Abstract
The gross generalization that East is collectivistic and West is individualistic overlooks the within-group variability among East Asians in the current era of social change and globalization. The aim of this study was to disentangle the role of sociodemographic factors, ethnic heritage culture, and immigration in shaping the individualistic–collectivistic value orientations of Korean mothers and their fifth-grade children residing in rural Korea, urban Korea, and Koreatown, Los Angeles. A European American sample in Los Angeles provided an approximation of mainstream United States values. In semi-structured interviews, participants’ value orientations were assessed using interpersonal dilemma scenarios in home and school situations. Participants’ household sociodemographic factors were the most significant contributor to their home values: Higher maternal education was associated with orientation towards individualism, and three-generation households was associated with orienttion towards collectivism. When the sociodemographic factors were taken into consideration, Koreans were not more collectivistic than European Americans in the home domain. Domain-specificity was found such that individualism was greater in the school domain than in the home domain across the three groups of Koreans, who were as individualistic as European Americans in the school domain. Generational difference was found such that children were more collectivistic than mothers in the home domain. Rural Koreans were not more collectivistic than urban Koreans, which should be interpreted in the context of the high level of technology and education in the rural Korean environment.
Keywords
Research has repeatedly shown that people of East Asian heritage are collectivistic, placing primacy on family and group orientations, whereas those of Western Anglo background are individualistic, prioritizing personal goals and desires. This pattern of results has been evident when comparing East Asian and Western nations, as well as comparing Asian Americans and European Americans (e.g., Chao & Tseng, 2002; Markus & Kitayama, 1991). East Asian values are thought to result in collectivistic socialization of children by their parents (e.g., U. Kim & Choi, 1994 [Korea]; Rothbaum, Pott, Azuma, Miyake, & Weisz, 2000 [Japan]).
Culture, however, is a dynamic system. Cultural values are not static; instead they respond to external forces (Markus & Kitayama, 2010; Twenge & Kasser, 2013). Given the global trend of increased urbanization, education, and nuclear-family structure, sociodemographic factors should be examined as a crucial force that shapes cultural values. Hofstede (1980, 2001) and Triandis (1993) noted that sociodemographics influence individualism and collectivism at the societal and individual levels. More recently, Greenfield (2009) theorized that the types of ecologies in which people live strongly influence the relative prioritization of individualism or collectivism. At one extreme, collectivistic values are adaptive in small communities with rural, farming ecology, and extended family structure, where families work together to fulfill their needs. Such communities tend to have the characteristics of informal home education, simple technology, ethnic homogeneity, and less wealth. Each of these environmental features reinforces collectivism. At the other extreme, individualistic values are advantageous in a large, urban ecology where individuals have greater contact with out-groups and compete in educational and economic realms. This type of ecology commonly features nuclear-family residence, complex technology, ethnic diversity, and greater wealth. Each of these ecological factors pushes people’s values in the individualistic direction (Greenfield, 2009). Indeed, in the United States and the United Kingdom, increasing urbanization has been associated with increasing individualism for two centuries (Greenfield, 2013).
If sociodemographic factors engender different sets of values, a critical question arises: To what extent is the widely endorsed notion that East is collectivistic and West is individualistic attributable to ethnic heritage culture or to variations in sociodemographic contexts? One must consider the possibility that national differences in values may have been exaggerated because nations and sociodemographics are often confounded (Park, Coello, & Lau, 2014). Ethnicity-based contrasts fail to consider variations among people of the same ethnic group living in different ecologies. The current study offers a way to examine these issues by comparing Korean families (mothers and their fifth-grade children) who live in rural Korea, urban Korea, and an urban ethnic enclave in Los Angeles, United States. Based on the notion that features of rural communities (e.g., farming, extended family structure, less formal schooling) foster collectivism among family members (Greenfield, 2009), we hypothesized that the rural Korean sample would have more collectivistic values than the urban Korean sample in the home domain (Hypothesis 1).
Socioeconomic Contexts of Korea and the United States in 1994 and 2010
South Korea (hereafter Korea) has undergone economically driven social change that has transformed the country’s ecology and lifestyles in the past few decades. In 1994—the average year of immigration for the Korean American sample in this study—the home country’s GDP per capita was $10,275 (United States $ throughout this article), 77% of the population lived in urban regions, and the post-secondary education rate was 45%. During the same year, the United States GDP per capita was $27,776, the urban population rate was 77%, and the post-secondary education rate was 78% (World Bank, 2014).
In 2010, when we completed our Korean data collection (2009-2010), Korea’s GDP per capita had risen to $22,151, 83% of the population lived in urban areas, and 100% of the population reported continuing after secondary school. In the same year, the United States reported its GDP per capita of $48,358, an urban population rate of 82%, and a post-secondary education rate of 88% (World Bank, 2014). The statistics demonstrate growth in economy and education in both Korea and the United States, with the two nations’ ecologies becoming more comparable in 2010 than in 1994. By 2010, post-secondary and urban population rates in Korea surpassed rates in the United States, although the United States GDP value remained higher than that of Korea.
Our cross-cultural comparison included the archival data of European American mothers and their fifth-grade children in Los Angeles from previous studies that utilized the same data collection method (Greenfield & Quiroz, 2013; Raeff, Greenfield, & Quiroz, 2000). The European American data served as a reference point for the Korean American sample, because the data collection year—1994—coincided with the average year when our Korean American participants immigrated to the same region in the United States, settling in an ethnic enclave within Los Angeles. Furthermore, the European American sample served as a comparison group for the Korean sample in urban Korea; this comparison was intended to challenge the notion that East is forever collectivistic and West is forever individualistic. Social change in Korea and the increased ecological similarities between urban Korea and urban United States led to our prediction that the following cultural essentialist hypothesis will not be supported: Urban Koreans in Korea will be more collectivistic than urban European Americans in the United States in the home domain (Hypothesis 2).
