Emotion attention and recognition of facial expressions among close friends and casual acquaintances

Abstract

The present study examined the link between emotion attention and accuracy of recognizing emotional facial expressions among close friend pairs and acquaintance pairs. Among acquaintance pairs, individuals high on emotion attention were more accurate than those low on emotion attention in reading the acquaintance’s negative expressions of sadness and anger but did not differ from the latter in reading the acquaintance’s happy expressions. Among close friend pairs, those high on emotion attention were also more accurate in reading the friend’s sad expressions and did not differ from the low emotion attention individuals in reading the friend’s happy expressions. However, they were actually less accurate in reading the friend’s angry expressions. This lower accuracy reflected a perceptual bias of systematically misperceiving the friend’s angry expressions as neutral or sad. This finding suggests a defensive mechanism in the perception of the friend’s angry expressions among individuals high on emotion attention. Overall, the findings highlight the importance of emotion attention in emotion recognition and the important role of the relationship context in this link.

Keywords

Casual relationships close relationships defensive mechanism emotion emotion attention facial expressions recognition accuracy

Facial expressions are ubiquitous in everyday life and constitute an important channel of nonverbal communication. People express feelings, desires, needs, and wants through the face; their facial expressions are often read by others as revealing their innermost emotions and intentions which in turn guide others’ reactions toward them. Therefore, to be able to decode and read facial expressions accurately is vital in social understanding and interaction. Severe impairment of this ability, such as in autism, schizophrenia, or depression, is a sign of a disorder and impedes normal social interaction and relationships (Bo’lte & Poustka, 2003). Expression and recognition of facial expressions is also an intrinsically social event (Mehta & Clark, 1994). People experience many emotions in the context of social relationships and report feeling more joy, anger, and sadness in the company of others than when alone in nonsocial settings (Babad & Wallbott, 1986; Csikszentmihalyi & Larson, 1984). Thus, how relationship contexts influence recognition accuracy is an interesting research question to explore. The present study investigated individual differences in recognizing facial expressions and the role of interpersonal relationships in recognition accuracy.

Over the past several decades, many studies have examined individual differences in accurately identifying emotional facial expressions and their relationships with personality correlates, but the results were mixed. For example, studies found that high self-monitoring individuals were more accurate than low self-monitors in reading facial expressions (Mufson & Nowicki, 1991; Zuckerman, Larrance, Hall, DeFrank, & Rosenthal, 1979). Field independents were found to be more accurate than field dependents at interpreting strangers’ facial expressions (Bastone & Wood, 1997). However, these results were not replicated in other studies (Cunningham, 1977; Riggio & Friedman, 1982; Sabatelli, Dreyer, & Buck, 1979). Recognition accuracy has also been linked to the big five personality traits and was reported to be positively correlated with openness, conscientiousness, extraversion, and neuroticism (Buck, Savin, Miller, & Caul, 1972; Cunningham, 1977; Matsumoto et al., 2000; Zuckerman et al., 1979). Even though the big five results were promising, they were not consistently found in other studies as well (e.g. Bastone & Wood, 1997; Cunningham, 1977; Matsumoto et al., 2000; Riggio & Friedman, 1982). In short, the inconsistent results in the previous research indicate there is a strong need to further investigate the relationship between individual differences in recognition of emotional facial expressions and personality correlates.

In our view, a key personality characteristic that was not directly investigated in the previous research but is intuitively related to recognition accuracy is attention to emotion. Attention to emotion (or emotion attention) reflects the degree to which individuals attend to and monitor their own feelings, generally regarded as an important aspect of emotional intelligence (Salovey, Mayer, Goldman, Turvey, & Palfai, 1995). Individuals who score high on emotion attention tend to pay close attention to feelings in not only oneself but also others and closely monitor these feelings (Davies, Stankov, & Roberts, 1998; Mayer, Salovey, Gomberg-Kaufman, & Blainey, 1991). Research has suggested that emotion attention plays an important role in accurately recognizing facial expressions and is vital in social interaction. Severe impairment in this ability disrupts normal interpersonal communication and often leads to difficulties in social relationships. For example, studies with individuals suffering from alexithymia found that these individuals showed marked deficiency in attending to and identifying emotions in both themselves and others and consequently experience a great deal of interpersonal problems and difficulties (Taylor, Bagby, Parker, & Grotstein, 1999). Studies with children and adults with attention-deficit hyperactivity disorder (ADHD) also found that the difficulty and poor social skills that individuals with ADHD showed in identifying emotions was linked to their failure to attend to appropriate emotional cues in both selves and others (e.g. Cadesky, Mota, & Schachar, 2000; Miller, Hanford, Fassbender, Duke, & Schweitzer, 2011). Children and adults with ADHD often struggle with relationships and have difficulties in maintaining relationships (Goodman, 2007). Thus, to be able to pay appropriate attention to emotional cues of both oneself and others is essential for accurate reading of facial expressions and important for normal social functioning. In addition to clinic research, personality research linking emotion attention to the big five personality traits found that low conscientiousness was related to poor goal-directed attention deployment (John & Gross, 2007; Nigg et al., 2002) and high levels of openness and neuroticism was related to greater emotional attention and awareness (Costa & McCrae, 1992; Extremera & Fernández-Berrocal, 2005). Given that high levels of conscientiousness, openness, and neuroticism were previously found to be associated with greater accuracy in recognizing facial expressions (at least in some studies, as reported earlier), it is reasonable to expect that high emotion attention would be associated with greater accuracy in recognizing facial expressions. In other words, individuals scoring high on emotion attention, who tend to be highly attentive to feelings, should be more accurate in reading facial expressions than individuals scoring low in emotion attention.

