The Effects of Social Anxiety on Recognition of Facial Expressions and Evaluation of Negative Emotions

PAGE 1

The Effects of Social Anxiety on Recognition of Facial Expressions and Evaluation of Negative Emotions By Hayden Bickerton A Thesis Submitted to the Division of Social Sciences New College of Florida in partial fulfillment of the requirements for the degree Bachelor of Arts Under the sponsorship of Dr. Steven Graham Sarasota, Florida May 2013

PAGE 2

ii This th esis is dedicated to all those affected by social anxiety in hopes that gathering more information and awareness will help these individuals lead a more comfortable life with reduced anxiety.

PAGE 3

iii Acknowledgements Thanks to Dr. Steven Graha m, m y thesis sponsor, whose encouragement, expertise, and insight helped make this thesis possible. Thanks to Dr. Heidi Harley, my academic advisor, for her consistent support and enthusiasm throughout my academic career and for serving on my thesis committee. Thanks to Dr. John Newman for serving on my thesis committee and providing me with insightful conversations and stress therapy during our tennis matches. Thanks to Dr. Duff Cooper, who helped me immeasurably with my statistics. Thank you to all my thesis participants for taking the time to complete my study. the Four Winds staff for their cooperation. Thanks to my extremely supportive family for their unconditional love and encouragement throug hout the thesis process. Thank you to my brother, Chad, for always believing in me. Thanks to Stacy Roudabush, my southern sister, whose support, ambition, and invaluable friendship motivated me in the thesis process. Finally, thank you to all my friends and peers who made my undergraduate experience fulfilling and meaningful.

PAGE 4

iv Table of Contents ii .. iii Table of iv v 1 Negative Interpretat 2 Attentional .. 3 5 7 7 9 1 2 1 3 1 6 2 3 2 9 ... 3 2 Table ... 3 3 Table ... 3 4 Table ... 3 5 ... 3 6 Table ... 3 7 Appendix A: Fear of Negative Eval ... 3 8 Appendix B: Sample Photographs of 3 9 Appendix C: Word St 40

PAGE 5

v THE EFFECTS OF SOCIAL AN XIETY ON RECOGNITION OF FACIAL EXPRESSIONS AND EVALUATION OF NEGATIVE EMOTIONS Hayden Bickerton New College of Florida, 2013 ABSTRACT The relationship between social anxiety and the identification of emotional facial expressions and words were examined in this study. The goal of this study was to determine whether highly socially anxious individuals would respond more quickly and accurat ely to negative facial expressions and investigate a n attentional bias toward fear or anger. Seventy six participants were divide d into one of three groups: low, moderate, or high social anxiety. Participants were asked to classify facial expressions associated with happiness, anger, fear, and neutral as well as respond to word pairs associated with anger and fear. No significant differences were found for high social anxiety on reaction time s and accuracies for emo tional facial expressions. Results did reveal that highly socially anxious participants reacted more so to anger words than fear words. Future research is needed in order to determine if those with high social anxiety are more reactive to anger than fear a cross multiple stimuli. ________________________ Dr. Steven Graham Division of Social Sciences

PAGE 6

1 The Effects of Social Anxiety on Recognition of Facial Expressions and Evaluation of Negative Emotions Social anxiety is characterized by intense anxiety and discomfort about social situations involving interaction with others and evaluation from others. Socially anxious individuals possess a heightened fear of negative evaluation from their peers which can affect cognitive processing of a social situation. Upon entering a social scenario, individuals with high social anxiety feel watched and evaluated by others to an (Leber, Heidenreich, Stangier, & Hofmann, 2009). Simultaneously, socially anxious individuals are participating in self focused attention and self judging cognitions that can interfere with social performance. By focusing attention on themselves, socially anxious individuals hinder their ability to process social cues that might contradict their maladaptive beliefs and eliminate unnecessary anxiety. Clark and Wells (1995) proposed the notion that a shift of attention inward toward interoceptive information was a principal characteristic of social phobia. Interoceptive information can be defined as internal physiological changes like heart rate. Social phobi a is the clinical form of social anxiety and is characterized by a persistent fear of social situations in which the person is exposed to possible scrutiny by others and exposure to these situations elicits anxiety (American Psychiatric Association, 2000). Clark and Wells (1995) suggested that socially anxious individuals shift their attention internally and use interoceptive information to formulate impressions of themselves that they believe depict how others perceive them. Socially anxious individuals ar e preoccupied with their somatic responses and negative cognitions which can interfere

PAGE 7

2 with processing of external social cues. This tendency may be an explanation for a bias found in individuals with heightened social anxiety when recognizing facial expre ssions. This tendency is important when considering how judging facial expressions incorrectly could lead to false impressions that may induce anxiety. The human face is an important tool of communication. It allows communication of approval and disapprov al to others (Leber et al., 2009). Facial expressions provide (Niedenthal, Halberstadt, Margolin, & Innes Ker, 2000). Since socially anxious individuals are overly concerned w ith disapproval and negative evaluation from others, several studies have been conducted relating social anxiety to the recognition and evaluation of facial expressions. Negative Interpretation Bias Cognitive theorists suggest the existence of cognitive processes that lead to attentional and interpretational biases in socially anxious individuals. Gilboa Schechtman, Foa, Vaknin, Marom, and Hermesh (2008) conducted a study assessing negative interpreta tion bias in individuals with social phobia and depression. Previous research has found that participants with both psychological disorders tend to rate facial stimuli as more negative than would a participant without either disorder (Heinrichs & Hofman, 2 001). In this experiment, participants were presented with facial expressions illustrating sad, angry, and happy emotions varying in the level of intensity; participants were instructed to label each facial expression as positive, negative, or neutral as quickly and as accurately as they could. Results revealed that socially anxious and depressed individuals performed as well as controls in labeling happy expressions as positive but