Heritage Culture and Individual Differences in Sociodemographic Characteristics
Korea is known for its strong Confucian heritage since the 900s. U. Kim and Choi (1994, 2014) note that Confucians do not consider individuals as independent entities but see them as linked in a web of interrelatedness. In Confucianism, the family is considered the model for all other relationships. This family-centered collectivistic value system was adapted to the agrarian environment of its time where formal education was not widely available and three-generation households were beneficial for farming and other subsistence activities (Kendall, 1996; Park & Cho, 1995). This heritage value system in Korea is very different from the heritage value system in the United States of rugged individualism. Nonetheless, as Korea has moved from subsistence farming toward a commerce-based economy with high levels of formal education and a majority of nuclear-family households (Cho & Shin, 1996), we theorize that collectivistic values have weakened and individualistic values have become stronger.
Yet, in conditions of rapid social change, changes are uneven within the population. Although an entire community can be transformed from rural to urban ecologies, individual families within a community can also encounter social mobility (e.g., higher education). Therefore, household sociodemographic factors must be considered when examining ethnic differences in cultural values. We focused on household differences in three-generation versus nuclear-family household status, maternal education level, and paternal occupations. Three-generation family households manifest greater family unity and interdependence, the most basic form of collectivism. Greater maternal education level and higher-status paternal occupation are sociodemographic features that engender individualism (Greenfield, 2009).
Our theoretical view is that heritage values remain strong, as long as the sociodemographics that have made them adaptive are intact. Thus, we expected the Korean heritage values of collectivism to distinguish our Korean samples in three regions from our European American sample in the United States, but only insofar as these values were supported by favorable household sociodemographic conditions (three-generation households, lower maternal education, and lower-status paternal occupations; Hypothesis 3). The other side of the coin was that individualistic values would be fostered by nuclear-family households, higher maternal education, and higher-status paternal occupations.
Immigration and Social Change
The global direction of immigration is generally from poorer to wealthier environments, from rural to urban environments, and from environments with less opportunity for formal education to environments with more opportunity for formal education (Suárez-Orozco & Sattin, 2007; Suárez-Orozco & Suárez-Orozco, 2009). Given that individualistic values are adaptive and dominant in the new environments, this pattern of immigration likely produces an environmental press on immigrants in the individualistic direction. For instance, Raeff and colleagues (2000) found that European American and Latino teachers and European American parents had more individualistic values than Latino immigrant parents of fifth graders. The findings suggest that immigrant families in the United States receive the individualistic press outside of their home setting.
Yet, ethnic enclaves can be an extension of immigrant families’ home environments where collectivism is valued. In addition, linguistically and culturally, ethnic enclaves are relatively isolated environments in which collectivism is adaptive (Greenfield, 2009). Thus, residence in ethnic enclaves may prompt cultural freezing of collectivism among immigrants, while their homelands experience cultural shift toward individualism as the environments are becoming wealthier, more educated, and more diverse after their departure.
Patterns of Korean Immigrant Adaptation in the United States
The dominant acculturation pattern for Korean immigrants has been described as “adhesive” adaptation (Hurh & Kim, 1984). That is, “certain aspects of American culture and social relations are added on to Korean immigrants’ traditional culture and social networks” (Hurh & Kim, 1984, p. 208). For example, Hurh and Kim found that the number of European American friends increased with length of residence, but family values—most relevant to the current study—did not decline: More than 90% of the respondents in the study agreed that family duty should always be given priority, and the numbers did not vary with length of residence. In addition, the sample was linguistically isolated (low proficiency and use of English, high proficiency and use of Korean), a sociodemographic characteristic associated with collectivistic values (Greenfield, 2009). However, the data were collected around 1980 in Los Angeles. Although a study done a few years later in Chicago came to the same conclusion (K. C. Kim & Hurh, 1993), more recent evidence is needed. Prior studies led to our fourth hypothesis: The values of our Korean American participants will be characterized by “freezing” at the time they left Korea, therefore, showing more collectivistic home values than the European American participants, whose values represent the dominant United States society at the time of immigration for our Korean American participants.
Home Versus School Values
Individualistic values are tied to school achievement (Greenfield, 2009). Students are typically evaluated based on their performance on tasks purported to assess individuals’ knowledge. Despite some collaborative work in classrooms, collectivistic behavior such as working together on a test would be considered cheating (Whiting & Whiting, 1973/1994). Thus, our fifth hypothesis was that values would be more individualistic in the school domain than in the home domain.
Development of Cultural Values
Within the urban environment, individualism may heighten between childhood and adulthood as the demands for competition and achievement increase (Greenfield & Quiroz, 2013). Even in individualistic cultures, cooperative behavior is thought to emerge before competitive behavior (Madsen, 1971). Using a subset of scenarios as the present study (see “Interpersonal Dilemma Scenarios” in the “Method” section), Greenfield and Quiroz (2013) found that the present sample of European American fifth-grade children had more collectivistic home values than their college-educated mothers. Interestingly, the home values did not differ between fifth-grade Latino American children and their immigrant mothers with five years of schooling; Latino mothers and children were equally collectivistic, and more collectivistic than their European American counterparts. These findings support the notion that education is an individualizer, but the confounding of maternal education level with ethnicity limits the interpretation of the results. It is unclear whether the lack of generational difference in the Latino families was due to the lower maternal education levels or their ethnic or immigrant background. In the current study, Korean mothers were expected to have education levels comparable with that of European American mothers. We hypothesized that Korean children would be more collectivistic than their mothers in the home domain, showing the same generational difference as the European American group (Hypothesis 6).