This general relationship between emotion attention and recognition accuracy, however, needs to be considered in the context of social relationships. Clark and Mills (1993, 2012) distinguished between two types of interpersonal relationships—communal and exchange relationships—and suggested that emotional experience in these two relationship types tended to differ (Mehta & Clark, 1994). In mutual communal relationships, individuals are genuinely concerned with one another’s welfare and pay attention to the other person’s needs and respond to these needs by engaging in behavior that benefits the other person. There is no expectation of receiving benefits in return. Individuals in strong mutual communal relationships are emotionally invested in one another. The well-being of the other person and the relationship is important to them. Close relationships, such as close friends, romantic relationships, and relationships among adult family members, are considered to be examples of strong communal relationships. In contrast to communal relationships, individuals in mutual exchange relationships do not feel particularly concerned with one another’s welfare, do not feel obligated to care for the other person, and expect favors and benefits to be returned. Individuals in mutual exchange relationships are not particularly responsive to one another’s needs and are not affected by the emotional needs of the other person as much as individuals in mutual communal relationships are. Examples of exchange relationships include relationships with classmates, coworkers, or other casual acquaintances.

These differing relationships suggest that goals and motivation for accurately recognizing facial expressions would differ for individuals in a close, communal relationship as opposed to in a casual, exchange relationship. Indeed, previous research has shown that closeness of a relationship influences how accurately individuals in the relationship recognize one another’s expressions. Whereas some research indicates that interpersonal closeness generally fosters greater accuracy in reading other’s expressions because individuals in close relationships are more familiar with one another’s emotional expressions (Beale & Keil, 1995; Zhang & Parmley, 2011) and are more motivated to accurately read the expressions (Thibault, Bourgeois, & Hess, 2006); other research indicates that closeness sometimes may induce a motivation to misread other’s emotions if these emotions are perceived to violate the relationship norm or are potentially harmful to the relationship (e.g. Sternglanz & DePaulo, 2004). Thus, close relationships have a strong relationship effect on recognition accuracy and pull for typically high but at times low recognition accuracy from individuals, regardless of their personality characteristics. This strong relationship effect suggests that personality characteristics of the individual—in this case, emotion attention—may be less predictive of recognition accuracy in close relationships than in casual exchange relationships.

The present study examined these hypotheses by comparing recognition accuracy in close friend pairs with accuracy in casual acquaintance pairs in a group of college students (relationship pair was a between-subjects variable). Close friendships and casual acquaintanceships represent, respectively, a mutual communal relationship and a casual exchange relationship. These two relationships are representative of college students’ social life as they typically have a large number of acquaintances and one or a few close friends. The present study focused specially on female students because substantial amount of evidence indicates that females experience greater intimacy and emotional closeness in friendships than males do (Jones, 1991; Wright, 1988). Research has shown that greater intimacy in female friendships begins at an early age and has been linked to evolutionary mechanisms (Hall, 2011). Because of females’ greater investment in offspring, Taylor et al. (2000) suggest that females respond to stress with a tend-and-befriend response. Befriending others involves pursuing supportive relationships as a form of protection from threat. This suggests that the difference in reading the emotional cues of close friends and acquaintances may be the greatest for women due to the greater intimacy in female close relationships.

The present study examined recognition of angry, sad, and happy expressions in these relationship pairs since these three emotions are more commonly communicated in interpersonal relationships and studied more in research than other emotions (Lipp, Price, & Tellegen, 2009; Sternglanz & DePaulo, 2004). Given that naturally occurring expressions in daily life are mostly subdued and low to moderate in intensity, the present study used the morph paradigm to create facial expressions that varied in intensity from low to moderately high. The morph paradigm uses each person’s neutral expression as a beginning anchor and his/her full, highly intense emotional expression as the final anchor to create facial composites of increasing intensities (e.g. Hess, Blairy, & Kleck, 1997). We selected the 15th, 30th, 45th, 60th, and 75th frames from the 100 facial composites to represent, respectively, expressions of 15%, 30%, 45%, 60%, and 75%¹ (e.g. from low to moderately high intensity) intensity of each emotional expression for each person.

To summarize, we predicted that high emotion attention would predict greater accuracy in recognition of facial expressions, but this effect would be moderated by the type of the relationship in which facial recognition takes place—specifically, emotion attention would predict individual differences in recognition accuracy more strongly in acquaintance pairs than in close friend pairs because close friends generally are highly motivated and attentive to one another’s emotional needs, regardless of whether the individuals are high or low on emotion attention. In addition to these two main predictions, we also predicted that emotion attention would predict differences in recognizing angry or sad expressions but not happy expressions. Happy expressions are easily recognized and were found to be decoded with high accuracy even at reduced intensities (Matsumoto et al., 2000; Thibault et al., 2006; Zhang & Parmley, 2011). The ceiling effect in decoding happy expressions may preclude variability of recognition accuracy.

Method

Participants

Forty-three pairs of female undergraduate close friends (n = 86) and 50 pairs of female undergraduate casual acquaintances (n = 99, one pair had missing data from one partner) participated for extra credit; thus a total of 185 participants completed the emotion recognition task. The acquaintance participants signed up for the study individually and listed classmates whom they considered as casual acquaintances. They were then paired randomly. Close friends signed up together or signed up individually and agreed to bring a close friend. Participants classified as close friends knew their partner for several months (M = 40.1 months, SD = 32.7) and perceived each other as close (M = 3.6, SD = 0.6, on a 4-point self-reported closeness scale). Participants ranged in age from 16 to 24 years (M = 19.58, SD = 1.37). All participants served as both the judge and the target in the facial recognition task. In total, 93.7% of participants were Caucasian, 1.6% were Black, 1.6% were Hispanic, 1.1% were Asian, 0.5% were Native American, and 1.6% did not report their ethnicity. Since the sample was homogeneous, we did not control for culture or race in the analyses.

