2013).
Lopes and colleagues (2004) theorized that the fourth construct, managing emotions,
would be the facet of EI that is most strongly associated with the quality of social interactions.
One theory behind this expectation is that emotion regulation likely influences people’s
Social Anxiety Moderates SEF and Eye Gaze
5
motivations and expectations for a social interaction and the emotional valence of social
interactions. Because of the ability to simultaneously influence the motivations and emotional
valence of an interaction, Lopes and colleagues (2004) posit that emotion regulation may create a
flexible attentional focus that can affect the quality of the interaction. They also argued that
emotion regulation may promote the underlying processes associated with coordination of the
many skills required for social behavior. For example, unregulated social anxiety inhibits
spontaneity and causes behavior to become over-controlled and inhibited. This assumption was
supported when Lopes and colleagues (2004) found that the managing emotions component of
the Mayer-Salovey-Caruso Emotional Intelligence Test (MSCEIT) was positively related to the
perceived quality of social interactions. These effects were unchanged when individual
differences due to big five characteristics, gender, and self-enhancement were controlled for.
Correlations between the quality of social interactions and the other three components of
emotional intelligence were also found, however these findings were not consistent (Lopes et al.,
2004). Managing emotions appears to play an important role in the regulation of behaviors, such
as anxiety driven behaviors, which impair an individual’s functioning in a social interaction.
These anxiety driven behaviors, such as avoidance of eye-gaze, can negatively impact the quality
of the interaction.
A second construct important to SEF is interpersonal sensitivity (IS). IS is defined as the
ability to make correct judgments about the abilities, traits, and states of others from nonverbal
cues (Carney & Harrigan, 2003). According to Carney and Harrigan (2003), there are two main
components of IS: emotional sensitivity and social sensitivity. Emotional sensitivity is the ability
to accurately assess nonverbal cues associated with emotion. These cues include facial displays
of emotion and body displays of emotion. Social sensitivity focuses more on social information
Social Anxiety Moderates SEF and Eye Gaze
6
such as emotion, personality, and social role. IS levels are also associated with accuracy in
judging another person’s interpersonal sensitivity. Carney and Harrigan (2003) found that a
person’s level of emotional sensitivity predicted their accuracy in judging a friend’s emotional
sensitivity, and that social sensitivity affected a person’s accuracy in judging a friend’s social
sensitivity. These abilities did not cross domains, however: The level of emotional sensitivity did
not increase accuracy in judging social sensitivity, nor did the level of social sensitivity affect the
accuracy in judging emotional sensitivity (Carney & Harrigan, 2003).
IS has many implications for understanding differences in the quality of social
interactions. In a study of college roommate pairings, Hodgins and Zuckerman (1990) found that
high nonverbal decoding pairs experienced greater emotional sharing than pairs with one or both
low decoding participants. This finding indicates that emotional sharing is not only enhanced by
verbal communication, but also by nonverbal cues, body cues, and tone of voice (Hodgins &
Zuckerman, 1990). In addition, Hodgins and Zuckerman (1990) also reported that the lower IS
person rated the interaction as more positive than the higher IS in high-high and high-low
pairings. Good decoders are more capable of accurately interpreting nonverbal cues, and
therefore are more likely to be better able to respond and provide positive social feedback to their
partner (Rosip & Hall, 2004). Although perception of nonverbal cues is an important component
of IS, expression of these cues is also an essential component of IS.
Expression and perception are both important components of IS. Based on this
observation, the “Snodgrass paradigm” was used to measure two types of IS (Snodgrass, Hecht,
& Ploutz-Snyder, 1998). Snodgrass and colleagues (1998) devised an experiment to determine
which construct was more important to IS by instructing observers to watch supervisorsubordinate interactions. The first measure was the impression one makes of another person, and
Social Anxiety Moderates SEF and Eye Gaze
7
the second was how the other person is feeling about themselves (Snodgrass et al., 1998). When
instructed to observe the supervisor, these investigators found that observers were more sensitive
to what the supervisor thought about the subordinate than how the supervisor felt about
themselves. In contrast, when told to watch the subordinate, the observers were more sensitive
to how the subordinate felt about themselves. Snodgrass and colleagues (1998) interpreted these
results as indicating that high IS is a function of clear encoding by the sender rather than the
target’s detection. For example, high quality nonverbal cues produced by the supervisor, such as
direct eye-contact, would contribute more to IS than the subordinate’s perception of these cues.