The Current Study and Hypotheses
The aim of this study was to disentangle the role of sociodemographic factors, ethnic heritage culture, and immigration in shaping the individualistic–collectivistic value orientations of European American families and native and immigrant Korean families. Utilizing vignettes about interpersonal dilemmas in home and school situations (explained and exemplified in the “Method” section and the appendix), we examined fifth-grade children’s and their mothers’ inclination toward collectivistic versus individualistic resolutions of cross-cultural value conflicts. To consider the effect of sociodemographic factors, we recruited participants sharing Korean heritage cultural background but residing in three distinctive regions: rural Korea, urban Korea, and urban ethnic enclave in Los Angeles, United States. We also assessed household sociodemographic factors of maternal education, paternal occupation, and family structure. To examine the role of heritage culture, we utilized European American data to make a comparison with the Korean samples. Because the European American data were collected at the approximate time when the Korean American families immigrated from Korea to the United States, the European American sample provided information about the dominant United States culture at the time of immigration for the Korean American sample. That is, the European American group provided benchmark values of the dominant host culture environment available for assimilation at the time of immigration. Finally, generational differences in cultural values between mothers and children were examined. Value orientations were examined separately for the home and the school domains because schools may particularly emphasize individualistic socialization (Greenfield & Quiroz, 2013). We had six hypotheses:
Their home values will represent cultural “freezing” in the home sphere in their process of “adhesive” or context-specific acculturation. At the time they immigrated, Korea was much poorer, less technological, less educated, and less urbanized than the United States, factors that would lead to a more collectivistic value system than the European American sample tested at that same time period (Greenfield, 2009, 2013; Zeng & Greenfield, 2015). The notion of cultural “freezing” in this case comes from the fact that, since the 1990s, Korea should theoretically have moved in the individualistic direction influenced by increased wealth, technology, education, and urbanization. Cultural freezing may also be reinforced by the cultural and linguistic self-containment of immigrants in an ethnic enclave, given that self-contained small communities also develop collectivistic value systems (Greenfield, 2009).
Method
Participants
Participants were 115 fifth-grade children (34 rural Korean, 30 urban Korean, 30 Korean American, 20 European American) and 70 of their mothers (13 rural Korean, 21 urban Korean, 20 Korean American, 16 European American). Mothers were less likely than their children to participate in the study, especially at the rural site where distance from school and work commitments often prevented parental involvement. None of the children were related. Korean participants were tested in 2009-2010. European American participants were tested in 1994.
Rural Korean participants were recruited from two neighboring elementary schools in Yang-Pyeong, a county known for farming and gardening. The urban Korean sample came from two local elementary schools in Incheon Metropolitan City, located adjacent to the capital of Korea, Seoul.
Korean American participants were recruited from children’s programs at a large Korean church in Los Angeles (in a neighborhood known as Koreatown). Korean churches in the United States function as an epicenter of the immigrant community (Min, 1992), a situation that is reflected in 75% to 80% Christian church membership or affiliation with Christianity among the Korean population in the United States, compared with only 20% to 30% in Korea (Connor, 2014; J. H. Kim & Min, 2002; “Report on Korean Christianity,” 2015). Non-Christian or non-religious Koreans commonly attend Korean churches to utilize social services or participate in activities (Min, 1992, 2010). For example, approximately 40% of Korean immigrants who were non-Christians in Korea were reported to have become Korean church attendees in the United States for various reasons such as meeting other Koreans, acquiring information concerning immigrant adjustment, and teaching their children Korean language and customs (Min, 1992). This figure indicates that there is little religious selection of Christians at the time of immigration; the movement to Christianity often occurs after arrival in the United States. In addition, Korean American participants were recruited from children’s programs at the church that did not preclude non-affiliated or non-Christian families. In sum, these facts indicate that the immigrant sample in our study would not deviate in terms of their original religious origins in Korea from a sample in urban Korea. Therefore, Korean church is a suitable recruitment site to obtain a representative sample of immigrant families situated in ethnic communities.
Yet, Korean immigrants tend to come from urban, middle-class backgrounds, and this was the case in our sample (see next paragraph and Table 1 for figures on urban origins and educational level of the Korean immigrant sample). In addition, Christianity in Korea is overrepresented among urban, middle-class Koreans, and it is also possible that Korean Christians who are more familiar with Western values may be more inclined to immigrate to the United States (Min, 1992). Both urban environment and higher income levels develop individualistic values (Greenfield, 2009; Zheng & Greenfield, 2015), and Western values are often equated with individualism. If urban, middle-class Christians had immigrated disproportionately and this sample bias had influenced our study, we would have found our Korean American sample to be more individualistic than urban Koreans. However, the opposite was the case: Although the difference was not significant, Korean Americans were slightly less individualistic than urban Koreans (see the “Results” section).
Sample Characteristics.
Note. Paternal occupation score was coded using an existing scale (Nam & Boyd, 2004). The scores consider the income level of the occupation and the education required for the occupation (Range = 1 [e.g., dishwasher] to 100 [e.g., surgeon]). Three-generation household represents whether or not at least one grandparent lived in the same household. Group differences are reported at p < .05 level. Least Significant Difference (LSD) post hoc tests were used. Rur Kor = Rural Korean; Urb Kor = Urban Korean; Kor Am = Korean American; Euro Am = European American.
Sample characteristics of the four groups are presented in Table 1. Maternal education level was significantly higher for European Americans than Korean Americans (p = .048, d = 0.43, small effect size), urban Koreans (p < .001, d = 0.90, large effect size), and rural Koreans (p < .001, d = 0.90, large effect size). Korean Americans had higher maternal education level than urban Koreans (p = .035, d = 0.51, medium effect size) and rural Koreans (p = .018, d = 0.56, medium effect size). Maternal education levels for urban and rural Koreans were not significantly different (p = .630, d = 0.10, small effect size). In addition, Korean American mothers had by and large grown up in urban environments (13 in urban, 1 in both urban and rural, 2 in rural, 4 missing information). Hence, we can characterize this community of Korean American participants in Los Angeles as urban in both its transnational roots and current state.