PAGE 8

3 better than controls in labeling faint angry and sad expressions as negati ve. Shown in their propensity to label faint emotional expressions as negative, the results suggest that Also, socially anxious individuals may be less sensitive to chang es in facial expressions because following the initial assessment of the expression, they shift their focus to internal feelings and sensations. Gilboa Schechtman et al. (2008) noted fear of rejection and of negative evaluation by others as key features of social anxiety; therefore, the tendency of socially anxious individuals to shift their focus internally may be an attempt to preserve social performance and avoid negative evaluation by others. Attentional Bias in Social Anxiety Research has also explore d the attentional biases of people with social anxiety. Pineles and Mineka (2005) performed a study examining whether socially anxious individuals possess an attentional bias toward cues for an internal source of potential threat (heart rate), toward cues for an external source of potential threat (threatening faces) or both. Participants completed a dot probe task to determine location of attention. A dot probe paradigm measures attentional allocation to threatening s timuli through measuring reaction times to these threatening stimuli (Mac Leod & Mathews, 1988). The dot probe paradigm contained stimuli that represented both external and internal sources of potential t hreat. Internal sources of threat were manipulated by showing the participant segments of a visually similar sound wave. External sources of threat were manipulated through exposure to threatening, happy, and neutral faces. Attention was measured by comparing reaction time s for pairings of different combinations of the three types of

PAGE 9

4 faces and for pairings of the heart rate information with sound wave segments. Results demons trated that socially anxious individuals possessed an attentional bias toward that highly socially anxious individuals exhibit preferential attention to internal, somatic information. By shifting their attention inward to monitor internal physiological cues, socially anxious individuals may believe that others notice signs of anxiet y and judge them negative ly and thus, reinforce their fear of negative evaluation and self focused negative cognitions. Other research has expanded the assessment of attentional bias by incorporating faces in addition to words as stimuli. Pishyar, Harris and Menzies (2004) performed a study assessing attentional bias for words and faces and social anxiety. The researchers conducted two studies comparing the performance of individuals with high and low social anxiety on dot probe tasks using words, faces photographed in front view, and faces photogra phed in profile, as stimuli. In S tudy 1, the dot probe task consisted of 40 experimental trials, 20 positive neutral pairs and 20 negative neutral pairs, presented in a new random order to each person. Attentio nal bias was measured utilizing a formula described by MacLeod and Mathews (1988). A positive value produced by this formula indicated a shift of attention towards the emotional stimuli relative to the neutral stimuli, whereas a negative valued indicated a shift of attention away from the emotional stimuli relati ve to the neutral stimuli (Pishyar, Harris, & Menzies, 2004). Results from Study 1 indicated those with high social anxiety possessed an attentional bias towards negative faces. The low social anxiety group displayed an attentional bia s towards positive faces while no significant effects were

PAGE 10

5 observed on the dot probe using words as stimuli. In Study 2, the dot probe stimuli were modified so that participants were shown pictures of two faces in profile looking at each other. Like Study 1, the expression on one of the faces was varied so that it was positive, negative or neutral. The second face presented always had a neutral expression, and in le vels of social anxiety were associated with attentional bias towards negative facial was present on the screen. The findings from these studies provide evidence tha t high social anxiety is associated with an attentional bias toward negative faces. Participants with high social anxiety responded faster to probes in the location of threatening facial stimuli. Thus, this finding suggests that highly socially anxious ind ividuals attend more to negative faces in comparison to neutral or positive faces. This research also aligns with the findings of Gilboa Schechtman et al. (2008) in which individuals with social phobia had shorter reaction times for detecting angry faces c ompared to happy faces. Attributional Bias Cognitive theories also mention the presence of an attributional bias in socially anxious individuals. During a social interaction, socially anxious individuals occupy much of their time thinking about what they will say next or the impression they are forecasting. There is speculation that too much self focused attention induces anxiety (Woody, 1996). So during a social interaction, socially anxious individuals are partaking in negative self directed cognitive a ctivity. The components of this cognitive process involve self focused attention, fear of negative evaluation, social performance, and negative self judging cognitions (Woody, 1996). Self focused attention in socially

PAGE 11

6 anxious individuals increases negative cognitions. Self directed attention may increase negative cognitions by changing the attributional bias of socially anxious individuals. An attributional bias is a cognitive bias in which individuals tend to take credit for success and blame others for fa ilure; however, for socially anxious individuals, the blame for social failure is placed on themselves (Woody, 1996). In a study conducted by Woody (1996), focus of attention was manipulated during a speech task to examine the effect of focus of attentio n on anxiety levels. Participants with social phobia were tested in pairs before an audience. Participant A was required to give a set of two speeches, and Participant B stood in front of the audience during t he delivery of these speeches. Those in t he passive, sil ent role constituted the control group. The content of the speeches consisted of anxious bodily sensations, cognitions, and emotions, but the person who was the focus of the speech varied for each trial. In the first trial, the content of Participant A's speech fo cused on their own current bodily sensations, cognitions, and emotions. For the second trial, Participant A's speech were measured by self report and rater observation. R esults revealed that increasing self focused attention increased anticipated anxiety and anxious appearance, regardless of the condition. Increased anticipated anxiety and anxious appearance was seen whether the participants were giving a spe ech or passively standing in front of the audience. These results suggest that self focused attention may play a causal role in intensifying social anxiety. By inducing self focused attention, anxiety levels were heightened These results provide strong ev idence that self focused attention is a

PAGE 12

7 component of social anxiety and those with high social anxiety are engaging in internal focus and ignoring external social cues. Information Processing Bias Kim and Oh (2010) conducted a study investigating the rela tionship among information processing bias in socially anxious individuals. The information processing of socially anxious individuals is characterized by the avoidance of threatening stimuli. Kim and Oh (2010) compared the performance of participants with high social anxiety and low social anxiety on three different tasks. In the primary task, participants were instructed to order three sequentially presented pictures of facial expressions according to the intensity of the emotions. In addition to the prim ary task, participants were given tasks requiring them to order pictures of neutral faces according to age and order pictures of geometrical shapes according to size. Results indicated that highly socially anxious participants performed more poorly on the primary facial expression task compared to the low socially anxious participants; however, the highly socially anxious group did not perform more poorly on the latter two tasks involving the nonemotional stimuli. The latter two tasks were used to determine whether the difficulties of socially anxious individuals are specific to processing emotionally charged facial expressions. The results indicate evidence that high social anxiety interferes with processing of emotionally charged facial expressions. Emoti onal State in Relation to Processing Emotional Facial Expressions Niedenthal et al. (2000) performed a study evaluating the effect of emotional state on the detection of change in emotional facial expressions. The goal of Niedenthal et al. (2000) was to as sess whether the emotional state of the perceiver affects the perception of