Procedure

Participants attended two sessions, approximately 1 month apart. During the first session, consent was obtained and happy, angry, and sad expressions were elicited in each participant and photographed. Participants also completed a battery of questionnaires. Participants returned for the second session in small groups. In the second session, they viewed and labeled facial expressions of their paired partner. Participants in the close friend pairs viewed expressions of their friend, whereas participants in the acquaintance pairs viewed expressions of the paired acquaintance. All of the participants viewed the paired partner’s expressions mixed with expressions of a female stranger. Finally, they were debriefed and dismissed. All ethical guidelines were respected during the course of the study.

Apparatus and materials

Photos of facial expressions

Each participant underwent an emotion induction procedure in the first session. She displayed a neutral expression, which was photographed, and then was asked to recall as vividly as possible an event that made her feel really angry, sad, or happy, reexperience the emotion, and then show the emotion through her facial expression, which was photographed.

Participants’ expressions were rated by judges who were unaware of the purpose of the study, using a 7-point intensity scale (1 = not at all to 7 = extremely). The mean intensity rating was M = 4.06, SD = 1.61 (α = .87 for interjudge reliability) for anger, M = 3.33, SD = 1.44 (α = .89) for sadness, and M = 4.85, SD = .86 (α = .88) for happiness. No difference in the intensity ratings was found between close friends and acquaintances.

Emotion recognition task

The photos were processed using morphing software. The morphing software uses each person’s neutral expression as the beginning anchor and his/her clear full emotional (happy, sad, or angry) expression as the final anchor to produce a digital 640 × 480 pixel movie composed of 100 facial composites for each of the emotional expressions, incrementally increasing in intensity. These expressions were used for the recognition task and conducted in E-Prime 1.0 (2002). Each participant viewed five intensities (15%, 30%, 45%, 60%, and 75%) of angry, sad, and happy expressions of her paired partner (either close friend or acquaintance), mixed with expressions of a female stranger,² and interspersed with neutral expressions of both the partner and the stranger (used as filler). The participants viewed each expression in randomized order and pressed a designated key (anger, sadness, happiness, and neutral) on the keyboard to label the expression. Each participant viewed 160 expressions in all, half exhibited by his/her paired partner (3 Emotions [sadness, anger, and happiness] × 5 Intensity Levels [15%, 30%, 45%, 60%, and 75%] = 15 expressions, each shown 4 times, plus the neutral expression, shown 20 times) and half by the stranger. Prior to completing the emotion recognition task, participants practiced by labeling the expressions of a male stranger.

Emotion attention

Emotion attention was measured using the Attention to Feelings subscale of the Trait Meta-Mood Scale (Salovey et al., 1995).³ This subscale consists of 13 items, such as “I pay a lot of attention to how I feel.” Although this scale measures attention to one’s own feelings, evidence has shown that appraisal of one’s own feelings is related to appraisal of others’ feelings (Mayer et al., 1991). Ratings were summed, with higher scores indicating greater attention to feelings. The scale is reported to have good internal reliability (Salovey et al., 1995) and α = .77 in the present study. No difference was found between the close friends and the casual acquaintances on emotion attention, t(183) = 1.12, n.s.

Data preparation

Table 1 summarizes the means and standard deviations of recognition accuracy (in raw percentage) for close friend and casual acquaintance pairs. As can be seen in Table 1, there is a ceiling effect in the recognition of happiness. Using the raw percentage, however, can represent a biased measure of accuracy because it fails to account for false positives. As such, an unbiased hit rate (H _u) was used to measure recognition accuracy (Wagner, 1993). H _u was computed by multiplying the raw hit rate (e.g. the number of times sadness was correctly identified and divided by the total number of times sad expressions were presented) and differential accuracy (e.g. the number of times sadness was correctly identified divided by the number of times participants used the sadness label across all stimuli).

Table 1.

Recognition accuracy (raw percentage) by relationship pairs and emotional expressions.

	Close friends	Casual acquaintances
Sad	55.45 (2.60)	43.14 (2.55)
Angry	59.08 (2.31)	47.21 (2.62)
Happy	82.27 (1.12)	81.49 (1.19)

Note. The numbers above represent the raw percentage of correct responses among close friends and acquaintances for each emotion. Standard errors are given in parentheses.

Results

Because the data has a three-level structure consisting of recognition accuracies at the five intensities (Level 1) nested within participants (Level 2) nested within pairs (Level 3), three-level hierarchal linear modeling was conducted (Raudenbush & Bryk, 2002).⁴ To test our hypotheses, the Level-1 model was specified as an unconditional model, as shown below.⁵

Y_{i j k} = π_{0 j k} + e_{i j k},

where Y_ijk represents recognition accuracy of a particular emotion at intensity level i of person j in pair k, π_0jk is the mean recognition accuracy of the person for that emotion, averaging across the five intensity levels, and e_ijk is a random observation effect.

The Level-2 model predicts the mean recognition accuracy of each person from her emotion attention score (grand centered). The model was specified as follows:

π_{0 j k} = β_{00 k} + β_{01 k} {(e m o t i o n a t t e n t i o n)}_{j k} + r_{0 j k},

where β_00k is the mean recognition accuracy of pair k, β_01k represents the effect of emotion attention of person j on recognition accuracy, and r _0jk is a random person effect.