For the European American sample, we utilized a preexisting archived data set (Greenfield & Quiroz, 2013; Raeff et al., 2000) collected in 1994. The year coincided with the average year of immigration for Korean American mothers in our study (Myear = 1994, Range = 1989-2004, SD = 4.70 years), providing a reference context for the mainstream values in the era of immigration. The European American group was part of the original sample for which the scenarios in this study were designed (Greenfield & Quiroz, 2013; Raeff et al., 2000).
All three Korean groups were more likely than the European American group to live in three-generation households (i.e., at least one grandparent living in the household; Korean American: p = .029, d = 0.64, medium effect size; urban Korean: p = .038, d = 0.63, medium effect size; rural Korean: p < .001, d = 0.86, large effect size). Paternal occupation score was higher for European Americans than Korean Americans (p = .003, d = 0.64), urban Koreans (p = .002, d = 0.82, large effect size), and rural Koreans (p < .001, d = 1.61, very large effect size). Rural Koreans had lower paternal occupation score than Korean Americans (p = .025, d = 0.82, large effect size) and urban Koreans (p = .004, d = 0.76, medium effect size).
Procedure
The first and the second authors of this study and another research assistant recruited and conducted the interviews with Korean families in rural Korea, urban Korea, and Los Angeles. These research staff were bilingual and bicultural Korean Americans who had immigrated from Korea to the United States. Using online search and community directory, the first and the second authors identified potential research sites (elementary schools and ethnic churches) and made the initial contact via phone calls and e-mails. Once the school or church staff expressed an interest in the study, the research staff met with the school principal or church pastor to answer questions and obtained permission to recruit participants at the site. During the meeting, the school or church leaders also served as community advisers to ensure the ecological validity of our research procedure and suggest any modifications. The suggestions included providing the option for participants to be interviewed in person at school/church or home, or by phone, depending on participants’ preference.
After the sites were identified and recruitment methods were agreed on, school or church personnel announced the study and invited mothers with fifth-grade children to participate in the study with their child. The announcement was made in parent meetings or class, and notes were also sent home. Research staff were present during announcements in parent meetings or class to answer any questions that parents or children might have and to schedule interviews. In addition, the research staff were provided with a list of fifth-grade students and their home contact information by the school or church staff. The information was used to reach potential participants who were not present in parent meetings or class, as well as those who were undecided about the study participation during the announcement. Research staff called mothers to discuss the study and scheduled the interview. Participants were given a choice to schedule for an in-person (school, church, or home) or a phone interview.
For in-person interviews, participants met individually—except for the rural Korean children who met in groups—with one of the three interviewers. Prior to data collection, the interviewers went through training and mock interviews to ensure procedural consistency among the interviewers. The first author was present at every research site to conduct interviews and supervise the other two interviewers. An interviewer explained that eight scenarios would be presented and obtained the consent to audio-record the interview. Each scenario introduced an interpersonal dilemma concerning collectivistic and individualistic ideas. Participants were told that the study purpose was to understand how people respond to various interpersonal situations. Participants were asked how they would handle each situation and why they gave the response. Interviewers emphasized that there were no right or wrong answers. Prior to the interview, mothers provided written consents for themselves and written permission for their children to participate in the study. Children provided written assents to participate in the study.
The procedure of individual interviews was modified for rural Korean children, whose school personnel raised the concern about children’s possible discomfort with being interviewed alone. Following the school’s recommendation, group interviews were conducted with rural Korean children. Steps were taken to ensure procedural compatibility with the other groups of children who were interviewed individually. First, rural Korean children were given paper copies of the scenarios and asked to prepare individual responses by writing down their responses to each scenario. This allowed them time to formulate individual responses and prevented them from changing responses when they later gathered in a group. After rural Korean children prepared their responses, interviews were conducted in groups of three to four children. At this time, an interviewer orally presented each of the eight scenarios, and rural Korean children in the small group took turns sharing the responses they had prepared. After the interview, their written responses were collected and compared with their interview responses. None of the children changed their individual written responses during the group interview.
Whereas all children were interviewed in person, 40% of all mothers chose to be interviewed by phone (all 13 rural Korean mothers, none of the 21 urban Korean mothers, none of the 20 Korean American mothers, and 15 of the 16 European American mothers). Rural Korean mothers commonly reported unavailability to meet in person (i.e., in-person was not a viable option), and European American mothers reported the phone to be a more convenient interview method for them (i.e., preference). Compared to rural Korean mothers, urban Korean and Korean American mothers tended to be more highly involved in their children’s school lives and were more eager to participate in the study. Rural Korean mothers infrequently visited school due to work commitments or lack of transportation between home and school. All rural Korean students took the school bus, which was the only method of commute for most of them.
By interviewing in the environment where participants felt most comfortable, we gave priority to functional equivalence (familiarity and comfort with the medium) over stimulus matching (all interviews by phone or all interviews face-to-face; all interviews individually or all interviews in a group; Greenfield & Quiroz, 2013; Greenfield & Zukow, 1978). After the interview, mothers filled out a sociodemographic questionnaire. Mothers who did in-person interviews completed the form after responding to the scenarios. For mothers who were interviewed by phone or did not participate in the interview, children took the form home to be filled out and returned it to school or church. Both Korean and English versions of the interview and the form were available for Korean American participants. All procedures were approved by the University of California, Los Angeles Institution of Review Board.
Measures
Interpersonal dilemma scenarios
Eight hypothetical scenarios introducing dilemmas between collectivistic and individualistic ideas were used to assess participants’ cultural value orientation. Four scenarios involved family members (e.g., two sisters are fighting over one t-shirt); these are termed home scenarios throughout this article. Four scenarios dealt with school situations (e.g., two students want to work together and submit one poster for a class project that requires an individual submission); these are termed school scenarios throughout. The scenarios had been used to examine the cultural value orientation of fifth-grade children and parents from European American and Latino backgrounds (Greenfield & Quiroz, 2013; Raeff et al., 2000). Hence, this instrument was unique in having been designed to assess individualism and collectivism in this particular age group of children and their mothers.