PAGE 13

8 the longevity of an expression; in other words, is an expression perceived as lasting longer when it is congruent with the emotional state of the perceiver? Film clips were used to induce happiness, sadness, or neutral emotion. Following the emotion induction period, participants completed the Brief Mood Introspection Scale. Emotions were maintained by having participants in the happiness and sadness conditions listen to classical m usic. Control condition participants heard no music. Participants then watched a computerized 100 frame movie in which the first frame displayed a face expressing a specific emotion such as happiness. The facial expression gradually became neutral over the course of the movie and participants were instructed to indicate at which frame the initial expression no longer appeared on the face. Results showed that emotion congruent expressions were perceived as persisting longer than emotion incongruent expressio ns. Furthermore, participants in the happy condition perceived the offset of a happy expression at a later frame in the movie than their perception of the offset in the sad expression. These findings have an important implication for the research of social anxiety since changes in facial expressions are monitored by others in order to assess the attitudes of others. Other studies have also examined the effect of emotional state on processing emotionally charged facial expressions. Leber and colleagues (2009 ) executed a study assessing the relationship between processing of facial expressions and social threat. Socially anxious and non anxious individuals were instructed to classify facial expressions associated with anger, sadness, fear, disgust, happiness, and surprise. Half of the participants performed this task after being faced with a social threat in which they performed a speech prior to completing the facial expression task. Participants also had to complete a separate task using household

PAGE 14

9 objects as neutral control stimuli. Results revealed that highly socially anxious participants were quicker than non anxious participants at distinguishing angry, sad, and fearful faces when confronted with a social threat. Group differences were not observed under t he no threat condition. These findings suggest that high ly socially anxious individuals are more vigilant toward threat related social cues and that the processing of anxio us participants responded more quickly to angry facial stimuli as well as to sad and fearful facial stimuli. This finding provides evidence that socially anxious individuals attend more to negative facial expressions. Also, the anticipation of a speech cou ld have triggered negative self focused cognitions making it easier to retrieve negative emotional information and apply it to the perception of the facial expressions. In other words, the angry, sad, and fearful faces may have been more easily recognized because they aligned Fear vs. Anger The culmination of these studies show research on attentional bias in socially anxious individuals has been inconsistent. Attentional bias towards some specific facial expressi ons have been examined for negative vs. positive facial expressions; however, research assessing the relationship between social anxiety and attentional bias towards fearful or angry faces has yet to be examined. Previous research has assessed this topic m ore broadly and has had mixed findings. Leber et al. (2009) indicated that socially anxious individuals are more reactive toward threat related social cues and the processing of facial expressions may be related to their current emotional state. The result s of this study suggest that angry, sad, and

PAGE 15

10 fearful facial expressions are more easily recognized because these emotional expressions Schechtman et al. (2009) found that people with social phobia were more likely to recognize cues of anger and sadness than happiness; however, people with social phobia were less sensitive to changes in facial expressions. Researchers proposed that this decreased sensitivity was due to individuals focus ing internally on their feelings and somatic sensations. Taylor, Bomyea, and Amir (2011) conducted an enlightening study in which they tested the malleability of attentional bias for positive emotional information. Taylor et al. (2011) stated that individ uals are distinguished in their tendency to allocate attention toward negative emotional cues under situations of stress This tendency suggests different susceptibility to anxiety. Researchers believed that social anxiety was affiliated with a diminished ability to access and perceive positive emotional information. Participants were instructed to complete a probe detection task designed to induce selective processing of positive stimuli or to a mock condition. Following this task, participants were expose d to a stressor by performing a videotaped speech. Anxiety levels and response to the stressor were examined. Results indicated that social anxiety was negatively associated with change in bias in the attentional training paradigm group proposing that as s ocial anxiety levels increased, the shift in attentional allocation toward positive stimuli decreased. Participants who illustrated the largest shift in attentional allocation toward positive stimuli following the attentional training procedure displayed t he least amount of anxiety reactivity to the stressor. Thus, higher levels of social anxiety were linked to diminished attentional allocation toward positive cues. These findings suggest that social anxiety may interfere with perceptual processing of posit ive

PAGE 16

11 information or socially anxious individuals might be more sensitive to negative information due to their fear of negative evaluation and tendency to engage in negative self judging cognitions. Blair et al. (2008) employed a study comparing the response to emotional facial expressions in generalized social phobia and generalized anxiety disorder. Generalized social phobia involves fear and/or avoidance of social situations while generalized anxiety disorder involves intense worry about various circumstan ces. Researchers utilized functional magnetic resonance imaging, fMRI, to measure the neural response to facial expressions in generalized social phobia and generalized anxiety disorder. The procedure involved participants viewing neutral, fearful, and ang ry expressions while making a gender judgment. Results indicated that individuals with generalized social phobia showed increased activity to fearful relative to neutral expressions in several regions of the brain, including the amygdala, compared to contr ols. Individuals with generalized anxiety disorder displayed less activity to fearful relative to neutral faces compared to controls in the amygdala but did show significantly increased response to angry expressions in comparison to controls in a lateral r egion of the middle frontal gyrus. The results suggest a difference in neural circuitry between individuals with generalized social phobia and those with generalized anxiety disorder. The previous studies mentioned have utilized negative and threatening s timuli to determine differences in reactivity among socially anxious individuals and non socially anxious individuals. While Gilboa Schechtman et al. (2008) found that socially anxious individuals performed better than controls at detecting angry and sad f aces, the researchers neglected to include fear faces as stimuli. Leber et al. (2009) included fear

PAGE 17

12 and anger stimuli in their study but failed to distinguish differences in reactivity to these stimuli. Thus, it is difficult to determine whether socially a nxious individuals would be more reactive to fear or anger. Blair et al. (2008) provides neurological evidence of differences in brain response activity among individuals with social phobia and generalized anxiety disorder. Since individuals with social ph obia showed increased activity to fearful expressions but not angry expressions in the amygdala, it can be suggested that individuals with high social anxiety might have an attentional bias to fearful expressions since social anxiety is a component of soci al phobia. In addition, Leber et al. (2009) suggested that the current emotional state of the individual may be related to the processing of facial expression s Since fear of negative evaluation is a component of heightened social anxiety, highly socially anxious individuals might process fearful expressions more quickly because their current emotional state is consistent with the presented expression. Overview of Current Research The present study has been designed to assess the associations between social anxiety and emotional facial recognition and to test the existence of an attentional bias to wards anger or fear. Previous research has found that highly socially anxious individu als are faster at detecting negative facial expressions in comparison to low socially anxious individuals (Leber et al., 2009; Pishyar et al., 2004). Therefore, it was hypothesized that individuals with high social anxiety would have higher accuracy and sh orter response time when recognizing negative facial expressions. A word probe detection task was utilized as a measure to assess an attentional bias towards fear or anger