Although perception is an important component of IS, expressivity and the production of
nonverbal behavioral cues contributed more to high IS.
Despite having conceptual points of contact, very few studies have examined the
interactions among eye-gaze, IS, and EI. In a meta-analysis, Hall and colleagues (2009) found
only a weak correlation between eye contact and IS (r=.14, z=.68). There are many approaches
that could be taken to expand this field by examining eye gaze further under the construct of
SEF. The introduction of SEF potentially allows for more a unified approach to the study of
nonverbal cues in social interaction. Investigation into any specific nonverbal behavior, such as
eye-gaze, could then provide a more complete understanding of the functional purpose of these
behaviors underlying constructs such as IS and EI.
Eye gaze in social interactions has been a well-researched topic. For the purpose of such
studies, Kleinke (1986) distinguished between eye gaze, eye contact, and averted gaze. Eye-gaze
was described as the direction of one’s gaze towards another persons’ eyes, whereas eye contact
is the mutual, direct looking at one another’s eyes between two individuals in a social
interaction. Eye contact is enhanced by face to face interactions, personalization, and minimal
Social Anxiety Moderates SEF and Eye Gaze
8
distractions during the interaction (Kleinke, 1986). Eye-contact behavior has been implicated in
many psychological phenomena. For example, Russo (1975) found greater amounts of eye
contact to be associated with higher levels of self-reported intimacy in social pairings. Among
these pairings, dyads composed of self-described friends were found to have longer mean
durations of eye contact compared to dyads of non friends (as reported by an authority figure
familiar with the participants; Russo, 1975). However, Kleinke (1986) found that people spent a
greater proportion of time using eye gaze towards strangers than friends during an interaction.
One possible explanation for this discrepancy would be that in an unfamiliar or ambiguous social
interaction, an individual uses greater amounts of eye gaze to better observe their partner’s
behavior, enabling the individual to make more accurate predictions of their partner’s intentions.
Due to the unfamiliarity between interactants in this situation, eye gaze would occur more often
but for shorter durations of time to avoid engaging in an uncomfortable level of intimacy with an
unfamiliar person. While eye contact has been found to increase intimacy in social pairings,
these findings indicate that eye gaze may also be a tool of information gathering in social
situations.
Averted gaze has been observed in the behavior of both submissive and anxious
individuals. Kleinke (1986) describes averted gaze, or eye-gaze avoidance, as the intentional
avoidance of eye contact. “Hypervigilance avoidance theory” predicts that anxious individuals
rapidly detect threat cues because they are hypervigilant of danger. After detection of the other
interactant’s eyes, the individual quickly avoids eye contact to reduce stress and prevent
confrontation (Terburg et al. 2012). This anxious-avoidant behavior is thought to differ from the
averted eye gaze behavior of submissive roles. Terburg et. al (2012) found that anxious
individuals detect, remember, and avoid social threats such as negatively valenced facial
Social Anxiety Moderates SEF and Eye Gaze
9
expressions, whereas submissive individuals detect and immediately avoid such social cues. In a
social interaction between dominant and submissive individuals, the submissive individual will
rapidly break eye contact with the dominant individual, whereas an anxious individual will avoid
making eye contact at all (Terburg et al. 2012). Anxious individuals also spend greater amounts
of time scanning the visual field than do submissive individuals (Terburg et al. 2012).