Using the method of back translation, English scenarios were translated into Korean by three bilinguals, two of whom were native Koreans and were familiar with Korean elementary school education (and they themselves had gone through it). The translators judged all the scenarios to be understandable and applicable to children and mothers in Korea and the United States. There were male and female versions of the scenarios and two scenario presentation orders. To prevent English versus Korean names priming Korean American participants’ responses to the scenarios, names were replaced with initials in the scenarios that were administered to the Korean American sample. Korean names were used in the scenarios given to the urban Korean and rural Korean samples, and English names were used for the European American sample.
For each scenario, we used existing categories capturing the central themes of the responses based on the combination of participants’ decision (i.e., what to do) and justification (i.e., why) toward the dilemma. Each category was classified as individualistic, collectivistic, or mixture of the two (Greenfield & Quiroz, 2013; Raeff et al., 2000). As with the prior studies, a score was assigned to each of the eight scenario (0 = individualistic, 0.5 = mixture, 1 = collectivistic). Scores were added across the four home and four school scenarios to produce cultural value composite scores (Range = 0-4) for the home domain and the school domain, where higher scores represented more collectivistic values. The composite score was equivalent to a 9-point scale as there were nine possible composite scores (0 to 4 in 0.5 increments). Item internal consistency was irrelevant because we were not interested in creating homogeneous scales. Our goal was to find out in how many home situations and school situations participants reported collectivistic or individualistic solutions to the interpersonal dilemmas.
In the first training stage of coding, two bilingual coders utilized some of the existing coded data of the European American sample from previous studies (Greenfield & Quiroz, 2013; Raeff et al., 2000) to familiarize with the coding scheme and to practice coding. Utilizing the original coding system ensured that comparison could be reliably made between the European American and the Korean data in the current study, given that the same coding was used for the archived European American data in this study. Once the two coders went through an extensive practice period using a set of European American data and felt comfortable with coding, each coder coded another set of the prior European American data. Both coders achieved interrater reliability with the original coding (κ = .71-1.00 for the home domain, κ = .70-1.00 for the school domain). The interrater reliability was based on the three possible scoring of the scenarios (0 = individualistic, 0.5 = mixture, 1 = collectivistic). In the subsequent actual coding stage, the two coders independently coded 33% of all Korean data in the current study and achieved interrater reliability with each other (κ = .72-.93 for the home domain, κ = .79-.92 for the school domain); discrepancies were resolved by reaching consensus through discussion. The remainder of the Korean data was equally split for coding between the two coders. The coding was identical for mother and child data.
Examples of the scenarios and coding are as follows:
Home domain: A and B are sisters. They both got $20 from their mother. A bought a t-shirt with the money. A week later, B wants to wear A’s t-shirt. A says, “This is my t-shirt and I bought it with my own money.” B says, “But you’re not using it right now.” What do you think the mother should do? Why?
A participant responded to the home scenario shown above as follows: “Since she’s just trying to borrow, not keep it . . . since they are sisters, it would be good if they helped each other and shared.” This response fit one of the thematic categories, “Share; sisters,” which was subsequently coded as collectivistic and given a score of 1. “Share” was based on the decision, and “sisters” was the reason for making the decision.
Another participant responded as follows: “Even if B wants to wear it, it’s A’s t-shirt . . . even though they are family members, that’s not quite right. B can use her money to borrow A’s t-shirt.” This response was classified as “Do not share; protect private property,” coded as individualistic and given a score of 0. “Do not share” was the decision, and “protect private property” was the reason for not sharing in the situation.
Responses were coded as mixture (0.5) when they had both collectivistic and individualistic themes. For example, another participant responded that A should share because A and B are sisters, but that they should also have a conversation first because the t-shirt is A’s property. That is, the rationale was both collectivistic (“sisters”) and individualistic (“personal property”).
School domain: A class of 5th grade students is working on posters in their art class. Next week some teachers will come to select five posters for an art show. Then, one poster will be chosen for a $50 prize. “A” and “B” realize that they have some similar ideas for a really neat poster, and they want to work together. What should the teacher do? Why?
In response to the school scenario presented above, a participant said, “Of course I will let them do that. Besides prizes and other things, if two children have the same idea, they will produce a good poster.” This response was coded as “Together; better final project.” Given its implication that working together for one better outcome was desirable, this theme was classified as collectivistic and assigned a score of 1.
Another participant responded, “They want to do it together? But I think it’s better to do it individually. Even if they have similar ideas, they have unique ways of portraying those ideas, so I don’t think they will produce the same exact pictures.” This response, emphasizing uniqueness and individual achievement, was coded as “Separately; individual achievement”; it was classified as individualistic and received a score of 0.
A sample mixture response was that working together is good because that is what the students want to do; this response, emphasizing working together (collectivistic) and personal choice (individualistic), received a score of 0.5. Our coding is further exemplified in the appendix that shows the complete range of subcategories for these two scenarios.
Household sociodemographic factors
Mothers reported the highest education level they attained, entered as their number of years of education. They were also asked about household composition to identify three-generation versus nuclear-family household status. Mothers also reported their husbands’ occupation, and the responses were converted into scores using an existing scale (Nam & Boyd, 2004). The scale assigned a score for different occupations based on the combination of income level of the occupation and the education required for the occupation. The score ranged from 1 to 100 for 975 different occupations (e.g., dishwasher = 1; surgeon = 100). Three coders were involved in the occupation scoring and ensured reliability by reaching 100% consensus. Because many mothers did not work outside the home, our analyses did not use mother’s occupation score. Even if mothers were not in the study, they filled out the family sociodemographic survey instrument. Family characteristics were entered for both mother and child as they applied to both.