PAGE 18

13 associated words. It was hypothesized that individuals with high social anxiety woul d attend more to fear than anger. Methods Participants Seventy six undergraduate students (57 females, 18 males, 1 other) over the age of 18 ( Mdn = 21.0 years, age range: 18 25 years) participated in this experiment. Students attending a small liberal arts college in southwest Florida were recruited utilizing multiple emails to the student forum inviting volunteers. Compensation included choosing ei ther a $5 gift card to an on campus student run cafe or a pen. Materials Participants completed the Fear of Negative Evaluation (FNE) scale, a standardized questionnaire which measures anxiety associated with perceived negative evaluation (Watson & Frien d, 1969). The scale contained 30 items with True/False riend, 1969; see Appendix A for all items). The FNE .98) and test retest reliability was .78 to .94 according to Watson and Friend (1969). The FNE and Brief Fear of Negative Evaluation (BFNE) ( Leary, 1983) correlate strongly, r = 0.96, which provides evidence of construct and criterion validity (Andrews & Robinson, 1991). The Demographic information such as age, gen der, AOC division, race, and ethnicity were also recorded.

PAGE 19

14 Stimuli Two types of stimuli were utilized in this study: faces and pairs of words. The facial stimuli were selected from the Montreal Set of Facial Displays of Emotion (Beaupre, 2005). The stim uli consisted of 32 pictures, 16 Caucasian males and 16 Caucasian females. The facial expressions displayed were anger, fear, happiness, and neutral (see Appendix B for sample items). The verbal stimuli were selected from three sets that have been used i n previous studies of emotional words (Dangleish, 1995; Johnson Laird & Oatley, 1989; Yang, Jackson, Gao, & Chen, 2012). Forty word pairs were created with 20 fear neutral word pairs and 20 anger neutral word pairs (see Appendix C). The emotion neutral wor d pairs were matched by word length. Emotional Facial Recognition Task The emotional facial recognition task was administered to participants on a computer using SuperLab 4.0. Participants were asked to rapidly and accurately decide which face on the sc reen was displaying a particular emotional expression. Emotional expressions assessed included happiness, fear, anger, and neutral. Participants completed a practice block before beginning the experimental blocks. The practice block asked participants to d ecide as rapidly and accurately Ten facial stimuli were presented, two of each displaying the four different emotional expressions. After completing the practice block, participants completed the experimental blocks One block was employed per emotional expression so participants completed a total of four blocks. Participants were shown the same 32 pictures in each block. Each facial expression stimulus was presented randomly, one at a time. Each facial expression

PAGE 20

15 wa s shown for 60ms (Leber et al, 2009). A fixation cross was shown in between each stimulus presented. The fixation cross was shown until a decision about the facial expression was indicated by the participant. To indicate a decision following the picture, p Word Probe Detection Task The word probe detection task was d esigned to study the presence of an attentional bias toward fear or anger. The task was administered to participants on a computer using SuperLab 4.0. Participants were presented with pairs of words, an anger neutral word pair or a fear neutral word pair. Participants were asked to decide which word relates to an emotion. The word pairs were shown until participants pressed a key indicating a d on the right relating to an emotion. Procedure All participants were provided a link via e mai l to access the FNE scale using SurveyMonkey. All participants were informed that they were participating in a study on social interactions and making judgments about faces and words. After receiving informed consent, participants were asked to decide whet her each statement was true or false as it pertained to each of them personally. Participants were also asked to provide demographic information. Following the completion of the survey, participants were invited to complete tasks in the lab and sign up for a time to come in. All participants were tested individually. Participants were assigned to complete three tasks, an emotional facial recognition task, a filler task, and a word probe task. Participants were randomly

PAGE 21

16 assigned to begin with either the faci al recognition task or the word probe task. The facial recognition task and word probe task were administered on the computer using SuperLab 4.0 while the filler task was administered at a desk behind the computer station. The filler task involved particip ants responding to a writing prompt asking them to describe their daily routine now in comparison to their daily routine in high school. Participants were allowed to write for 5 minutes. At the completion of the experiment, participants were debriefed, tha nked, and given compensation for their time. Results Self Report Measures Participants were allocated to one of three groups: low, moderate, or high social anxiety based on scores from the FNE scale. Participants were allocated to these groups based on st andards reported by Watson and Friend (1969). Participants with FNE scores of 12 or less were assigned to the low social anxiety group while participants with scores of 21 or more were assigned to the high social anxiety group; participants scoring between 13 and 20 were assigned to the moderate social anxiety group (Watson & Friend, 1969). Table 1 lists the results of the comparison between the low, moderate, and highly socially anxious groups for various self report measures. Scores on the FNE scale range d from 2 30 ( M = 16.5, SD = 7.39). Data Analysis For each participant, correct reaction times were identified for each facial expression block. Using SAS 9.2, proc univariate was run for each photograph shown in each facial expression block to identify re action time outliers. Outliers were classified as values that exceeded the lower quartile, Q1, and upper quartile, Q3, by 3 interquartile

PAGE 22

17 ranges. Accuracy percentages were computed for each participant by identifying correct reaction times for each facial expression block and dividing the number of correct reaction times by the total number of photographs for which correct responses were given. R eaction times for incorrect responses were not used in data analysis. Accuracy percentages were translated to accuracy scores r ounded to two decimal places. For the word probe task, correct reaction times were identified for each participant for each emotion neutral word pair presented. Outliers were identified utilizing the same method as in the facial expression task for each fear neutral and anger neutr al word pair. R eaction times for incorrect responses were not used in data analysis. Accuracy percentages were computed for each participant by identifying correct reaction times for the fear neutral and anger neutral word pairs and dividing the number of correct reaction times by the total number of fear neutral and anger neutral word pairs for which correct responses were given. Accuracy percentages were translated to accuracy scores rounded to two decimal places. Classification of Facial Expressions Was social anxiety associated with faster reaction times when identifying negative facial expressions? A one way analysis of variance (ANOVA) was performed for social anxiety and the mean reaction times of each facial expression. Table 2 repor ts the means and standard deviations for the mean reactions times of each facial expression. The results indicated there was no main effect across low, moderate, and high social anxiety on the mean reaction times when identifying happy facial expressions, F (2, 73) = 0.59, p = .56, 2 = .02. There was no main effect for social anxiety on the mean reaction times when identifying angry facial expressions, F (2, 73) = 0.59, p = .56, 2 = .02, or

