The concept of empathy has undergone a period of continuous construction with the emphasis on its single dimensions, such as affective empathy or cognitive empathy, to both dimensions ^[1][10]. Davis suggested that empathy is a kind of personality or a stable trait of feeling and understanding others’ emotions, including affective and cognitive factors. Affective empathy refers to emotional sharing in terms of other people’s situations, while cognitive empathy means emotional recognition and understanding ^[2][11]. However, some scholars believe that empathetic behavior, such as prosocial behavior owing to affective or cognitive empathy, is also a component of empathy ^[3][12]. Other researchers also point out that empathy can be divided into state empathy and trait empathy ^[4][5][6,13]. The former is caused by a certain situation, while the latter is a relatively stable individual difference. Regardless of the above diverse definitions of empathy, the two-dimensional component view of empathy, which consists of affective and cognitive sections, is widely accepted by most scholars ^[6][3].

Empathy measurement methods are diverse based on different definitions of empathy, including subjective and objective ones. Subjective measurements include questionnaire surveys and various behavioral evaluation methods. Then, the former can be divided into three categories. The first category is the affective empathy scale, including the emotional response scale (questionnaire measure of emotion empathy, QMEE) ^[7][14] and the empathy concern scale (ECS) ^[8][15]. The second category is the scale that measures cognitive empathy, including the Hogan empathy scale (HES) ^[9][16]. The third category is the scale that measures emotional and cognitive empathy, including interpersonal relation index (IRI) ^[10][17] and the basic empathy scale (BES) ^[11][18]. Objective measures of empathy include neuroimaging, electroencephalography, facial electromyographic activity, heart rate, and skin electrical response, which are mainly used on the study of state empathy ^[12][19]. In this study, empathy can be divided into affective and cognitive empathy from two dimensions. Empathy obtained from the scales that measure both emotional and cognitive empathy is called multidimensional empathy.

Attachment is the emotional connection formed between individuals and significant others in the early infant–caregiver interaction ^[13][20]. According to attachment characteristics, attachment is divided into different patterns, including secure attachment, anxious attachment (or contradictory attachment), avoidant attachment, and disordered attachment ^[14][21]. Early attachment studies mainly focus on the parent–child attachment. However, with the whole-life attachment theory and the multiple attachment theory arising, increasing attention has been paid to the attachment relationship between individuals and other significant figures, which is considered an essential part of attachment ^[15][22]. For the above reason, we classified attachment into parent–child attachment, peer attachment, and adult attachment, according to one’s past or current experiences with significant others, i.e., parents, brothers, sisters, friends, and partners ^[16][23].

In response to differences in attachment types and figures, various measurement methods have also been developed, including behavioral observation, story material, and questionnaire measures. Specifically, the behavioral observation method includes attachment Q-sort (AQS) and strange situation procedure (SSP) ^[17][24]. These two methods aim to measure children’s attachment sorts in real or experimental scenarios. Another type of measurement is performed by assessing children’s internal working models (IWMs) to determine their attachment styles, such as eliciting attachment-related dilemmas by telling stories and then asking children to solve those dilemmas ^[18][25]. The resolution of children to their dilemmas shows their IWMs, reflecting their attachment styles. The last category is the questionnaire survey. For example, the inventory of parent and peer attachment (IPPA) is wildly used for adolescence ^[19][26]. In addition, there are some other commonly used tools for measuring attachment, such as experience in close relationships (ECR) ^[20][27]. Although the attachment theory states that it has two aspects, the dynamic characteristics of the attachment behavior system and individual differences, namely state and trait attachment ^[21][28], the current study focuses on attachment as a stable trait.

The empathy–attachment relationship exists in three possible patterns. The first pattern is that attachment serves as the antecedent variables of empathy. The attachment theory reveals that infants develop their social emotions based on the attachment ^[12][19]. The infant–caregiver interaction is important for infants to recognize and understand the feelings and emotions of caregivers ^[13][20]. The theory also points out the specific mechanism by which children’s attachment quality affects empathy. Secure attachment enables children to develop positive IWMs, including self-affirmation and trust in others, which enable individuals to pay less attention to their feelings but more attention to others’ feelings ^[12][19]. Insecure IWMs involve the denial of the self and distrust of others, hindering individuals from understanding and feeling the emotions of others ^[22][23][29,30]. In addition, attachment also affects empathy through emotional regulation ^[24][31]. The second pattern is that empathy influences attachment. From childhood to adolescence, individuals develop peer attachment ^[25][32]. The equal status of individuals and their peers allows them to share and understand feelings for each other to meet their needs, thus maintaining and promoting peer relationships. This may further enhance peer secure attachment ^[26][27][33,34]. The third pattern is that there has been an interaction effect across time. According to the cascade model ^[28][35], empathy and attachment may influence each other in different development stages. The main pattern in early childhood is that secure attachment promotes empathy, while insecure attachment hinders empathy. In late childhood, especially in adolescence, empathy plays an important role in developing peer attachment.