Analysis
Our analytic strategy was to restrict statistical tests to those required for testing each specific hypothesis, thus avoiding the accumulation of Type I error. Our main analytic tool was the independent-samples t test, supplemented when necessary for hypothesis testing by bivariate correlations. ANOVA and ANCOVA were required to test Hypothesis 3. Cohen’s d or r was used to assess the effect size of all differences. We used a paired-samples t test to compare values expressed in response to home and school scenarios, in which case Cohen’s d was adjusted to take into account the possibility of correlated means.
Given that 39% of children participated without their mothers, we could not run paired-samples t tests with mother–child pairs. In fact, when we examined the bivariate associations between the child and the mother for participating pairs, we did not find a significant association for the cultural value orientation scores between the child and the mother, either for the home domain or for the school domain. The lack of association provided strong justification for considering mothers and their children to be independent participants.
Preliminary analysis revealed that child’s gender, gender in the scenario, and scenario order were not significantly associated with value orientation, in either the home or school domain. Hence, these variables were eliminated from our analyses.
Results
This hypothesis was not confirmed; an independent-samples t test showed no significant difference between urban Koreans and rural Koreans in their home values, t(95) = −1.30, p = .199, two-tailed, d = 0.26, small effect size. Table 2 shows group means and standard deviations for the urban and rural Korean groups in the home domain. The table also shows group means and standard deviations for all four groups in both the home and school domains.
Cultural Value Mean Scores in the Home Domain and the School Domain.
Note. Scale goes from 0 (all individualistic responses) to 4 (all collectivistic responses). Each scenario is scored 0 (individualistic), 1 (collectivistic), or 0.5 (mixed), yielding 9 possible scale points for the four home scenarios and for the four school scenarios (0-4 in 0.5 increments). Means below 2 are on the individualistic side of the scale. Means above 2 are on the collectivistic side of the scale.
An independent-samples t test revealed a borderline group difference in the home values between urban Koreans in Korea and urban European Americans in the United States, t(84) = −1.89, p = .062, two-tailed, d = 0.41, small effect size.
To test this hypothesis, we first conducted a one-way ANOVA where participants’ ethnic group (Korean or European American) was entered as the independent variable and cultural value in the home domain was the dependent variable. This analysis was followed with a one-way ANCOVA to examine the role of sociodemographic factors.
The one-way ANOVA yielded a main effect of ethnic group on the home value orientation score, such that Korean ethnicity was associated with greater collectivism in the home domain (M = 2.34, SD = 0.94) than the European American ethnic group (M = 1.77, SD = 1.02, d = 0.60, medium effect size), F(1, 145) = 9.74, p = .002. However, once the covariates of three-generation residence and maternal education were entered into the one-way ANOVA, the effect of ethnicity was no longer significant. Instead, family sociodemographic factors emerged as significant predictors of value scores in the home domain (see Table 3).
ANOVA and ANCOVA: Dependent Variable = Home Domain Cultural Value Mean Score (N = 146).
Note. Sample was restricted to those who reported information on three-generation household and maternal education.
Specifically, collectivism was higher in three-generation households compared with two-generation households. Furthermore, each of the ethnic Korean groups had a significantly higher rate of three-generation households than the European American sample, which had none. For instance, almost one fourth of Korean ethnic families across the three Korean groups lived in three-generation households, compared with none of the European American families. A series of t tests showed that the frequency of three-generation households was significantly higher in each of the three Korean ethnic groups, in comparison with the European American group, which had zero three-generation households. In order not to count households twice, only children were used in this particular analysis; European American versus Korean American: M = 21%, SD = 42%, t(27) = 2.71, p = .012 (two-tailed, corrected for unequal variances), d = 0.64, medium effect size; European American versus urban Korean: M = 20%, SD = 41%, t(29) = 2.69, p = .012 (two-tailed, corrected for unequal variances), d = 0.63, medium effect size; European American versus rural Korean: M = 32%, SD = 47%, t(33) = 3.97, p < .001 (two-tailed, corrected for unequal variances), d = 0.86, large effect size.
Because none of the European American households contained three generations, we did a further analysis within the Korean ethnicity (urban Korean, rural Korean, and Korean American) to assess whether three-generation households were associated with more collectivistic home values. We carried out a t test comparing Korean participants living in three-generation households with Korean participants living in two-generation households. Combining the three Korean groups, we found that participants in three-generation households had more collectivistic home values (M = 2.65, SD = 0.87) than those in two-generation households, M = 2.25, SD = 0.94; t(140) = 2.21, p = .028, two-tailed, equal variances, d = 0.17, small effect size.
Here is an example of the response to the t-shirt scenario by a rural Korean family living with grandparents, contrasted with a rural Korean family living in a nuclear-family household:
Rural Korean child living in a three-generation household: “I think the mother should scold them [for not sharing]. Tell them to just share.” (Share, 1 = Collectivistic) Rural Korean child living in a nuclear-family household: “The mother should scold the sister [who asked to borrow the t-shirt], because she wants to wear her sister’s t-shirt that the sister bought with her own money. Each property belongs to each individual.” (Protect private property, 0 = Individualistic)
The covariate of maternal education was significant at the .023 level, with higher maternal education associated with greater individualism, as predicted (r = −.24, small effect size). Here is an example of a rural Korean mother with less education, compared with a rural Korean mother with more education responding to the t-shirt scenario:
Rural Korean mother with less education (12 years): “I would tell her [owner of the t-shirt] to let her sister borrow the t-shirt for today. Then to the sister [who borrows the t-shirt], I would tell her to take a good care of the t-shirt. As long as they wear about the same size, sharing clothes is not a big deal.” (Share, 1 = Collectivistic) Rural Korean mother with more education (19 years): “I would ask her [owner of the t-shirt] if she would like to let her sister borrow the t-shirt. If she insists that she does not want to lend the t-shirt to her sister, I will not push her because it is her own decision whether or not she decides to lend her t-shirt to her sister. Even for family members, it is not right to ask each other for difficult requests just because they are family.” (Choice, 0 = Individualistic)
The average educational levels of the Korean samples were lower than the average education level of the European American sample (Table 1). Nevertheless, maternal education level differentiated values within the three Korean ethnic groups as well; the correlation between maternal education level and collectivistic values in the home domain was significantly negative (r = −.22, p = .018, two-tailed) among participants of Korean ethnicity (n = 147).