PAGE 23

18 fearful facial expressions, F (2, 73) = 0.57, p = .57, 2 = .02. There was a marginally significant main effect for social anxiety on the mean reaction times when identifying neutral facial expressions, F (2, 73) = 2.48, p = .09, 2 = .06. Participants with lower social anxiety were marginally faster ( M = 703.98, SD = 138.03) at identifying facial expressions as neutral than participants with moderate ( M = 728.98, SD = 138.59) or high ( M = 795.41, SD = 154.73) social anxiety. To investigate the relationship between mean reaction times of each facial expression and soci al anxiety, Pearson r correlations were performed. No significant correlations were found for happiness, r (74) = 0.07, p = .56, anger, r (74) = 0.10, p = .37, or fear, r (74) = 0.09, p = .46. A marginally significant positive correlation was found for the ne utral mean reaction times, r (74) = 0.22, p = .06, indicating that as social anxiety increased, mean reaction times increased. Thus, participants with higher social anxiety had slightly slower reaction times compared to participants with lower social anxiet y when identifying neutral facial expressions. Was social anxiety associated with faster reaction times when iden tifying facial expressions in each block? The mean reaction times for each facial expression within each block were investigated to find diff erences among participants with low, moderate, and high social anxiety. For example, the mean reaction times were computed for the happiness block when participants were identifying the happy, angry, fearful, and neutral faces presented in this block. Tabl e 3 displays the means and standard deviations for the mean reaction times within each block. A one way analysis of variance (AN OVA) was executed to examine for differences in each block. The results indicated no significant main effects for social anxiety on reaction time within the happiness, angry, or

PAGE 24

19 fearful block. The results revealed a significant main effect for social anxiety on reaction time in the neutral block when responding to happy faces, F (2, 73) = 4.01, p = .02, 2 = 0.10. A post hoc Tukey HSD test showed that participants with low social anxiety were faster ( M = 605.93, SD = 151.50) at identifying happy faces as not neutral in the neutral block as opposed to participants with high social anxiety ( M = 738.44, SD = 175.95). Participants wit h high social anxiety were not as fast as participants with moderate ( M = 694.93, SD = 143.81) social anxiety. Results also indicated a significant main effect for social anxiety on reaction time in the neutral block when responding to fearful faces, F (2, 73) = 3.48, p = .04, 2 = .09. A post hoc Tukey HSD test revealed that participants with low social anxiety were faster ( M = 640.21, SD = 155.71) at identifying fearful faces as not neutral in the neutral block as opposed to participants with high ( M = 7 66.79, SD = 172.45) social anxiety. Participants with high social anxiety did not respond more or less quickly than participants with moderate ( M = 680.59, SD = 165.64) social anxiety. To examine the relationship between social anxiety and mean reaction t imes within each block, Pearson r correlations were computed. No significant correlations were found for the happiness, anger, or fear blocks. A significant positive correlation was found in the neutral block when reacting to happy faces, r (74) = 0.27, p = .02, and fearful faces, r (74) = 0.26, p = .02. As social anxiety increased, mean reaction times in the neutral block when responding to happy and fearful faces increased as well. Thus, participants with higher social anxiety had a slower reaction time compared to participants with lower social a nxiety when identifying the happy and fearful faces as not neutral in the neutral block. Although it was predicted that high social anxiety would

PAGE 25

20 yield faster reaction times to negative facial expressions compared to low social anxiety, the results did not support this hypothesis. Was social anxiety associated with higher accuracy scores when identifying negative facial expressions? A one way analysis of variance (ANOVA) was computed for social anxiety and the mean accuracy scores for each facial expressi on. Table 4 reports the means and standard deviations for the mean accuracy scores of each facial expression. The results indicated no significant main effect for social anxiety on accuracy for happiness, F (2, 68) = 1.40, p = 0.25, 2 = .04, anger, F (2, 7 3) = 0.19, p = .83, 2 = .00, fear, F (2, 71) = 0.27, p = .76, 2 = .00, or neutral, F (2, 71) = 0.89, p = .42, 2 = .02. Pearson r correlations were performed to assess the relationship between mean accuracy scores for each facial expression and social anxiety. No significant correlations were found for happiness, r (69) = 0.18, p = .13, anger, r (74) = 0.03, p = .79, fear, r (72) = 0.02, p = .84, or neutral, r (74) = 0.27, p = .02. Thus, there was no relationship found between mean accuracy scores and social anxiety when identifying each facial expression. Although it was predicted that high social anxiety would generate higher accuracies when identifying negative facial expressions, the results did not support this hypothesis. Classification of Emotional Words: Fear vs. Anger Did those with high social anxiety attend more to fear or anger? A one way analysis of variance (ANOV A) was performed for social anxiety and the mean reaction times for the fear neutral and anger neutral word pairs. Table 5 shows the mean reaction times for low, moderate, and high social anxiety. The results revealed a significant main effect for social a nxiety on reaction time when responding to the fear neutral word pairs,