Anxious gaze avoidance, however, does not necessarily relieve anxiety, but rather is
maladaptive to relieving anxiety. In fact, Langer and Rodebaugh (2013) found that avoiding eye
contact was the most anxiety-provoking behavior for people high in social anxiety. These
authors interpreted their results as indicating that individuals with higher levels of social anxiety
orient their attention inwards, rather than focusing on external cues such as a conversation
partner’s nonverbal behavior (Kleinke, 1986; Langer & Rodebaugh, 2013). This repeated gaze
avoidance causes missed opportunities for social information such as positive feedback, and
anxiety is therefore maintained over time (Langer & Rodebaugh, 2013). Eye-contact avoidance
has long been regarded as characteristic of social anxiety, and it now appears that this is a
maladaptive behavior that only maintains a person’s anxiety over time. Rapid gaze aversion,
however, is considered to be a comparatively reflexive behavior that is not accounted for (not the
focus of) by Terburg et al.’s “hypervigilance avoidance” theory. Instead, rapid gaze aversion is
an overt social signaling behavior used to communicate subordination to a socially dominant
individual. It is not associated with the scanning of the visual field or an anxious state, but
instead it is treated as a nonverbal behavior used to avoid confrontation and communicate
submissiveness with a more dominant individual (Terburg et al. 2012). Although different in
their intentions, both of these gaze behaviors impact the quality of social interactions.
Social Anxiety Moderates SEF and Eye Gaze
10
The main question to be addressed in this study is how high-SEF individuals use eye gaze
in social interactions and how this association may be affected by social anxiety. Social anxiety
has been well-documented in its impairment of behaviors relative to smooth social functioning,
in particular eye gaze (Langer & Rodebaugh, 2013; Terburg, Aarts, & van Honk, 2012). While
it seems likely that social anxiety would impair the effect of SEF on anxious-avoidant eye gaze,
it remains unclear exactly what form this interaction could take and what the strength of this
interaction would be. We hypothesize that social anxiety will have a moderately strong
moderating effect on the interaction of social emotional fluency on eye-gaze behavior.
Most studies to date have examined the direction of a person’s gaze, which can provide
information about important areas for gathering information in the visual field (Kleinke, 1986).
Eye-gaze direction can also be indicative of certain personality indicators such as submissiveness
and social anxiety. In contrast very little is known about how frequency or duration of eye gaze
can affect the social interaction. As stated earlier, it would not necessarily be predicted that a
high SEF person would have a greater frequency of direct eye gaze. Too frequent eye-gaze
could have a diminishing effect on the power of the nonverbal signal and decrease the
effectiveness of using eye gaze as an indicator of behavior such as aggressive tendencies. For
instance, overly frequent eye-gaze could make the partner in a social interaction uncomfortable,
thereby impairing social-emotional communication during the interaction. Maintaining direct
eye-gaze for too long could also make the social partner uncomfortable and impair the social
emotional fluency. It would be predicted that a high SEF individual would attempt to maintain a
moderately high level of eye-gaze with their partner that would enable them to engage their
partner and promote rapport by using their gaze to confirm they are listening to the partner,
gather social cue information, and provide positive feedback without making the partner
Social Anxiety Moderates SEF and Eye Gaze
11
uncomfortable with the intensity of the gaze. A high SEF individual would also not be expected
to display anxious gaze avoidant or submissive gaze behaviors, as both of these behaviors would
arguably inhibit the establishment of rapport and negatively impact the quality of the social
interaction.
Equally as important as understanding how SEF promotes social-interaction quality is
understanding how social anxiety affects SEF in a social interaction. Although many studies
have confirmed the negative impact that social anxiety has on nonverbal behaviors and the
quality of social interaction, very little work has been done to examine the moderating role of
social anxiety on these variables. We predict that social anxiety will have a moderating effect on
SEF during social interactions. We predict that SEF will have trait like features making it
relatively stable over time. Social anxiety fluctuates based on the situation, and therefore we
predicted social anxiety would act as the moderator as it is less stable. We chose to measure the
effects of social anxiety by specifically examining eye-gaze behavior. Differences in eye-gaze
influenced by social anxiety are more prevalent than other nonverbal behaviors impacted by
social anxiety and exert greater influence on the resulting quality of the interaction. In this
experiment, we aim not only to take a preliminary behavioral step towards construct validation
of SEF, but also to test whether social anxiety moderates the association between SEF and social
interactions.