Contrary to the hypothesis, paternal occupation score was not a significant covariate and was dropped from the analysis.
A t test confirmed this hypothesis, t(84) = 3.12, p =.004, two-tailed. In the home domain, the mean collectivism score of the Korean Americans was 2.43 out of 4 (SD = 0.91); the mean collectivism score of the European Americans was 1.77 out of 4 (SD = 1.03; d = 0.68, a medium effect size).
Paired-samples t test revealed that there was significantly greater collectivism in the home domain and greater individualism in the school domain for urban Koreans (p = .026, d = 0.44, small effect size), rural Koreans (p = .014, d = 0.47, small effect size), and Korean Americans (p = .001, d = 0.69, medium effect size). Scores were not different between the two domains for European Americans, whose values were individualistic in both the home and school domains (see Table 2 for means and standard deviations).
This hypothesis was confirmed by a series of t tests. In each group, children were significantly more collectivistic than mothers in the home domain; urban Korean: Mchild = 2.45, SD = 0.87; Mmother = 1.75, SD = 0.80; t(48) = 8.06, p = .006, d = 0.84, large effect size; rural Korean: Mchild = 2.73, SD = 0.78; Mmother = 1.62, SD = 1.08; t(45) = 3.93, p < .001, d = 1.18, very large effect size; Korean American: Mchild = 2.88, SD = 0.70; Mmother = 1.75, SD = 0.77; t(48) = 5.38, p < .001, d = 1.54, very large effect size; European American: Mchild = 2.29, SD = 0.87; Mmother = 1.13, SD = 0.84; t(34) = 4.03, p < .001, d = 1.36, very large effect size.
Discussion
To what extent is the widely endorsed notion that East is collectivistic and West is individualistic attributable to ethnic heritage culture or to variations in sociodemographic contexts? This question, important in the present era of economically driven social change, was addressed by examining individualistic–collectivistic orientations of rural and urban native Koreans in Korea and urban immigrant Korean Americans. Comparing these families of Korean ethnicity with urban European American families, results underscore that household sociodemographic factors shape home values, more so than heritage value systems.
Korean value system, rooted in the Confucian heritage, emphasizes family-centered collectivism (U. Kim & Choi, 1994, 2014). Indeed, in the home domain, Korean Americans were more collectivistic than European Americans; there was also a borderline difference between urban Korean and urban European American samples with the urban Korean sample responding in a more collectivistic fashion. Nevertheless, Koreans were no more collectivistic than European Americans in their home values once household differences in sociodemographic factors were taken into consideration. Higher maternal education predicted individualism, whereas living with grandparents predicted collectivism.
Formal schooling, which is often accompanied by urbanization and globalization, provides individualistic socialization (Greenfield, 2009). Our study adds to the literature by showing that increased education attainment may reduce the value differences between collectivistic East and individualistic West. It is also possible that the East–West difference may have been exaggerated due to the entanglement of nations and sociodemographic factors (Park, Coello et al., 2014). Further supporting the notion that school is an individualizer, all groups of Korean participants preferred more individualistic resolutions in school situations, although they preferred more collectivistic resolutions in home situations. Koreans and European Americans were equally individualistic in the school domain, demonstrating that individualistic values were perceived to be more adaptive in school situations by both ethnic groups.
Living with at least one grandparent in the household appears to be a vehicle that reinforces and maintains more traditional collectivistic values at home, as grandparents contribute to child socialization and overall dynamics of the family (Bengtson, 2001; Silverstein, Cong, & Li, 2006). The role of grandparents in child socialization and family dynamics has particular implications in rapidly changing societies such as Korea, as there would be sizable generational gaps between grandparents and parents, both of whom reinforce cultural values for children in their household. Grandparents, who grew up in a less urban, wealthy, educated, and technological world, may represent the more collectivistic child rearing described in the 1990s (U. Kim & Choi, 1994, 2014). Similarly, grandparents may play an important role in immigrant households as family members face the need to navigate both collectivistic heritage and individualistic United States cultural value systems.
The overriding effect of household sociodemographic factors above and beyond the effect of ethnicity on home values should be understood in the context of contemporary Korea. Korea has undergone rapid social change, resulting in transformation of the entire nation’s socioecology. As an example of this transformation, nationwide social change in Korea has introduced features of urban environments to rural communities such as high-technology and high education attainment, obscuring the distinctions between rural versus urban ecologies at the community level. In such situations, it may be more meaningful to examine household sociodemographic within the broader ecologies classified as rural or urban. For instance, we did not find the hypothesized group difference in the home values between rural Koreans and urban Koreans, whereas lower maternal education and three-generation households were consistently associated with collectivistic home values. It is possible that the city-like characteristics of the environment in which rural Koreans resided washed out the effect of urbanization. The rural community had advanced technology in its school and exposure to some ethnic diversity; a Latino American from the United States was the children’s English teacher, as part of the Korean government’s effort to increase the quality of education in rural schools. Thus, the rural ecology included other individualizing sociodemographic elements, possibly yielding no value difference between the rural Korean sample and the urban Korean sample. To test such alternatives, future research needs to add rural sites with less individualizing sociodemographic elements or implement the rural–urban contrast design in other traditionally collectivistic nations that have encountered less social change than Korea.