PAGE 26

21 F (2, 72) = 5.37, p = .00, 2 = 13. A post hoc Tukey HSD test revealed that highly socially anxious participants responded more quickly ( M = 1022.69, SD = 192.56) than those with low ( M = 1196.16, SD = 192.49) social anxiety. The reaction times of participants with low social anxiety did not differ from moderately ( M = 1142.67, SD = 168.43) socially anxious participants. A significant main effect for social anxiety on reaction time was found when responding to the anger neutral word pairs, F (2, 73) = 4.87, p = .01, 2 = .12. A post hoc Tukey HSD test revealed that highly socially a nxious participants responded faster ( M = 1002.36, SD = 150.12) than those with moderate ( M = 1179.21, SD = 268.43) social anxiety but not faster than participants with low ( M = 1127.26, SD = 166.67) social anxiety. Spearman correlations were performed t o evaluate the nonlinear relationship between social anxiety and mean reaction times to the fear neutral and anger neutral word pairs. Results revealed a significant negative correlation for fear, (73) = 0.31, p = .00. Thus, as social anxiety increase d, the reaction time to the fear words decreased. A significant negative correlation was also found for anger, (74) = 0.26, p = .02. As social anxiety increased, the reaction time to the anger words decreased. A one way analysis of variance (ANOVA) was implemented for social anxiety on mean accuracy scores for each emotional word pair. Results indicated a marginally significant main effect for social anxiety on accuracy scores for the anger, F (2, 70) = 2.93, p = 0.06, 2 = .08. Participants with high social anxiety were slightly less accurate ( M = 0.96, SD = 0.06) than those with moderate ( M = 0.97, SD = 0.06) and low ( M = 1.00, SD = 0.00) social anxiety for anger words. There was not a main effect for social a nxiety on accuracy scores for the fear, F (2, 71) = 2.03, p = 0.14, 2 = .05.

PAGE 27

22 Spearman correlations were computed to examine the nonlinear relationship between social anxiety and accuracy scores for fear and anger words. Results revealed a marginally sig nificant negative correlation between social anxiety and accuracy scores for the anger word pairs, (71) = 0.22, p = .06. Thus, as social anxiety slightly increased, accuracy slightly decreased. No significant correlation was found for accuracy scores and the fear word pairs, (72) = 0.17, p = .15. Although it was predicted that people with high social anxiety would attend more to fear than anger, the results did not support this hypothesis. Highly socially anxious participants responded faster to a nger words than fear words, and as social anxiety increased, accuracy scores for anger words marginally decreased while no differences in accuracy scores were found for fear words. Secondary Analyses Although not hypothesized, further analyses were conducted to compare social anxiety with age, gender, AOC, race, and ethnicity. A Pearson r correlation was performed to assess the relationship between age and social anxiety. Results revealed a significant negative correlation between age and social anxiety, r (74) = 0.33, p = .00. As age decreased, social anxiety increased. A chi square test was performed to examine the relationship between gender and social anxiety. Results revealed that the relationship was approaching significance, (2, N = 75) = 5.25, p =.07. A point biserial correlation was conducted to further assess the relationship between gender and social anxiety. Results indicated a significant negative correlation, (73) = 0.26, p = .02. Thus, females were more socially anxious than males. A one way analysis of variance (ANOVA) was computed to assess the relationship for social anxiety across AOC. Results revealed no significant differences in social anxiety among humanities, social scien ce, and natural

PAGE 28

23 science AOCs, F (2, 68) = 0.94, p = .40, 2 = .03. Point biserial correlations were computed to assess the relationship between race and ethnicity on mean reaction times and accuracies for each facial expression. No significant differences were found. Discussion None of the hypotheses were supported by the data. Analyses revealed that social anxiety had no effect on reaction time when identifying positive or negative facial expressions. Social anxiety did yield significant differences for reaction time when identifying neutral facial expressions. Participants with low social anxiety were significantly quicker at identifying happy and fearful faces as not neutral as opposed to participants with high social anxiety. Analyses also indicated t hat social anxiety had no effect on accuracy when identifying happy, angry, fearful, or neutral facial expressions. Social anxiety did reveal significant differences when identifying anger and fear words. Highly socially anxious participants responded fast er to anger words than fear words. Higher social anxiety did yield slightly less accuracy for anger words but accuracy did not differ across social anxiety for fear words. Inconsistent with the original hypothesis, individuals with high social anxiety re acted more quickly to anger words than fear words. Since no significant differences were found across social anxiety when identifying the angry and fearful facial expressions, it cannot be suggested that higher social anxiety attended more to anger than fe ar. It can be suggested that high social anxiety was more reactive to anger words than fear words. Highly socially anxious participants were also marginally less accurate when responding to anger words suggesting that while they may have been more reactive to

PAGE 29

24 anger, the presentation of aggressive stimuli may interfere with the processing of threatening cues. Alternative Explanations The results of the current study do not align with previous findings produced by Leber et al. (2009) in which highly sociall y anxious individuals were quicker at classifying negative facial expressions compared to low socially anxious individuals. In this study, half of the participants were confronted with a social threat prior to the experimental procedure. No differences in reaction times across social anxiety were found for the non social threat condition which suggests that the introduction of a stressor to induce anxiety may be necessary to produce expected results. Pinelas and Mineka (2005) also included a social stressor in their methodology but their results revealed that the absence or presence of the stressor did not influence the performance of highly socially anxious participants. Thus, a stressor was neglected to be included in the experimental procedure to avoid po tential harm to participants but the results of the current study suggest that anxiety induction might have been needed to generate data supportive of my hypotheses. The results of the current study partially parallel the results produced by Pineles and M ineka (2005). Pineles and Mineka (2005) found the presence of an attentional bias toward internal sources of threat but not toward external sources of threat which were threatening faces. The results of the current study found no significant differences ac ross social anxiety when reacting to happy, angry, and fearful facial expressions. The results found by Pineles and Mineka (2005) provide support for Clark and Wells (1995) cognitive model of social phobia which states that socially anxious individuals all ocate

PAGE 30

25 their attention inward to monitor internal physiological cues. This model could be a possible explanation as to why the current study did not find significant differences for high social anxiety when responding to the different facial expressions. Hi ghly socially anxious participants may have been participating in an internal shift of attention which could have interfered with processing the external facial stimuli. Alternative explanations as to why no significant differences were found when identi fying facial expressions could be attributed to an information processing deficit possibly present in highly socially anxious individuals. The results revealed that low socially anxious participants were significantly faster at identifying happy and fearfu l faces as not neutral than highly socially anxious participants. This finding reveals that social anxiety may interfere with processing and identifying emotionally charged faces. Kim and Oh (2010) suggested this information processing deficit after highly socially anxious individuals performed significantly more poorly than low socially anxious individuals in ordering facial expressions based on intensity; however, highly socially anxious individuals performed just as well on tasks involving non emotional stimuli. In the current study, participants with high social anxiety may have taken longer to identify emotionally charged faces as not neutral due to an information processing deficiency that possibly interfered with their ability to effectively process a nd interpret external cues. This information processing deficit could also explain why highly anxious participants were slightly less accurate when responding to anger words. Future Directions While highly socially anxious participants did not seem very reactive to emotionally charged faces, they were reactive to emotionally charged words. Socially