Level-3 models compare close friend pairs with acquaintance pairs. Because we were mainly interested in comparing the effect of emotion attention in the two types of relationships, we created a dummy variable, relationship pair (0 = casual acquaintance pairs; 1 = close friend pairs) to differentiate the two relationship status. Relationship pair was used to predict the effect of emotion attention on recognition accuracy at Level 3. The Level-3 models were specified as follows:

β_{00 k} = γ_{000} + γ_{001} {(r e l a t i o n s h i p p a i r)}_{k} + u_{00 k},

β_{01 k} = γ_{010} + γ_{011} {(r e l a t i o n s h i p p a i r)}_{k} + u_{01 k},

where γ₀₀₀ represents the grand mean recognition accuracy of acquaintance pairs (relationship pair = 0), γ₀₀₁ represents grand mean accuracy differences between acquaintance and close friend pairs, γ₀₁₀ represents the effect of emotion attention on recognition accuracy among acquaintance pairs, γ₀₁₁ represents the recognition accuracy difference between acquaintance and close friend pairs in relation to emotion attention, and u _00k and u _01k represent random pair effects.

The following mixed model summarizes the above-mentioned three-level models.

\begin{aligned} Y_{i j k} = γ_{000} + γ_{001} {(r e l a t i o n s h i p p a i r)}_{k} + γ_{010} {(e m o t i o n a t t e n t i o n)}_{j k} \\ + γ_{011} {(e m o t i o n a t t e n t i o n)}_{j k} {(r e l a t i o n s h i p p a i r)}_{k} \\ + r_{0 j k} + u_{00 k} + u_{01 k} {(e m o t i o n a l a t t e n t i o n)}_{j k} + e_{i j k} . \end{aligned}

Analyses were run separately for each of the three emotions. Results for the fixed effects are presented in Table 2. As shown, the γ₀₀₀ coefficients were positive and statistically significant for all three emotions, indicating that the grand mean recognition accuracy among acquaintance pairs was significantly higher than 0 for all emotions. The γ₀₀₁ coefficients were also positive and statistically significant for all three emotions, indicating that, compared to acquaintance pairs, close friend pairs showed significantly higher recognition accuracy, suggesting that close friends perceived one another’s expressions more accurately than acquaintances did.

Table 2.

Fixed effects of three-level hierarchical linear modeling predicting recognition accuracy from EA and RP.

	Sad			Angry			Happy
	Coefficient	t	Effect size^a	Coefficient	t	Effect size	Coefficient	t	Effect size
Intercept γ₀₀₀	.19	16.16***	.86	.28	15.65***	.85	.64	34***	.96
RP γ₀₀₁	.09	3.51***	.34	.11	3.55***	.35	.06	2.26*	.23
EA γ₀₁₀	.005	2.32*	.24	.01	2.66**	.27	−.005	−1.34
EA × RP γ₀₁₁	−.001	−0.24		−.02	−3.65***	.36	.003	0.55

Note. Relationship pair was a dummy variable with 0 being acquaintances and 1 being close friends. df = 92 dyads; EA: emotion attention; RP: relationship pair.

^aEffect sizes were computed with the formula: r = square root of (t ²/t ² + df; Rosenthal & Rosnow, 1984) and reported only for significant effects.

*p < .05; **p < .01; ***p < .001.

Consistent with the hypotheses, the γ₀₁₀ coefficient of emotion attention was positive and statistically significant for sad expressions, indicating that higher emotion attention predicted greater accuracy in recognizing sad expressions among acquaintance pairs. The γ₀₁₁ coefficient of Emotion Attention × Relationship Pair interaction was not significant for sad expressions, suggesting that the effect of emotion attention did not differ between acquaintance and close friend pairs; in other words, higher emotion attention similarly predicted greater accuracy in recognizing sad expressions for both close friend and acquaintance pairs. For angry expressions, the γ₀₁₀ coefficient was also positive and statistically significant, suggesting again higher emotion attention predicted greater accuracy in recognizing angry expressions among acquaintance pairs. However, interestingly, the γ₀₁₁ coefficient was negative, indicating that the effect of emotion attention on recognition accuracy differed for close friends and acquaintances. The negative γ₀₁₁ coefficient indicated that while higher emotion attention predicted greater accuracy among acquaintance pairs, higher emotion attention actually predicted less accuracy among close friends in recognizing the friend’s angry expressions. Finally, for happy expressions, neither γ₀₁₀ nor γ₀₁₁ was significant, indicating that emotion attention had no effect on the recognition of happy expressions for both close friend and acquaintance pairs.

Mislabeling of emotional expressions

To better understand these findings, additional analyses were computed on incorrect responses to assess the pattern of mislabeling emotional expressions (e.g., mislabeling angry expressions as sad, happy, or neutral) among close friends and acquaintances. Table 3 summarizes the means and standard deviations of mislabeling (in raw percentage) each emotion for each relationship. A series of multiple regression analyses were run, separately for each emotion, predicting the mislabeling of the emotional faces (using H _u scores) from emotion attention (grand centered), relationship pair, and the interaction between the two variables. No significant main effects or interaction effect of emotional attention and relationship pair was found for happy expressions. However, significant main effects and/or an interaction were found for sad and angry expressions. As shown in Table 4, relationship pair status significantly predicted the tendency to mislabel sad expressions as neutral. Specifically, close friends were less likely than acquaintances to mistake one another’s sad expressions for neutral expressions. Close friend and acquaintance pairs did not differ in the tendency to mislabel sad expressions as either angry or happy expressions. An interesting pattern of mislabeling was found for angry expressions. A main effect of emotion attention was found in the neutral and sad mislabeling of the angry expressions qualified by a significant interaction between emotion attention and relationship pair. The results indicate that high emotion attention predicted fewer errors in mislabeling the acquaintance’s angry expressions as either neutral or sad among acquaintances. However, among close friends, high emotion attention predicted more errors in mislabeling the friend’s angry expressions as neutral or sad.