In studying immigrant populations, it is critical to understand their contexts, both before and after immigration. Our Korean immigrant mothers had by and large grown up in urban environments prior to immigration, after which they settled in an ethnic enclave, Koreatown, Los Angeles. It was in this context that the Korean American families were recruited at a large, local Korean church community. The site represented a common social sphere for this population beyond one’s religious affiliation (J. H. Kim & Min, 2002; Min, 2010).
In line with prior work (Hurh & Kim, 1984, K. C. Kim & Hurh, 1993), our findings supported the notion of “adhesive” or context-specific acculturation for Korean immigrants. Our Korean American families with immigrant mothers had more collectivistic home values compared with European American families. This group difference suggests that the heritage value system of collectivism, which was strong at the time of immigration (U. Kim & Choi, 1994), remains intact in the home domain for Korean immigrants. Yet, Korean Americans were as individualistic as European Americans in the school domain, supporting the idea that adaptation of the individualism takes place in certain mainstream contexts outside the family. That is, it appears that certain aspects of the host culture (e.g., individualistic school values) are added on to Korean immigrants’ traditional culture (e.g., collectivistic home values; Hurh & Kim, 1984).
Moreover, we can infer that in the decades since the period of immigration in the 1990s, Koreans in Korea have been undergoing adaptation to a wealthier, more technological, and more industrialized environment with greater opportunities for formal education. On the other hand, Korean Americans were the only group to be significantly more collectivistic in the home domain than the European American sample, assessed at the time of the Korean immigration. Thus, our findings suggest the possibility of cultural “freezing” in the home sphere in the process of “adhesive” or context-specific acculturation for Korean immigrant families. Collectivism is likely adaptive and reinforced in immigrant experience and struggles, combined with both the linguistic and cultural isolation of ethnic enclaves (Liu & Geron, 2008).
A robust pattern of generational difference was found, with children responding more collectivistic than mothers in each of the four groups (rural Korean, urban Korean, Korean American, European American). It may be that the development of social cognition creates more individualism between childhood and adulthood in relatively highly educated families where independence and individualism are the cultural targets. Our findings strengthen a similar claim made in a prior study; Greenfield and Quiroz (2013) found that the present sample of European American fifth-grade children had more collectivistic home values than their college-educated mothers, unlike their Latino American sample for whom the fifth-grade children and their immigrant mothers with five years of schooling were equally collectivistic. However, maternal education level was confounded with ethnicity, limiting the interpretation of the results. In our study, Korean American mothers were immigrants with collectivistic heritage culture (similar to immigrant Latino mothers) who had high education levels (similar to European American mothers). By showing that Korean American families evidenced the same generational difference as European American families, our study supports the notion that formal schooling may contribute to increased individualism between childhood and adulthood. Future research should look at this issue longitudinally.
One limitation of our study was that the European American archival data were collected about 15 years before the Korean samples. This time lag provided an advantage for the comparison with the Korean American immigrants, as the European American sample represented the environment into which they immigrated. However, it also meant that the comparison of the European American group with Korean groups living in Korea involved participants tested in different time periods. It is likely that the ethnic difference in cultural values would have been larger if the European American sample had been more recent, given that research has shown that individualism has climbed steadily in the United States up through 2010 (Park, Twenge, & Greenfield, 2014). Therefore, our findings are a conservative estimate of the ethnic group difference in values.
Using a set of hypothetical scenarios that had been designed to assess individualism and collectivism in this particular age group of children and their mothers (Greenfield & Quiroz, 2013; Raeff et al., 2000), we were able to assess participants’ individualistic–collectivistic value orientations in a developmentally appropriate way. Furthermore, our multicultural team of bilingual and bicultural researchers, as well as the school and church staff who served as community advisers, ensured the cultural sensitivity of our instruments. Nevertheless, it will be meaningful for future research to expand the scope of investigation and examine specific values both within and outside of the individualism-collectivism framework (Park, Coello et al., 2014) by using less structured and more qualitative methods.
Future research should also increase the sample size to enhance the generalization of the findings to the study populations. For example, the borderline difference between the home values of Koreans in urban Korea and European Americans in urban United States, with the Korean group manifesting greater collectivism, could become statistically significant with a larger sample.
In sum, we implemented a unique four-group design to extricate the role of sociodemographic factors, heritage culture, and immigration in shaping individualistic–collectivistic value orientations of European Americans in the United States and native and immigrant Koreans. Results highlight the influence of household sociodemographic factors on home values, domain-specificity of home collectivism and school individualism, and generational difference in home values between more collectivistic children and more individualistic mothers. Results also imply that shifting sociodemographics augment individualism in the home country, although collectivism, rooted in the heritage value system, may remain in Korean immigrant households. Together, findings provide an important outlook during this opportune time of worldwide social change and globalization.
Footnotes
Appendix
Coding Categories for T-Shirt (Home) and Poster (School) Scenario Responses
| Category (Score) | T-shirt (Home) | Poster (School) |
|---|---|---|
| Individualistic (0) | Protect private property | Separate; individual achievement |
| Choice | Separate; get more money | |
| Negotiation | Separate; be fair to others and follow the rule | |
| Collectivistic (1) | Share; brothers | Together; better final project |
| Share; don’t be selfish | Together; social reasons | |
| Share; brothers and don’t be selfish | Together; no competition/cooperation | |
| Preserve relationship | Together; no reason given (this is just the right way) | |
| Share; save money a | ||
| Mixture (0.5) | Share, but protect private property a | Together; choice |
Added for the Korean data to the categories used in previous research on European American and Latino samples.
Authors’ Note
Heejung Park is now at the Department of Psychology, Bryn Mawr College. Jenna Joo is now at the Department of Education, University of California, Santa Barbara. Blanca Quiroz is now at the Literacy Coalition of Central Texas.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by FPR–UCLA Center for Culture, Brain, and Development, UCLA Graduate Division, and UCLA Asian American Studies Center.