Age-related changes in empathy-related responding

Empathy is commonly defined as a response that stems from the apprehension or comprehension of another’s emotional state or condition and is identical or very similar to what another is feeling or would be expected to feel. An example of empathy is if a girl views a sad boy and experiences sadness herself, even though she is aware that the boy’s emotion is not her own emotion. Unlike emotional contagion, empathizers must realize that another person is experiencing emotion (or would be expected to do so), not just themselves. Empathy is generally believed to be based on cognitive processes such as identifying another’s emotion, cognitively taking the role of the other (i.e., perspective taking), or assessing information in memory that is relevant for understanding another person’s emotion or situation. Although many investigators also require the affective component to label a response as empathy, others label the cognitive processes contributing to empathy as empathic accuracy or cognitive empathy. Empathy, when defined as having an affective as well as cognitive component (i.e., as defined previously), is believed to frequently lead to sympathy or personal distress. Sympathy is an affective response that can stem from empathy, perspective taking, or other cognitive processing including retrieval of information from memory. It consists of feelings of sorrow or concern for another (rather than feeling the same emotion as another). Thus, if an observer feels concern for someone who is sad or distressed, or for someone who is in a situation likely to evoke sadness or distress, the observer is experiencing sympathy. Like sympathy, personal distress also frequently stems from exposure to another’s state or condition but is a self-focused, aversive emotional reaction to the vicarious experience of another’s emotion (e.g., discomfort, anxiety; see Eisenberg and Fabes, 1998). For example, if a girl feels distress when observing someone who is sad because it makes her uncomfortable, she is viewed as experiencing personal distress. In general, sympathy has been positively related to prosocial behaviors performed for selfless reasons, whereas personal distress tends to be negatively or unrelated to such prosocial behavior (Eisenberg and Fabes, 1998; Knafo et al., 2008). There has been considerable interest in understanding the emergence of empathy and its related responding. Based on both theory and empirical findings, it is believed that empathy-related responding is exhibited early in life and continues to improve with age, especially in the toddler, preschool, and elementary years. Hoffman (2000) proposed a theoretical model including a series of phases, delineating the development of empathy-related responding and prosocial behavior. Specifically, rudimentary forms of empathy are evidenced by the newborn’s reactive or contagious crying in response to the cries of other infants, although researchers have questioned whether these findings support the notion of rudimentary empathy or if infants may simply find a novel cry to be particularly aversive (Eisenberg and Lennon, 1983). Although this issue has not been resolved, it is clear that infants are responsive to others’ emotional signals. According to Hoffman’s theory, beginning around the end of the first year of life, infants typically seek comfort for themselves when exposed to another’s distress. Because infants at this age cannot fully differentiate between their own distress and that of another, they are likely to respond with personal distress. In the second year of life, toddlers can begin to express concern (sympathy) for another, rather than simply seeking comfort for themselves. However, toddlers’ prosocial behaviors may involve giving the other person what they themselves find comforting (e.g., bringing a favorite teddy bear to a distressed adult) because they have difficulty differentiating factors that affect their own versus others’ emotions. Hoffman argued that as children develop cognitively, they are increasingly capable of understanding another person’s needs and that others’ needs may differ from their own. In early childhood, however, these empathic responses are limited to another’s immediate distress. With increasing perspective-taking skills and cognitive understanding of others’ emotional states, older children begin to experience empathy towards people who are not physically present (e.g., if they hear about someone in distress), and by later childhood (around 9 or 10 years of age), the ability to experience empathy for another’s life condition or general plight develops (Hoffman, 2000). Thus, the adolescent is viewed as capable of comprehending and responding to the plight of an entire class of people, such as the impoverished. In brief, Hoffman argued that with increasing cognitive maturation, children are better able to respond with concern to others’ distress, be it observed or presented symbolically (e.g., through the written word), and are better able to empathize and sympathize with a broad range of people. Others (e.g., Eisenberg and Fabes, 1998) have argued that the ability to regulate vicariously induced affect is important for children to experience sympathy rather than personal distress. This is because high levels of empathically induced negative emotion are likely to be experienced as aversive, which results in children focusing on their own aversive emotional state rather than on the emotions and needs of the other person. Because self-regulation develops rapidly in the first four years of life and continues to emerge more gradually across childhood and adolescence, one would expect sympathy to increase with age. Empirical research provides some support for Hoffman’s ideas. At six months of age, children are rarely upset by others’ distress. At about a year of age, children may attend to others’ emotional displays but react with emotion relatively infrequently. By about 12-18 months of age, toddlers sometimes respond to another’s distress with empathy and prosocial reactions, although such responses are not