Table 3.

Incorrect emotion labeling of the expressions (raw percentage) by relationship pairs and emotions.

Expressions	Sad label	Angry label	Happy label	Neutral label
Close friends
Sad		17.33 (1.95)	1.45 (0.36)	28.02 (1.84)
Angry	14.07 (1.58)		1.80 (0.65)	29.77 (1.87)
Happy	2.50 (0.51)	0.52 (0.19)		14.42 (1.07)
Casual acquaintances
Sad		18.35 (1.77)	1.46 (0.30)	39.90 (1.83)
Angry	16.41 (1.62)		3.35 (0.70)	37.38 (1.97)
Happy	2.62 (0.73)	1.31 (0.33)		14.95 (0.98)

Note. The numbers above represent raw percentage of mislabels among close friends and acquaintances for each emotion. Standard errors are given in parentheses.

Table 4.

Multiple regression analyses predicting mislabeling of sad and angry expressions from EA, RP, and EA × RP.

	Predictor coefficients
Label	Emotion attention	Relationship pair	EA × RP	R ²
Mislabeling sad expressions
Neutral	−.001	−.014**	.001	.07**
Angry	.001	−.002	.000	.01
Happy	<.0001	.00	<.0001	.002
Mislabeling angry expressions
Neutral	−.001*	−.008	.003**	.06**
Sad	−.002*	−.014	.004*	.04^†
Happy	.000	−.003	.000	.01

Note. Relationship pairs were a dummy variable with 0 being acquaintances and 1 being close friends; EA: emotion attention; RP: relationship pair.

^† p < .10; *p < .05; **p < .01.

Discussion

The present study examined the link between emotion attention and recognition accuracy in two relationship contexts, among close friend pairs and casual acquaintance pairs. Consistent with the predictions, individuals in acquaintance pairs who scored high on emotion attention were more accurate than those scoring low on emotion attention in reading the other person’s sad and angry facial expressions, and emotion attention did not predict accuracy in reading happy expressions. Casual acquaintance relationships are a type of exchange relationship (Clark & Mills, 1993, 2012). Acquaintances are individuals whom we know or socialize with but do not consider as close, such as classmates, coworkers, and neighbors. People generally have little emotional involvement with acquaintances, do not feel obligated to care for them, and are not particularly concerned about the future of the relationship. Our results indicate that in such casual relationships, emotion attention of the individual influences how accurately he or she reads the other’s negative emotional expressions. Those high on emotion attention are more skilled than low counterparts in reading acquaintances’ sad or angry expressions (Cadesky et al., 2000; Matsumoto et al., 2000).

Among close friend pairs, an interesting pattern emerged. Similar to the acquaintances, individuals who scored high on emotion attention were found to be more accurate than those scoring low on emotion attention in reading the friend’s sad expressions. However, in reading the friend’s angry expressions, interestingly, individuals high on emotion attention actually were less accurate than their low counterparts. This result is surprising because close friends were found to be better skilled and more accurate in reading one another’s expressions than acquaintances in general. Our data showed, for example, that close friend pairs were more accurate than acquaintance pairs in reading all three emotional (i.e. sad, angry, or happy) expressions, consistent with the closeness advantage reported in the previous research (Zhang & Parmley, 2011). Close friends were also found to make fewer errors than acquaintances in confusing their partner’s sad expressions with her neutral expressions. If the skill levels of close friends are generally high in reading one another’s expressions, then this surprising finding suggests that factors other than recognition skills affect the reading of angry expressions. Our analyses with mislabeling of the angry expressions provided insights into this hidden factor. Emotion attention and relationship pair were found to interact to predict mislabeling of the angry expressions. High emotion attention predicted fewer errors in mislabeling the acquaintance’s angry expressions among acquaintance pairs. Individuals high on emotion attention in the acquaintance pairs made fewer errors than their low counterpart of mistaking the acquaintance’s angry expressions for neutral or sadness, in line with their generally better recognition skills. However, in close friends pairs, the pattern was opposite. High emotion attention predicted more errors in mislabeling the friend’s angry expressions. Specifically, individuals high on emotion attention made more errors of mistaking the friend’s angry expressions for neutral or sad than those low on emotion attention did. Thus, the lower accuracy of these individuals in reading the friend’s angry expressions reflects a systematic perceptual bias to misperceive the friend’s angry expressions as neutral or sad. These results suggest a defensive mechanism among them in identifying the friend’s angry expressions. Angry expressions communicate hostility and threat and are hurtful to the perceiver in close relationships (e.g. Fischer & Manstead, 2008). Individuals high on emotion attention may have misperceived the friend’s anger as neutral or sad to reduce anxiety. This interpretation is consistent with Ickes and Simpson’s (1997) model of motivated inaccuracy. According to them, motivated inaccuracy is more likely to occur when there are ambiguous cues in situations where the perceiver may perceive threat. In close friendships, instances of anger may be both confusing and upsetting at the same time because close friends usually feel a great deal of affection and attachment toward one another, thus creating an ambiguous situation for misperception of the expressions. This may be particularly true for individuals high on emotion attention because they are more sensitive to emotional cues, in contrast to individuals low on emotion attention, who tend to be less attentive in monitoring emotions.

