As global migration between countries and regions continues to accelerate (OECD, 2023a), societies around the world are becoming increasingly diverse, resulting in frequent professional and informal interactions between individuals and groups that perceive one another as being culturally different. For organizations and corporations, comparative models employing various cultural characteristics have been utilized to facilitate understanding between members of various cultural backgrounds (Murdock, 1967; Hofstede, 1991). However, such models grant minimal consideration to the impact of ingroup/outgroup perception, which research has demonstrated results in individuals experiencing more empathy for individuals in their ingroup and less for individuals perceived as outgroups (Cialdini et al., 1997; Galinsky and Moskowitz, 2000). Acknowledgment of the role of ingroup/outgroup perception has subsequently expanded to other fields, such as healthcare and foreign-language pedagogy, where educators are tasked with preparing learners for interactions with individuals and groups to whom they may perceive as members of outgroups. Consequently, in addition to analyses of cultural differences, nurse/patient interactions, which frequently involve intercultural encounters, may also be understood by analyzing the presence or absence of interlocutor empathy (Reynolds, 2017).

The belief that increased empathy can mitigate attitudes of intolerance and discrimination while positively affecting attitudes of respect and understanding has resulted in studies demonstrating that mediation of subject engagement in cognitive empathy can improve attitudes towards members of perceived outgroups (Batson et al., 1997; Batson et al., 2002) including differing cultural groups (Stephan and Finlay, 1999; Finlay and Stephan, 2000; Vescio et al., 2003). The ability of researchers to both elicit and suppress cognitive and affective empathic responses demonstrates the malleability of subject empathy in cross-cultural situations, indicating the potential pedagogical efficacy of targeting learner empathy for increased awareness and understanding of groups who may be perceived as culturally different (i.e., as outgroups). While evidence suggests that empathy differs between individuals (Davis, 1983) and possibly between sexes (Baron-Cohen, 2020), that empathy can be mediated in subjects indicates the possibility of facilitating interventions to increase empathic ability for improved outcomes in interpersonal interactions. At present, the belief that learner ability to empathize with outgroups can be developed underscores the emergence of empathy-centered pedagogies in the fields of narrative medicine (Kuhnigk et al., 2012; Shankar, 2019; Cambra-Badii et al., 2020) and foreign-language acquisition (Lasa Álvarez, 2017; Chen, 2018), fields in which the outcomes of patient/caregiver interactions and intercultural encounters can be positively impacted by interlocutor ability to engage empathically.

Pedagogical attempts seeking to develop learner empathy towards outgroups necessitate tools capable of accurately measuring learning outcomes. While numerous empathy instruments exist, most focus on measuring subject empathy towards members of their ingroup (Rasoal et al., 2011). Increasingly, there is a need for tools to measure ethnocultural empathy or empathy in intercultural contexts. This is important, as “taking the perspective of a person from a different culture may be more difficult than taking the perspective of someone with the same cultural background” (Rasoal et al., 2011, p. 2). Developed by Wang et al. (2003), the Scale of Ethnocultural Empathy (SEE) represents a measurement tool designed to measure “empathy directed towards members of racial and ethnic groups different from one’s own” (p. 222) and tested in a multicultural learning environment (the United States). The present research attempts to validate the SEE in a novel learning environment (undergraduate students at Japanese universities), to consider the applicability of the SEE in societies that are comparatively less multicultural.

With increasing numbers of foreign residents entering Japan to reside in urban and rural environments, there is a need for an instrument to measure ethnocultural empathy to determine the efficacy of curricula targeting empathy in healthcare education and foreign-language learning. The aim of this study was to adapt the English version of the SEE for use in the Japanese context and to assess the psychometric properties of the SEE in a sample of students attending university in a rural region where the percentage of foreign-born residents is relatively low. Differences in scores on the subscales of the SEE and difference in scores between males and females were also examined.

The results from the initial CFA did not substantiate Wang et al.’s (2003) original four-factor solution in the Japanese sample. Consequently, a series of EFAs and CFAs were undertaken to determine the structure underlying the scores. This process revealed a two-factor solution explaining 42.1% of the variance. The two factors are moderately positively correlated, and both factors displayed a sufficient degree of reliability, with Cronbach’s alpha values in Study 2 substantially higher than those in study 1.