In order to test these ideas, we designed an experiment using two paradigms. In the first
scenario, the investigator leaves the stranger participants alone in the lab to create a socially
awkward situation. The purpose of this paradigm was to create a mild social-anxiety
provocation. The second paradigm involved a trip planning task based on the Bernieri Plan-ATrip paradigm, which was designed to measure IS (Bernieri, Gillis, Davis, & Grahe, 1996)
Social Anxiety Moderates SEF and Eye Gaze
12
Although we focused on those aspects most directly related to eye gaze, a few additional
measures (e.g., arm movement, trunk movement, and smiling) were also tested for exploratory
purposes. Hand/arm movement (.71) and fidgeting (.60) were both highly correlated with social
anxiety in previous studies (Monti, Kolko, Fingeret, & Zwick, 1984). Gesturing (.34) was
differentiated from other hand/arm movements in this code as it did not correlate strongly with
social anxiety (Monti et al., 1984). However, because the movement of gesturing is associated
with speech it was predicted that this movement would be a more prosocial behavior, potentially
correlating with a social function construct such as SEF. A relative absence of facial expressions
of emotion has previously been associated with anxious states, and therefore Duchenne and nonDuchenne smiling was also included in the behavioral code as a measure of facial expression of
emotion.
Methods
Participants
Participants were 48 students recruited from Vanderbilt University who participated
towards course credit in an introductory psychology course. Forty-eight participants were tested
as part of 24 stranger dyads, with an equal number of male and female participants (n = 24
females). There were 12 female-female dyads, and 12 mixed-gender dyads. Data was missing
Social Anxiety Moderates SEF and Eye Gaze
13
from two participants self-report measures, reducing the final number of participants for the selfreport analyses to n = 46 participants. Behavioral data from all 48 participants was still used,
and analyses were adjusted to account for the missing data.
Measures
The Social-Emotional Fluency Scale (Motlong & Bachorowski, in preparation) was used
to measure participant’s level of SEF. The SEF scale is composed of 26 items; responses are
collected using a 5-point Likert scale. The Liebowitz Social Anxiety Scale (LSAS; Heimberg et
al., 1999) is a 24-item scale used to measure fear and avoidance in social situations. Each item in
the LSAS is rated on a 4 point Likert scale with an alpha coefficient value of 0.96. The LSAS
can be scored using two indices of social anxiety: social fear and social avoidance. Participants
also completed the Emotional Intelligence Scale (Schutte et al., 1998) and the Interpersonal
Sensitivity Measure (Harb, Heimberg, Fresco, Schneier, & Liebowitz, 2002). The NEO Five
Factor Inventory (NEO-FFI; McCrae & Costa, 2004)), Social Desirability Scale, and SelfMonitoring Scale were also administered, but were used in the present analyses.
Design
The independent variable for this experiment was the dyad manipulation, which was
varied as sex of social partner. A within-subjects factor was the behavioral paradigm, specifically
the socially awkward deception paradigm and the trip-planning paradigm. All participants
experienced both paradigms in the same order. The dependent variable was the amount of eye
gaze directed toward each participant’s partner. The dependent variable eye-gaze behavior was
measured by the SEF behavioral coding scheme. Cameras were placed behind the each
participant to capture a frontal view of the other participant. Research assistants coding eye gaze
Social Anxiety Moderates SEF and Eye Gaze
14
in the behavioral coding scheme were able to bring up video images of both participants
simultaneously to ensure accuracy when measuring eye contact.