Another possible explanation for this surprising result is that anger may be particularly important in relationships (Sternglanz & DePaulo, 2004). As discussed earlier, anger conveys hostility and aggression, different from sadness, which conveys submissiveness and elicits comforting behavior (Hendriks, Nelson, Cornelius, & Vingerhoets, 2008). Frequent expressions of anger are found to be associated with negative relationship outcomes (e.g. Fischer & Manstead, 2008). Thus, angry expressions are more hurtful to the perceiver in close relationships and poses as a greater threat to the relationship. Thus, individuals high in emotion attention may have a greater motivation to mislabel angry expressions when reading a close friend in order to preserve the friendship. This may be particularly true for women who are more concerned about relational contexts (Timmers, Fischer, & Manstead, 1998). Consistent with this premise, women modify their emotional expressions in order to maintain close relationships and have different emotion regulation goals than men (Deaux & Major, 1987; Timmers et al., 1998). For example, women have been found to modify their expressions of anger depending on the relationship context (Fischer & Evers, 2011) but not other emotions like sadness (Timmers et al., 1998). Women may not only modify their expressions of anger in relationships, but findings from the present study suggest that they may be motivated to “misperceive” angry expressions in those who are close to them to something less threatening to protect the relationship, particularly if they are high on emotion attention.

The present finding highlights the importance of the relationship context in the recognition of facial expressions. In close relationships, individuals are heavily invested in one another emotionally and identify themselves with the relationship, unlike in causal acquaintanceships (Aron, 2003; Aron, Aron, Tudor, & Nelson, 1991; Clark & Mills, 1993, 2012; Clark, Mills, & Corcoran, 1989). Relationship goals trump self-goals in close relationships and the individual’s desire to maintain and promote the relationship may introduce biases in perceiving and interpreting the other’s behavior, especially when the behavior is deemed threatening and potentially harmful to the relationship (Sternglanz & DePaulo, 2004). The present finding leads support to our general premise that different relationship contexts introduce goals and motivations that moderate the connection between emotion attention and recognition accuracy, although it shows that emotion attention is associated with recognition accuracy in a more complex way in close friendships than in acquaintance relationships, different from our prediction of a mere reduced effect of emotion attention in close friendships. The fact that individuals high on emotion attention in close friend pairs were more accurate than those scoring low on emotion attention in reading the friend’s sad expressions corroborates our results for the acquaintance pairs and lends further support to the important role of emotion attention in emotion recognition. The inability to attend to appropriate emotional cues has been linked to difficulties in identifying emotions among individuals suffering from alexithymia or ADHD (Cadesky et al., 2000; Miller et al., 2011; Taylor et al., 1999). More recently, research on mindfulness has shown that mindfulness, a state of consciousness characterized by attention to the present experience with open acceptance, is associated with reduced stress and increased well-being. Mindfulness training has been linked to a host of positive effects including increasing self-awareness, improving attention, increasing interpersonal skills and relationships, and more positive emotions (Brown, Ryan, & Creswell, 2007; Smalley & Winston, 2010). The present finding suggests that attention to emotion underlies individual differences in accuracy of emotion recognition in both casual acquaintance relationships and close friendships, with individuals high on emotion attention showing greater accuracy by and large, and the beneficial effects of mindfulness training on interpersonal relationships may be due partly to the effect of increased emotion attention.

Finally, as predicted, emotion attention did not predict accuracy of reading happy expressions among both close friend pairs and acquaintance pairs. No systematic mislabeling of happy expressions was found either among close friends or among acquaintances. As can be seen by the mean recognition accuracy in Table 1, participants were highly accurate in identifying happiness (raw accuracy percentages were above 80 for happy expressions). Ceiling effects probably precluded detecting individual differences (Matsumoto et al., 2000; Thibault et al., 2006).

Three limitations of the present study should be mentioned. First, we examined expression recognition only among female undergraduates because of the greater closeness found in female friendships (Jones, 1991). It is unclear, however, whether these findings would generalize to male friendships or cross-sex friendships. Future studies need to expand the investigation to samples that include both males and females and also age-groups other than the undergraduate population. Second, we tested only the recognition of sad, angry, and happy facial expressions. Given that recognition may vary as a function of emotion, future studies should expand the investigation to include other emotions such as fear, surprise, and disgust to provide additional insights. Third, the present study used a self-report measure of emotion attention that focuses on attention to self-feelings. Attention to self-feelings was previously found to be interconnected with appraisal of feelings of others (Mayer et al., 1991). Individuals who show deficiency in identifying self-feelings have also been found to have difficulty in perceiving and identifying feelings of others (Taylor et al., 1999). Nevertheless, a more inclusive and more behavior-based measure of emotion attention that assesses both attention to feelings of self and others would be desirable.

Despite these limitations, the present findings suggest that emotion attention is an important psychological characteristic that links to recognition accuracy in reading expressions. Individuals higher on emotion attention in general were more accurate in identifying others’ negative expressions. But the interpersonal context in which the recognition takes place also influences accuracy, and in close relationships, when expressions signal threat, individuals high on emotion attention showed defensive inaccuracy in recognition. Given that emotion recognition takes place in interpersonal settings, the role of relational context and motivations in emotion recognition needs to be further explored.

Footnotes

Funding

This research was supported by a Psi Chi Faculty Advisor Research Grant and an Assumption College Faculty Development Grant awarded to Fang Zhang.

Notes

References

Aron

(2003). Self and close relationships. In Leary

M. R.

Tangney

J. P.

(Eds.), Handbook of self and identity (pp. 442–461). New York, NY: Guilford Press.