The first factor on the Japanese version of the SEE comprises 6 items relating to emotional experiences, such as anger, pride and support, and communicating with individuals from other ethnic groups (e.g., I get disturbed when other people experience misfortunes dure to their racial or ethnic backgrounds). All six items were from the Empathic Feeling and Expression subscale on the original SEE, and so this factor was named similarly, EFE_J. The second factor consisted of 5 items concerning awareness and understanding of the position of culturally different others (e.g., I am aware of how society treats racial or ethnic groups other than my own.). This factor contained items from both the Empathic Awareness and Empathic Perspective Taking subscales from the original scale, and thus it was named EAPT_J.

Correlation analyses between the SEE_J, EFE_J and EAPT_J and two subscales of a measure of general empathy, the IRI, provide evidence for the convergent validity of the Japanese version of the SEE. Test–retest reliability results suggest that scores on the SEE_J and its two subscales are sufficiently stable.

Respondents scored significantly higher on EFE_J than EAPT_J, with a large effect size. One possibility for this is, that as (Wang et al., 2003) hypothesize, the monocultural experience of the Japanese respondents may have delayed development of empathic perspective taking. Concerning gender differences, females scored higher on the SEE_J overall and on EFE_J, both with small effect sizes. Scores on EAPT_J were similar for both genders, however.

Responses from Study 2 revealed that the majority of respondents have had very little contact experience with individuals from other cultures. Those respondents with foreign friendships reported higher total SEE_J, EFE_J and EAPT_J scores than those without non-Japanese friends, however other cultural contact variables had no significant effect on respondents scores.

The factor structure of the Japanese version of the SEE differs from the original version (Wang et al., 2003), as well as the Swedish (Rasoal et al., 2011), Turkish (Özdikmenli-Demir and Demir, 2014), and Spanish (Albar et al., 2015) versions of the SEE, with the elimination of one subscale, Awareness of Cultural Differences, and the combination of two subscales, Empathic Awareness and Empathic Perspective Taking, into one on the Japanese version. However, the items comprising the two factors of the Japanese version, remain similar to those in Wang et al. (2003). Furthermore, similarities can also be seen in the trends in respondents’ scores. In both studies, respondents scored higher on the Empathic Feeling and Expression subscales than on Empathic Awareness or Perspective Taking subscales, and this was also the case in the Swedish and Turkish samples. In addition, females reported higher scores on the SEE overall and the Empathic Feeling and Expression subscales in both this study and Wang et al. (2003). These similarities suggest that despite their differences, both versions of the SEE are tapping similar sources.

One significant reason for the difference in its structure is the number of items that were eliminated from the Japanese version during the analysis process. The majority of these items were reverse scored items, with all 12 such items on the original version eliminated. The inclusion of reverse scored items on measures, their influence on the psychometric properties of instruments, and respondents’ interpretation of these items has been a topic of interest in the past few years both outside (e.g., Carleton et al., 2011; Rodebaugh et al., 2004) and inside Japan (e.g., Nihei et al., 2018). These items have been found to obscure the factor structure of instruments (e.g., Brown, 2003; Nihei et al., 2018; Rodebaugh et al., 2011), to confuse respondents, especially those with less education (Rodebaugh et al., 2011; Weeks et al., 2005), to exhibit weaker correlations with convergent measures than non-reversed items (Nihei et al., 2018; Rodebaugh et al., 2007; Weeks et al., 2005), and even to correlate less strongly with straightforwardly worded items on the same scale (Weeks et al., 2005). Issues with reverse-scored items had a significant impact on the structure of the Japanese version, where many of the reverse scored items correlated only very weakly with the other items on the SEE, failed to load on either factor, or loaded only very weakly.