Procedure
Participants were greeted in the laboratory and the experimenter confirmed that the
participants did not know each other. Two pairs were not tested because they described
themselves as friends. Participants were seated at a table across from each other, and were given
oral and written consent information. Once written consent was provided, demographic
information was collected. Participants were then instructed that they were to plan a trip around
the world that they would take together if they were given $25,000. Next, the experimenter
turned on three cameras that were used to record audio and visual signals. The participants were
then falsely informed that a battery was missing from the one of the cameras. The experimenter
then excused him- or herself from the room to pretend to search for the missing battery pack;
participants were left alone for 5 min, during which time they were recorded. After 5 min had
passed, the experimenter returned to the room to announce that s/he could not find the battery
pack and would proceed with the experiment without it. Fake money, a globe, and a scratchpad
were then provided to the participants. Two rules were mandated for the trip: 1) Both
participants had to travel to the same locations together, and 2) they should continue planning
until all of the money had been spent. The experimenter then left the room and the participants
were given 12 minutes complete the task. Participants were given an additional three minutes if
they had not completed the task during the allotted time. Once the participants had finished, the
experimenter returned to the room, turned off the video cameras, and set up the participants to
complete the seven self-report questionnaires via MAC Airbooks. When finished, the
Social Anxiety Moderates SEF and Eye Gaze
15
participants were fully debriefed, provided consent to use the video data that had been used
during the social-anxiety task, and then were dismissed.
Coding
Three undergraduate students coded analyzed the video recordings. Two measures of eye
gaze were collected. The occurrence of eye gaze and eye gaze avoidance was microcoded in ½-s
increments for four 15-s units. Four 30-s measures of the quality of eye gaze were also coded,
with specific 5-point Likert ratings of overall eye-gaze quality, eye-gaze shifting, and dominant
or submissive eye-gaze behavior. Two of these time segments were coded from the socialanxiety paradigm in which the experimenter left the room searching for the “missing battery
pack.” The other two time segments were coded from the trip-planning task. Three other
behaviors were chosen for their correlation with social anxiety and coded in 4 30-s segments.
Hand/arm movement, gesturing, trunk movement such as fidgeting, Duchenne smiling, and nonDuchenne smiling were also coded. These behaviors were coded for being present and for
intensity using a 5-point Likert scale.
Results
Correlational Analyses
We used correlations to examine relationships between behavioral and self-report
measures. Correlations for SEF with the LSAS, EI scale, and the IS scale were conducted with
the results reported in Table 1. As expected, SEF was significantly negatively correlated with
social anxiety and significantly positively correlated with EI. No significant correlation between
Social Anxiety Moderates SEF and Eye Gaze
16
SEF and IS was found. Additionally, correlations for LSAS with the EI scale and the IS scale
were investigated, which were also reported in Table 1. Social anxiety scores on the LSAS were
significantly negatively correlated with EI and significantly positively correlated with IS.
Table 1
Self-Report Measure Correlations
Measure
1. SEF Scale
1
-
2
2. Social Anxiety
-.353*
-
3. Emotional Intelligence
.623**
-.434**
-
-.188
.602**
-.203
4. Interpersonal Sensitivity
3
4
-
*p< .05; **p< 0.1
SEF was also examined for possible gender effects. An independent samples t-test was
conducted on the data, with the results reported in Table 2 and Table 3. No significant
differences between male (M= 89.42, SD = 14.02) and female (M = 90.62, SD = 12.07) SEF
scores were found t(44) = -.284, p > .05.
We also examined how measures of eye gaze from each of the 4 coding bins correlated
with the measures of SEF and social anxiety. Self-reported SEF and social anxiety scores from
the LSAS were predictors of the association between social context and eye-gaze. We also
examined overall eye gaze quality due to its’ high correlation with eye contact r(46) =.508, p<
.01. results of these correlations are reported in Table 4. No significant correlations were found.