Aron

E. N.

Tudor

Nelson

(1991). Close relationships as including other in the self. Journal of Personality and Social Psychology, 60, 241–253. doi:10.1037/0022-3514.60.2.241

Babad

E. Y.

Wallbott

H. G.

(1986). The effects of social factors on emotional relations. In Scherer

K. R.

Wallbott

H. G.

Summerfield

A. B.

(Eds.), Experiencing emotion: A cross-cultural study (pp. 154–172). Cambridge, UK: Cambridge University Press.

Bastone

L. M.

Wood

H. A.

(1997). Individual differences in the ability to decode emotional facial expressions. Psychology: A Journal of Human Behavior, 34, 32–36.

Beale

J. M.

Keil

F. C.

(1995). Categorical effects in the perception of faces. Cognition, 57, 217–239. doi:10.1016/0010-0277(95)00669-X

Bo’lte

Poustka

(2003). The recognition of facial affect in autistic and schizophrenic subjects and their first degree relatives. Psychological Medicine, 33, 907–915. doi:10.1017/S0033291703007438

Brown

K. W.

Ryan

R. M.

Creswell

J. D.

(2007). Mindfulness: Theoretical foundations and evidence for its salutary effects. Psychological Inquiry, 18, 211–237. doi:10.1080/10478400701598298

Buck

R. W.

Savin

W. J.

Miller

R. E.

Caul

W. F.

(1972). Communication of affect through facial expressions in humans. Journal of Personality and Social Psychology, 23, 362–371. doi:10.1037/h0033171.

Cadesky

E. B.

Mota

V. L.

Schachar

R. J.

(2000). Beyond words: How do children with ADHD and/or conduct problems process nonverbal information about affect? Journal of the American Academy of Child and Adolescent Psychiatry, 39, 1160–1167. doi:10.1097/00004583-200009000-00016

10.

Clark

M. S.

Mills

(1993). The difference between communal and exchange relationships: What it is and is not. Personality and Social Psychology Bulletin, 19, 684–691. doi:10.1177/0146167293196003

11.

Clark

M. S.

Mills

(2012). A theory of communal (and exchange) relationships. In Van Lange

P. A. M.

Kruglanski

A. W.

Higgins

E. T.

(Eds.), Handbook of theories of social psychology (Vol. 2, pp. 232–250). Thousand Oaks, CA: Sage.

12.

Clark

M. S.

Mills

Corcoran

(1989). Keep track of needs in communal and exchange relationships. Personality and Social Psychology Bulletin, 15, 533–542. doi:10.1177/0146167289154007

13.

Costa

P. T.

McCrae

R. R.

(1992). NEO PI-R professional manual. Odessa, FL: Psychology Assessment Resources.

14.

Csikszentmihalyi

Larson

(1984). Being adolescent: Conflict and growth in teenage years. New York, NY: Basic books.

15.

Cunningham

M. R.

(1977). Personality and the structure of the nonverbal communication of emotion. Journal of Personality, 45, 564–584. doi:10.1111/1467-6494.ep7380563.

16.

Davies

Stankov

Roberts

R. D.

(1998). Emotional intelligence: In search of an elusive construct. Journal of Personality and Social Psychology, 75, 989–1015. doi:10.1037/0022-3514.75.4.989

17.

Deaux

Major

(1987). Putting gender into context: An interactive model of gender-related behavior. Psychological Review, 94, 369–389. doi:10.1037/0033-295X.94.3.369

18.

E-Prime (Version 1.0) [Computer program]. (2002). Pittsburgh, PA (http://www.pstnet.com): Psychology Software Tools.

19.

Extremera

Fernández-Berrocal

(2005). Perceived emotional intelligence and life satisfaction: Predictive and incremental validity using the Trait meta-mood scale. Personality and Individual Differences, 39, 937–948. doi:10.1016/j.paid.2005.03.012

20.

Fischer

A. H.

Evers

(2011). The social costs and benefits of anger as a function of gender and relationship context. Sex Roles, 65, 23–34. doi:10.1007/s11199-011-9956-x

21.

Fischer

A. H.

Manstead

A. S. R.

(2008). Social functions of emotion. In Lewis

Haviland-Jones

J. M.

Barrett

L. F.

(Eds.), Handbook of emotion (3rd, pp. 456–470). New York, NY: Guilford Press.

22.

Goodman

D. W.

(2007). The consequences of attention-deficit hyperactivity disorder in adults. Journal of Psychiatric Practice, 13, 318–327. doi:10.1097/01.pra.0000290670.87236.18

23.

Hall

J. A.

(2011). Sex differences in friendship expectations: A meta-analysis. Journal of Social and Personal Relationships, 28, 723–747. doi:10.1177/0265407510386192

24.

Hendriks

M. C. P.

Nelson

J. K.

Cornelius

R. R.

Vingerhoets

A. J. J. M.

(2008). Why crying improves our well-being: An attachment-theory perspective on the functions of adult crying. In Vingerhoets

Nyklicek

Denollet

(Eds.), Emotion regulation: Conceptual and clinical issues (pp. 87–96). New York, NY: Springer Science.

25.

Hess

Blairy

Kleck

R. E.

(1997). The intensity of emotional facial expressions and decoding accuracy. Journal of Nonverbal Behavior, 21, 241–257.

26.

Ickes

Simpson

J. A.

(1997). Managing empathic accuracy in close relationships. In Ickes

(Ed.), Empathic accuracy (pp. 218–250). New York, NY: Guilford Press.

27.

Jones

D. C.

(1991). Friendship satisfaction and gender: An examination of sex differences in contributors to friendship satisfaction. Journal of Social and Personal Relationships, 8, 167–185. doi:10.1177/0265407591082002

28.