PAGE 31

26 anxious participants may have reacted more so to the words because both word pair stimuli were ass ociated with only negative emotions while the facial stimuli included negative faces as well as positive and neutral faces. Future studies should employ the usage of words as stimuli across multiple emotions to investigate the relationship between social a nxiety and the emotions associated with the words. Another future direction involves the incorporation of self focused attention. The results of Woody et al. (1996) suggested that self focused attention may be a component in heightening social anxiety. F uture studies could include the induction of self focused attention through techniques such as using mirrors or videotaping participants making a speech in order to examine the relationship between social anxiety, self focused attention, and processing emo tional facial expressions and words. Leber et al. (2009) noted that emotional state affected the performance of highly socially anxious individuals when identifying facial expressions while being confronted with a social threat. Future studies could not only include the presence of a social stressor but in addition, measures of mood or current emotional state could be incorporated to assess the connections between social anxiety, current emotional state, and the reactivity to emotional facial expressions. The inclusion of these measures may be important in understanding the processing of information in socially anxious individuals. The results of this study could be strengthened through altering methodology. As previously noted, the introduction of a soc ial stressor to induce anxiety may be necessary to evoke predicted results. The brief stimulus duration for the facial stimuli could also have attributed to the outcome of this study. Pineles and Mineka (2005) and Pishyar et al. (2004) used longer stimulus durations and yielded significant results. The duration may

PAGE 32

27 Furthermore, this could explain why highly socially anxious participants were no less fast or accurate than low socially anxious participants when facial stimuli were presented briefly but are so when more time is allotted for processing. There also may have been a slight ceiling effect for the accuracy scores across social anxiety for each facial expression wh ich suggests that participants performed well at the task regardless of stimulus duration. Future studies should extend the stimulus duration to allow for adequate processing and explore its effect on reaction time and accuracy. Limitations in the Current Study Limitations were present in the current study which provides more avenues for future directions. The population this study investigated is somewhat far from being representative of the general population. All participants were connected directly to t he researcher or indirectly through the method of participant selection. The majority of participants were white females attending a small, liberal arts school in Florida. Future studies could investigate the same research questions utilizing a broader and more representative sample of college students. Additionally, the participants recruited were not diagnosed social phobia patients which could explain why the results were not consistent with previous research findings that utilized individuals with socia l phobia. Another limitation in the current study was that the results could describe only correlations. In order to apply the results to a practical setting, causation needs to be shown. Since participants were not assigned to groups prior to beginning t he study, they could only be measured on where they fell on the FNE scale. In the future, participants

PAGE 33

28 should be separated into the appropriate social anxiety groups prior to completing any tasks so data analysis can provide results suggesting possible cau sal relationships. If the hypotheses of this study are eventually confirmed, it would add further support to findings associating high social anxiety with faster responses and accuracies to negative facial expressions as well as reveal whether highly socia lly anxious individuals are more reactive toward anger or fear. This information is useful for evaluating the difficulties socially anxious individuals encounter in their interpersonal relationships and can provide insight to the origin of these difficulti es. Thus, this information can be beneficial to the treatment of social anxiety and may help decrease incorrect judgments of facial expressions and therefore, increase satisfactory social interactions and hopefully decrease social anxiety.

PAGE 34

29 Ref erences American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders. (4th ed., text rev.). doi: 10.1176/appi.books.9780890423349 Andrews, F. M., & Robinson, J. P. (1991). Measures of subjective well being. Measures of personality and social psychological attitudes 1 61. Beaupr, M. G. (2005). Cross cultural emotion recognition among Canadian ethnic groups. Journal of Cross Cultural Psychology, 36 (3), 355. doi: 10.1177/0022022104273656 Blair, K. (2008). Response to emotional expressions in generalized social phobia and generalized anxiety disorder: Evidence for separate disorders. The American Journal of Psychiatry, 165 (9), 1193 1202. doi: 10.1176/appi.ajp.2008.07071060 Clark, D. M., & Wells, A. (1995). A cognitive m odel of social phobia. Social phobia: Diagnosis, assessment, and treatment 41 68 91. Dalgleish, T. (1995). Performance on the emotional stroop task in groups of anxious, expert, and control subjects: A comparison of computer and card presentation format s. Cognition and Emotion, 9 (4), 341 362. doi: 10.1080/02699939508408971 Gilboa Schechtman, E., Foa, E., Vaknin, Y., Marom, S., & Her mesh, H. (2008). Interpersonal sensitivity and response b ia s in social phobia and depression: Labeling e motional e xpressions. Cognitive Therapy & Research 32 (5), 605 618. doi: 10.1007/s10608 008 9208 8

PAGE 35

30 Heinrichs, N., & Hofman, S. G. (2001). Information processing in social phobia: A critical review. Clinical Psychology Review 21(5), 751 770. doi: 10.1016/j.bbr.2011.03.031 Johnson Laird, P., & Oatley, K. (1989). The language of emotions: An analysis of a semantic field. Cognition & Emotion, 3 (2), 81 123. doi: 10.1080/02699938908408075 Kim, S., & Oh, K. (2010). Information processing bias against emo tional facial expressions in social anxiety. Japanese Psychological Research 52 (1), 23 32. doi: 10.1111/j.1468 5884.2009.00419 .x Leary, M. R. (1983). A brief version of the fear of negative evaluation scale. Personality Social Psychology Bulletin, 9 (3), 3 71. doi: 10.1177/0146167283093007 Leber, S., Heidenreich, T., Stangier, U., & Hofmann, S. G. (2009). Processing of facial affect under social threat in socially anxious adults: mood matters. Depression & Anxiety (1091 4269) 26 (2), 196 206. doi: 10.1002/da.20525 MacLeod, C., & Mathews, A. (1988). Anxiety and the allocation of attention to threat. Quart erly Journal of Experimental Psychology: Human Experimental Psychology 38 659 670. Niedenthal, P. M., Halberstadt, J. B., Margolin, J., & Innes Ker, H. (2000). Emotional state and the detection of change in facial expression of emotion. European Journal o f Social Psychology 30 (2), 211 222. doi: 10.1002/(SICI)1099 0992(200003/04)30:2<211::AID EJSP988>3.0.CO;2 3