Social Anxiety Moderates SEF and Eye Gaze
17
Moderator Analyses
We used regression analyses to test for moderator effects. The dyad condition was
entered first, followed by the SEF and LSAS total scores. To test for moderation effects, the SEF
x LSAS interaction term was entered third. Evidence of moderation was taken as a significant
increment in R2 change between levels 2 and 3 of the regression analyses. Only one regression
analysis yielded evidence of a significant moderation effect: of Overall Eye Gaze Quality Time
2, R2 change =.123, F(1, 41) = 6.535, p=.014, yielded significant support for social anxiety as a
moderator variable. This effects is shown in Figure 1. Although the results were largely
insignificant, the presence of a statistically significant regression indicates that social anxiety
could act as a moderator variable.
Figure 1.
SEF x Eye Gaze Quality Time 2
Moderator Effects of Social Anxiety on SEF x Overall Eye Gaze Quality
Social Anxiety
Social Anxiety Moderates SEF and Eye Gaze
18
Exploratory Analyses
Although eye gaze was the focus of the study, correlations between self-report outcomes
and other behavioral measures were examined for exploratory purposes. When correlated with
SEF, LSAS, EI, and IS scores, both Duchenne and non-Duchenne smiles yielded various
significant results, which are reported in Table 5. Neither type of smile significantly correlated
with SEF. During various time measures both Duchenne and non-Duchenne smiles correlated
positively with social anxiety, as well as both negatively correlating with EI. One measure of
Duchenne smiling correlated positively with IS.
Table 5
Related Behavior Correlations
Behaviors
SEF
Social Anxiety
Emotional
Intelligence
Interpersonal
Sensitivity
Smiling: non-Duchenne
Time 1
-.152
.393*
-.069
.142
Smiling: Duchenne
Time 2
-.218
.330*
-.366*
.342*
Smiling: Duchenne
Time 3
-.099
.130
-.353*
.135
Smiling: Duchenne
Time 4
-.270
.327*
-.363*
.189
Smiling: non-Duchenne
Time 4
-.233
.237
-.291*
.129
*p<.05
Social Anxiety Moderates SEF and Eye Gaze
19
Discussion
One important outcome of this experiment was a preliminary step towards construct
validation of SEF. Although SEF was highly correlated with EI, it was also distinguished as a
separable construct. In other words, although there appears to be a good bit of similarity in the
extent to which individuals vary in SEF and EI, there is enough unexplained variance between
the two constructs to support the idea of independent explanatory power in social interactions.
SEF scores were not significantly correlated with IS scores, indicating that SEF and IS are
independent constructs. Additionally, and as expected, SEF was negatively correlated with social
anxiety, which was expected. Exploratory analyses did not find any gender differences
associated with SEF, which is consistent with other findings in the field of social functioning.
In this study eye-gaze behavior in two socially demanding situations, were examined in
order to test whether social anxiety moderates the relationship between SEF and eye-gaze
behavior. Little evidence was found to support this hypothesis, however, with only one measure
at one time point supporting the moderator hypothesis. Despite the lack of significant results it
appears that social anxiety may still be a moderator and the results could be due to a flaw in the
conceptualization. When examining the values resulting from the moderator analysis there was a
noticeable trend in the data indicating very weak statistical support for the interaction between
SEF and social anxiety on eye-gaze behavior. The statistical outcomes were stronger for the first
rather than the second paradigm. This overall trend is intriguing because the first paradigm
intended to provoke mild social anxiety, whereas the second trip planning paradigm that was not
Social Anxiety Moderates SEF and Eye Gaze
20
intended to provoke any anxiety. This discrepancy in statistical outcomes could indicate that
while the first paradigm did induce some social anxiety, the effect was not strong enough to
cause a significant enough amount of anxiety related behaviors to be detected by the model. This
could explain the failure of eye contact to negatively correlate with social-anxiety scores.
Furthermore, if the social anxiety condition was too weak, any residual anxiety during the trip
planning task could have diminished to an immeasurable point, which could explain the blatant
absence of effects from the second paradigm. Correlations with LSAS and other self-report
measures such as the SEF scale indicate that recruiting participants who experience social
anxiety was not a problem in this study. Therefore it appears more likely that the anxietyinducing paradigm was strong enough to elicit some anxiety-driven behaviors but not enough to
reach statistical significance. Further testing with a stronger anxiety manipulation may be
necessary to reveal more consistent behavioral effects and associations with individual
differences in SEF and social anxiety.