John

O. P.

Gross

J. J.

(2007). Individual differences in emotion regulation. In Gross

J. J.

(Ed.), Handbook of emotion regulation (pp. 351–372). New York, NY: Guilford Press.

29.

Lipp

O. V.

Price

S. M.

Tellegen

C. L.

(2009). Emotional faces in neutral crowds: Detecting displays of anger, happiness, and sadness on schematic and photographic images of faces. Motivation and Emotion, 33, 249–260. doi:10.1007/s11031-009-9136-2

30.

Matsumoto

LeRoux

Wilson-Cohn

Raroque

Kooken

Ekman

… Coh

(2000). A new test to measure emotion recognition ability: Matsumoto and Ekman’s Japanese and Caucasian Brief Affect Recognition Test (JACBART). Journal of Nonverbal Behavior, 24, 179–209. doi:10.1023/A:1006668120583

31.

Mayer

J. D.

Salovey

Gomberg-Kaufman

Blainey

(1991). A broader conception of mood experience. Journal of Personality and Social Psychology, 60, 100–111. doi:10.1037/0022-3514.60.1.100

32.

Mehta

Clark

M. S.

(1994). Toward understanding emotions in intimate relationships. In Weber

A. L.

Harvey

J. H.

(Eds.), Perspectives on close relationships (pp. 88–109). Needham Heights, MA: Allyn & Bacon.

33.

Miller

Hanford

R. B.

Fassbender

Duke

Schweitzer

J. B .

(2011). Affect recognition in adults with attention-deficit/hyperactivity disorder. Journal of Attention Disorders, 15, 452–460. doi:10.1177/1087054710368636

34.

Mufson

Nowicki

(1991). Factors affecting the accuracy of facial affect recognition. Journal of Social Psychology, 131, 815–822. doi:10.1080/00224545.1991.9924668

35.

Nigg

J. T.

John

O. P.

Blaskey

Huang-Pollock

Willcutt

E. G.

Hinshaw

S. P.

Pennington

(2002). Big five dimensions and ADHD symptoms: Links between personality traits and clinical symptoms. Journal of Personality and Social Psychology, 83, 451–469. doi:10.1037/0022-3514.83.2.451

36.

Raudenbush

S. W.

Byrk

A. S.

(2002). Hierarchical linear models: Applications and data analyses methods. Thousand Oaks, CA: Sage Publications Inc.

37.

Riggio

Friedman

H. S.

(1982). The interrelationships of self-monitoring factors, personality traits, and nonverbal social skills. Journal of Nonverbal Behavior, 7, 33–45. doi:10.1007/BF01001776

38.

Rosenthal

Rosnow

R. L.

(1984). Essentials of behavioral research: Methods and data analysis. New York, NY: McGraw Hill.

39.

Sabatelli

R. M.

Dreyer

Buck

(1979). Cognitive style and sending and receiving of facial cues. Perceptual and Motor Skills, 49, 203–212. doi:10.2466/pms.1979.49.1.203

40.

Salovey

Mayer

J. D.

Goldman

Turvey

Palfai

(1995). Emotional attention, clarity, and repair: Exploring emotional intelligence using the trait meta-mood scale. In Pennebaker

J. W.

(Ed.), Emotion, disclosure, and health (pp. 125–154). Washington, DC: American Psychological Association.

41.

Smalley

S. L.

Winston

(2010). Fully present: The science, art, and practice of mindfulness. Cambridge, MA: Da Capo Lifelong Books.

42.

Sternglanz

R. W.

DePaulo

B. M.

(2004). Reading nonverbal cues to emotions: The advantages and liabilities of relationship closeness. Journal of Nonverbal Behavior, 28, 245–266. doi:10.1007/s10919-004-4158-7

43.

Taylor

G. J.

Bagby

M. R.

Parker

J. D. A.

Grotstein

(1999). Disorders of affect regulation: Alexithymia in medical and psychiatric illness. Cambridge, UK: Cambridge University Press.

44.

Taylor

S. E.

Klien

L. C.

Lewis

B. P.

Gruenewald

T. L.

Gurung

R. A. R.

Updegraff

J. A.

(2000). Biobehavioral responses to stress in females: Tend-and-befriend, not fight-or-flight. Psychological Review, 107, 411–429. doi:10.1037//0033-295X.107.3.411

45.

Thibault

Bourgeois

Hess

(2006). The effect of group-identification on emotion recognition: The case of cats and basketball players. Journal of Experimental Social Psychology, 42, 676–683. doi:10.1016/j.jesp.2005.10.006

46.

Timmers

Fischer

A. M.

Manstead

A. S. R.

(1998). Gender differences in motives for regulating emotions. Personality and Social Psychology Bulletin, 24, 974–985. doi:10.1177/0146167298249005

47.

Wagner

H. L.

(1993). On measuring performance in category judgment studies of nonverbal behavior. Journal of Nonverbal Behavior, 17, 3–28. doi:10.1007/BF00987006

48.

Wright

P. H.

(1988). Interpreting research on gender differences in friendship: A case for moderation and a plea for caution. Journal of Social and Personal Relationships, 5, 367–373. doi:10.1177/0265407588053006

49.

Zhang

Parmley

(2011). What your best friend sees that I don’t see: Comparing close friend dyads and casual acquaintance dyads on the perception of emotional facial expressions of varying intensities. Personality and Social Psychology Bulletin, 37, 28–39. doi:10.1177/0146167210388194

50.

Zuckerman

Larrance

D. T.

Hall

J. A.

DeFrank

R. S.

Rosenthal

(1979). Posed and spontaneous communication of emotion via facial and vocal cues. Journal of Personality, 47, 712–733. doi:10.1111/j.1467-6494.1979.tb00217.x