PAGE 36

31 Pineles, S. L., & Mineka, S (2005). Attentional biases to internal and external sources of potential threat in social anxiety. Journal of Abnormal Psychology, 114 (2), 314 318. doi: 10.1037/0021 843X.114.2.314 Pishyar, R., Harris, L. M., & Menzies, R. G. (2004). Attentional bias for words and faces in social anxiety. Anxiety, Stress & Coping: An International Journal, 17 (1), 23 36. doi: 10.1080/10615800310001601458 Taylor, C. T., Bomyea, J., & Amir, N. (2011). Malleability of attentional bias for positive emotional information and an xiety vulnerability. Emotion, 11 (1), 127 138. doi: 10.1037/a0021301 Watson, D., & Friend, R. (1969). Measurement of social evaluative anxiety. Journal of Consulting and Clinical Psychology, 33 (4), 448 457. doi: 10.1037/h0027806 Woody, S. R. (1996). Effects of focus of attention on anxiety levels and social performance of individuals with social phobia. Journal of Abnormal Psychology, 105 (1), 61 69. doi: 10.1037/0021 843X.105.1.61 Yang, Z., Jackson, T., Gao, X., & Chen, H. (2012). Identifying selective visual attention biases related to fear of pain by tracking eye movements within a dot probe paradigm. Pain, 153 (8), 1742 1748. doi: 10.1016/j.pain.2012.05.011

PAGE 37

32 Footnotes Participants were also separated into social anxiety groups based on the population FNE mean and standard deviation ( M = 16.5, SD = 7.39). Participants falling one standard deviation above the mean were in the high social anxiety group and participants falling one standard deviation below the mean were in the low social anxiety group. Remaining participants were placed in the moderate social anxiety group. Additional analyses were run to determine significance across these social anxiety groups for reaction time and accuracy for facial expressions and words. Results revea led that allocating participants to social anxiety groups based on the population FNE scores did not make any difference in obtaining significant results.

PAGE 38

33 Table 1: Characteristics of participants in each social anxiety (SA) group Self Report Measures Low SA Moderate SA High SA Gender Male 7 9 2 Female 12 23 22 Other 1 0 0 AOC Social Sciences 8 16 10 Humanities 6 7 4 Natural Sciences 4 8 8 Race White 13 19 15 Non white 4 4 7 Ethnicit y White 13 17 13 Non white 4 6 9

PAGE 39

34 Table 2: Mean reaction times (in milliseconds) for each facial expression (standard deviations in parentheses) Facial Expression Low SA Moderate SA High SA Happy 709.20 (120.05) 752.14 (148.82) 730.86 (143.02) Anger 759.84 (149.89) 745.97 (121.79) 794.12 (220.45) Fear 697.10 (134.92) 702.30 (127.03) 737.30 (162.65) Neutral 703.98 (138.03) 728.98 (138.59) 795.41 (154.73)

PAGE 40

35 Table 3: Mean reaction times (in milliseconds) of each facial expression in each block (standard deviations in parentheses) Block Low SA Moderate SA High SA Happy Happy 645.81 (112.27) 678.59 (120.23) 657.68 (151.43) Anger 676.66 (116.15) 701.91 (147.58) 658.69 (136.27) Fear 679.92 (116.83) 723.93 (143.29) 700.48 (131.74) Neutral 689.44 (128.29) 756.82 (161.04) 736.63 (176.63) Anger Happy 620.93 (120.66) 703.63 (145.60) 680.42 (192.05) Anger 775.97 (147.56) 729.17 (178.09) 746.28 (212.42) Fear 763.60 (175.49) 756.08 (179.24) 810.3 4 (206.00) Neutral 839.98 (207.03) 796.39 (176.00) 813.61 (221.42) Fear Happy 630.15 (155.21) 665.69 (138.51) 665.15 (143.39) Anger 807.26 (215.19) 780.72 (177.11) 814.41 (217.62) Fear 677.95 (99.90) 694.55 (160.93) 715.61 (198.16) Neutral 659.40 (189.08) 680.73 (191.36) 726.41 (167.96) Neutral Happy 605.93 (151.50) 694.93 (143.81) 738.44 (175.95) Anger 726.05 (163.63) 784.97 (174.33) 778.85 (176.57) Fear 640.21 (155.71) 680.59 (165.64) 766.79 (172.45) Neutral 790.47 (137.79) 745 .32 (152.30) 794.22 (165.54)

PAGE 41

36 Table 4: Mean accuracy scores for each facial expression (standard deviations in parentheses) Facial Expression Low SA Moderate SA High SA Happy 0.98 (0.03) 0.95 (0.05) 0.96 (0.06) Anger 0.82 (0.11) 0.84 (0.13) 0.84 (0.11) Fear 0.88 (0.14) 0.88 (0.15) 0.90 (0.09) Neutral 0.90 (0.09) 0.91 (0.07) 0.88 (0.08)

PAGE 42

37 Table 5: Mean reaction times (in milliseconds) of fear and anger words (standard deviations in parentheses) Emotional Word Low SA Moderate SA High SA Fear 1196.16 (192.49) 1142.67 (168.43) 1022.69 (192.56) Anger 1127.26 (166.67) 1179.21 (268.43) 1002.36 (150.12)

PAGE 43

38 Appendix A Fear of Negative Evaluation Scale (Watson & Friend, 1969)

PAGE 44

39 Appendix B Sample Photographs of Facial Stimuli

PAGE 45

40 Appendix C Word Stimuli Fear Neutral 1. Fear Copy 2. Dread Pivot 3. Apprehension Illustration 4. Consternation Differentiate 5. Afraid Chalky 6. Panic Purse 7. Terror Peanut 8. Scare Clock 9. Frighten Drilling 10. Worry Stool 11. Terrify Harvest 12. Petrify Artisan 13. Embarrassed Upholstered 14. Alarm Straw 15. Horror Jersey 16. Threaten Scaffold Anger Neutral 1. Scorn Light 2. Offend Barrel 3. Animosity Frequen cy 4. Indignant Manuscript 5. Cross Floor 6. Furious Pedaled 7. Peeve Glass 8. Irritate Chanting 9. Annoy Brick 10. Enrage Ladder 11. Infuriate Calculate 12. Aggravate Translate 13. Inflame Clothes 14. Irate Brake 15. Rage Corn 16. Mad Ru g

PAGE 46

1