We failed to support our hypothesis that eye gaze behavior would correlate with SEF.
Eye gaze is enhanced by minimal distractions, which would have been impeded by the apparatus
supplied during the trip planning task such as the globe and trip planning task. This limitation
could indicate that eye gaze is not the most ideal behavior to focus on for this study. Other
behaviors, such as smiling, which are not as easily affected by the presence of potential
distractors may be more suitable than eye gaze. Additionally, nonverbal cues associated with
affective expression, such as smiling, may be more informative when exploring social anxiety as
a moderator. Although eye gaze can be associated with anxious- driven behaviors, eye gaze
behavior itself does not provide any information regarding the affective state of the individual.
Social Anxiety Moderates SEF and Eye Gaze
21
Nonverbal cues that do provide affective indications, such as Duchenne smiling, may better
illustrate the potential moderator role of social anxiety.
The results of the correlations for smiling could be due to several factors. First, it is
possible that the investigators overestimated the occurrence of Duchenne smiling when coding
the video. It may also be possible that non-Duchenne smiling could be a behavior employed by
those with mild forms of social anxiety. Unlike Duchenne smiling, non-Duchenne smiling is less
tightly coupled with the experience of positive affect. Consistent with previous research that
socially anxious individuals are less likely to express emotion, non-Duchenne smiling could be a
noncommittal form of socially engaging behavior. If a person produces a non-Duchenne smile
towards another individual, they are providing their social partner with feedback without having
to verbally communicate or really involve themselves in the interaction. Because the smile
provides social feedback, this signal could be construed as a pro-social behavior more typical of
a high SEF person, yet it may be likely this could actually be a subtle social-avoidance behavior.
This could account for the positive correlations of smiling with social anxiety and interpersonal
sensitivity, particularly if there was a coding error misrepresenting non-Duchenne smiles as
Duchenne smiles. Future studies could examine whether non-Duchenne smiling is actually a
socially avoidant behavior. If this is hypothesis is supported, then smiling could also be
investigated to determine if social anxiety moderates the relationship between SEF and smiling
in a similar way to the mechanism proposed in this study.
In this study, we took a preliminary step towards examining the behavioral correlates of
SEF. Although we failed to that eye-gaze behaviors are significantly predicted by individual
differences in SEF, it may well be the case that other behaviors – especially those more
susceptible to mimicry and imitation – are more closely related to SEF. Our test of the
Social Anxiety Moderates SEF and Eye Gaze
22
interaction of SEF and social anxiety on eye gaze behavior indicate that social anxiety can
sometimes moderate SEF; however, the results also show that the linkage is unlikely to be a
strong one. Further tests should investigate whether social anxiety is a true moderator variable, or
whether social anxiety affects the production and quality of nonverbal behavior cues in some
other way. Lastly, if social anxiety is found to moderate the association between SEF and SEFrelevant behaviors, future studies should investigate whether social anxiety also moderates the
production of other nonverbal behavior cues important to social functioning and the quality of
social interactions.
Social Anxiety Moderates SEF and Eye Gaze
23
References
Bernieri, F. J., Gillis, J. S., Davis, J. M., & Grahe, J. E. (1996). Dyad rapport and the accuracy of
its judgment across situations: A lens model analysis. Journal of Personality and Social
Psychology, 71(1), 110-129. doi: http://dx.doi.org/10.1037/0022-3514.71.1.110
Carney, D. R., & Harrigan, J. A. (2003). It takes one to know one: Interpersonal sensitivity is
related to accurate assessments of others' interpersonal sensitivity. Emotion, 3(2), 194200. doi: 10.1037/1528-3542.3.2.194
Hall, J. A., Andrzejewski, S. A., & Yopchick, J. E. (2009). Psychosocial Correlates of
Interpersonal Sensitivity: A Meta-Analysis. Journal of Nonverbal Behavior, 33(3), 149180. doi: 10.1007/s10919-009-0070-5
Harb, G. C., Heimberg, R. G., Fresco, D. M., Schneier, F. R., & Liebowitz, M. R. (2002). The
psychometric properties of the Interpersonal Sensitivity Measure in social anxiety
disorder. Behavior Research And Therapy, 40(8), 961-979.