The absence of these items impacted the structure of the Japanese version of the SEE in two ways. First, the ACD subscale from the original version is composed entirely of reverse scored items, all of which were eliminated from the analysis in this study, leaving only items from three factors in the original version to be analyzed. Second, four of the seven items on the EPT subscale from the original version were reverse scored items, which were also eliminated from the analysis. As a result, items from two subscales on the original scale, EA and EPT, form one factor in the Japanese version. Similarly, Items 4, 6, 20 and 25 (four of the five items on this subscale in the Japanese version) all loaded on a single factor on the Turkish version of the SEE (Özdikmenli-Demir and Demir, 2014) and Items 6, 20 and 25 loaded on a single factor (with Item 4 removed due to low item correlation) on the Swedish version of the SEE (Rasoal et al., 2011). One possible explanation for the tendency of these two aspects to combine in some versions of the SEE while being separate in the original study is the complex nature of the intellectual aspect of ethnocultural empathy which may include both awareness and perspective taking (Wang et al., 2003). Another interesting possibility is that the aspects of intellectual empathy formed two subscales in Wang et al.’s (2003) study due to the influence of a method effect (Brown, 2003) caused by the inclusion of reverse scored items on the EPT subscale rather than the independence of these two aspects. Determining if either these possibilities is correct requires further investigation in both the original and translated versions of the SEE.

Furthermore, the question of cultural plurality—or the lack thereof—bears consideration when comparing results between studies investigating the SEE. As mentioned above, the other contexts to which the SEE has been adapted, America, Sweden, Spain and even Turkey, have higher levels of cultural and ethnic diversity than Japan. Keeping this in mind, it is necessary to consider the potential environmental effects on subjects completing the SEE, as interactions with representatives of non-Japanese cultural backgrounds are rare.

Regarding the relative scarcity of foreigners, in presupposing degrees of intercultural interaction, some SEE items may present challenges for Japanese respondents. Although subjects in the first study were not queried as to their degree of intercultural interaction, considering the small percentage of foreign-born residents in the area in which this study was conducted (Kumamoto City; 0.98%), some subjects may have experienced difficulties in responding to some items. For example, Item 4 presupposes that the subject has experiences as an ethnic or cultural minority, a rare occurrence for typical individuals in Japanese society, while Item 11 assumes that subjects have friends from differing ethnic backgrounds who they have had opportunities to support vocally. Similarly, Items 13 and 21 assume that the subject has had opportunities to experience and praise cultural diversity or advocate against discrimination, situations that may be implausible for some Japanese subjects. Lacking personal experiences, some respondents may have engaged in thought exercises, where they imagined what they would do if they were in situations presupposed by the statements, or they may have responded based on idealized actions in theoretical situations. The fact that Japanese respondents’ scores on these particular items tended to be higher than those of Wang et al.’s (2003) American respondents is suggestive of this phenomenon, as overestimating one’s ability to react in a hypothetical situation has been noted in past research (e.g., Lee, 1984).

This gives rise to one of the limitations of this study. All of the data was collected using a self-report measure. This allowed for the collection of the large data set necessary for the validation of the Japanese version of the SEE, but as Fan et al. (2006) have reported, it is not the case that self-reported responses necessarily imply similar behaviors. That is to say, respondents who report that they would act with ethnocultural empathy, may not behave that way in actual situations of cross-cultural encounter. In addition, the attitudes of the sample population examined in this study, university students, may not be representative of those of the Japanese population as a whole. Furthermore, these students were enrolled in universities in a smaller, non-multicultural city. Future research should examine the structure of the SEE in a more heterogeneous sample, including students from a more metropolitan area, such as Tokyo or Osaka, both of which have much higher percentages of foreign residents, as well as individuals of differing ages and occupations. A further objective for future research is the development of additional items for the Japanese version of the SEE so that the instrument can serve as a better measure of ethnocultural empathy in the Japanese context.

The purpose of this study was to adapt the English version of the SEE for use in the Japanese context and to determine the validity and reliability of the adapted version to allow for the measurement of cultural empathy and promote further research into this concept in Japan. In addition to its use in research, the Japanese version of the SEE can be very useful in practice. First of all, as Wang et al. (2003) point out, the use of empathy can be effective in altering individuals’ attitudes towards cultural outgroups, and the SEE could serve as a tool for measuring these changes. Secondly, as Albar et al. (2015) indicate, the ability to measure ethnocultural empathy allows for the design of courses and programs aimed at increasing individuals’ sense of ethnocultural empathy, and the measurement of their effectiveness. Such programs could be of importance in Japan with its relatively homogenous culture, and in particular in less urbanized areas of the country such as that where this study took place.