very frequent and are directed mostly at people they know well (e.g., their mother). Moreover, children of this age often respond with aggression, ignoring, and distress reactions (Eisenberg et al., 1998; Zahn-Waxler et al., 2001). The frequency of empathy and sympathy has been found to increase with age in the first years of life (e.g., Knafo et al., 2008). For example, using longitudinal data and observed measures of empathy and empathy-related responding, Zahn-Waxler and colleagues (2001) studied toddlers’ reactions to an unfamiliar female and to their mother feigning injuries at 14, 20, 24, and 36 months of age. An increase in toddlers’ concern was found, and self-focused distress decreased with age, particularly from 14 to 24 months. This is consistent with the finding in a meta-analysis that there was a significant increase in prosocial behavior within the infant and preschool periods (effect sizes = 0.24 and 0.33 for infant and preschool, respectively; Eisenberg and Fabes, 1998; see Hay and Cook, 2007, for the argument that prosocial behavior does not increase systematically in the first two years of life). In a number of the studies in the meta-analysis, empathy/ sympathy was part of the measure of prosocial behavior. The authors qualified their meta-analytic findings, however, by noting that they were typically based on relatively small samples and cross-sectional data. Consistent with Hoffman’s arguments about the importance of differentiating one’s own and others’ emotions for young children’s empathy, at about 1.5-2 years of age, children who are able to recognize themselves in a mirror-a behavior considered indicative of the development of a sense of self that is separate from that of others-are more likely to exhibit empathy and prosocial behavior than are their peers without such self-recognition. Moreover, in the second year of life and the preschool years, children who respond with more cognitive forms of empathy (e.g., hypothesis testing, inquisitiveness whereby the child tries to understand others’ distress) appear especially likely to assist others due to empathy (Knafo et al., 2008; see Eisenberg and Fabes, 1998). Despite some increases in empathy and prosocial behavior, empathy during early childhood appears to be relatively stable across time and context. For example, Knafo et al. (2008) found rank-order (intraindividual) stability in empathy across the second and third years of life. In addition, the researchers showed evidence that children’s empathy toward the feigned distress of the mother and an unfamiliar female experimenter were positively related. Moreover, the indices (empathy and hypothesis testing, across the two victims) loaded on a single factor at each age, supporting the existence of an overall empathic disposition in young children. Findings on age-related changes in empathy-related responding past the early years are somewhat less consistent than findings for the first few years of life, likely due in part to the measures used to assess empathy and sympathy. Most of the measures of empathy/sympathy used in elementary school involve selfreports on questionnaires or in response to empathy-inducing stories. In studies of empathy-related responding before approximately 1985, there was evidence that self-reports of empathy increase in the school years until age 11; however, findings were not very consistent after that age (see Lennon and Eisenberg, 1987). Facial/gestural indices appeared to be either inversely related or unrelated to age in the early school years, perhaps due to increases with age in the ability to mask emotion. In many early studies of empathy, children were asked their feelings in response to a series of very short vignettes about other children in emotionevoking contexts. Studies of this sort have been criticized for not being evocative and for assessing the desire to behave in socially appropriate ways rather than actual empathic responding (Eisenberg and Lennon, 1983). Therefore, it is wise to be cautious when drawing conclusions from early studies using picture-story methods. The pattern of findings in more recent childhood studies using better measures is somewhat clearer. Eisenberg and Fabes (1998) reported a meta-analysis of age differences in empathy/sympathy (rather than prosocial behavior) in studies published from 1983 to about 1995 and found an overall unweighted effect size of 0.24 (favoring older children). Effect sizes for age-related increase in empathy varied with the method of assessing empathy-related responding; they were larger for observational (usually behavioral) and self-report indices than for nonverbal (facial/physiological) or other-report measures (for which the effects sizes were not significant). Studies of empathy-related responding across adolescence are limited. In studies before about 1985, findings were inconsistent (Lennon and Eisenberg, 1987). However, most early studies did not assess sympathy, which one might expect to increase along with adolescent gains in emotional understanding. In studies since the 1980s, researchers have found evidence for an increase in selfreported empathy-related responding, especially sympathy, across the junior and high school years (see Eisenberg, Morris, McDaniel, and Spinrad, 2009). Unfortunately, there are few studies of age-changes in adolescence using other types of measures in adolescence. Thus, we cannot be sure that sympathy actually increases rather than adolescents’ self-perceptions of being sympathetic. Studies of change in empathy-related responding across the adult lifespan are more limited than studies with children. The majority of the research has been cross-sectional and has found either lower empathy in older adults than in younger persons or equivalence across ages. For instance, of two narrative studies, one showed age equivalence and the other found higher empathy in adolescents but no difference between midlife and older adults (see review in Grühn et al., 2008). Four questionnaire studies showed lower empathy in older adults, but two showed age equivalence (see review in Richter and Kunzmann, 2010).