Heimberg, R. G., Horner, K. J., Juster, H. R., Safren, S. A., Brown, E. J., Schneier, F. R., &
Liebowitz, M. R. (1999). Psychometric properties of the Liebowitz Social Anxiety Scale.
Psychological Medicine, 29(1), 199-212.
Hodgins, H., & Zuckerman, M. (1990). The effect of nonverbal sensitivity on social interaction.
Journal of Nonverbal Behavior, 14(3), 155-170. doi: 10.1007/BF00996224
Kassymzhanova, A., & Mun, M. (2013). Emotional Intelligence as the Object of Research in
Modern Psychology. Procedia - Social and Behavioral Sciences, 82, 892-894. doi:
10.1016/j.sbspro.2013.06.367
Kleinke, C. L. (1986). Gaze and Eye Contact: A Research Review. Psychological Bulletin,
100(1), 78-100.
Social Anxiety Moderates SEF and Eye Gaze
24
Langer, J. K., & Rodebaugh, T. L. (2013). Social Anxiety and Gaze Avoidance: Averting Gaze
but not Anxiety. Cognitive Therapy and Research, 37(6), 1110-1120. doi:
10.1007/s10608-013-9546-z
Lopes, P. N., Brackett, M. A., Nezlek, J. B., Schutz, A., Sellin, I., & Salovey, P. (2004).
Emotional intelligence and social interaction. Personality And Social Psychology
Bulletin, 30(8), 1018-1034. doi: 10.1177/0146167204264762
Mayer, J. D., Salovey, P., & Caruso, D. R. (2008). Emotional intelligence: new ability or eclectic
traits? The American Psychologist, 63(6), 503-517. doi: 10.1037/0003-066X.63.6.503
McCrae, R. R., & Costa, P. T. (2004). A contemplated revision of the NEO Five-Factor
Inventory. Personality and Individual Differences, 36(3), 587-596. doi: 10.1016/s01918869(03)00118-1
Monti, P., Kolko, D., Fingeret, A., & Zwick, W. (1984). Three levels of measurement of social
skill and social anxiety. Journal of Nonverbal Behavior, 8(3), 187-194. doi:
10.1007/BF00987290
Rosip, J., & Hall, J. (2004). Knowledge of Nonverbal Cues, Gender, And Nonverbal Decoding
Accuracy. Journal of Nonverbal Behavior, 28(4), 267-286. doi: 10.1007/s10919-0044159-6
Russo, N. F. (1975). Eye Contact, Interpersonal Distance, and the Equilibrium Theory. Journal
of Personality and Social Psychology, 31(3), 497-502.
Schutte, N. S., Malouff, J. M., Hall, L. E., Haggerty, D. J., Cooper, J. T., Golden, C. J., &
Dornheim, L. (1998). Development and validation of a measure of emotional
intelligence. Personality and Individual Differences, 25(2), 167-177. doi: Doi
10.1016/S0191-8869(98)00001-4
Social Anxiety Moderates SEF and Eye Gaze
25
Snodgrass, S. E., Hecht, M. A., & Ploutz-Snyder, R. (1998). Interpersonal sensitivity:
expressivity or perceptivity? Journal of Personality and Social Psychology, 74(1), 238249.
Terburg, D., Aarts, H., & van Honk, J. (2012). Memory and attention for social threat: anxious
hypercoding-avoidance and submissive gaze aversion. Emotion, 12(4), 666-672. doi:
10.1037/a0027201
Download