Doing Ethnographic Research in Chinese Families - Reflections on Methodological Concerns from Two Asian Cities

Abstract

This paper compares and contrasts the ethnographic practices of two non-native researchers - a Singaporean researcher studying families in mainland China and a Swedish researcher studying Chinese families in Singapore. A novel conceptual frame of ‘radius of observation positions’ has been proposed to explicate the extent of intrusion and intimacy to which researchers may venture in the private family domain. The opportunities and challenges of two positions of observation within this radius are discussed. The choice of position is largely influenced by the interacting forces of the contextual and cultural factors as well as the personhood of the researcher. The authors call for special attention to cultural sensitivity in conducting Chinese family research. Families are embedded in culture, and the possibility of accessing family spaces hinges on one's awareness of the intricacies of family cultures and realistic assessment of one's strengths and limitations in handling complex family dynamics.

Introduction

We are two researchers who have done independent ethnographic fieldwork on Chinese family relations in contemporary Asia, more specifically in Singapore and in Xiamen, China. Esther is a Singaporean researcher who has conducted fieldwork in Xiamen; Kristina, a Swedish researcher, has conducted fieldwork in Singapore. As we are not native members of the communities we chose to study, we both entered our field sites as ‘outsiders.’ One significant similarity in these two pieces of ethnographic work is that both set out to understand intergenerational dynamics within Chinese families. ¹ Esther was interested in understanding how grandparents and parents jointly raise ‘only’ children in urban Xiamen. Kristina, on the other hand, concentrated her fieldwork in Singapore on how obligations between generations are being renegotiated in the face of rapid societal change. Ethnography was chosen as it is committed to the first-and experience and exploration of a cultural setting through participant observation (Atkinson, Coffey, Delamont, Lofland, & Lofland, 2001). This method allowed the two researchers to access family dynamics embedded in the two different cultural settings; it was done through observation as well as participating, interacting, and conversing with families over an extended period of time (see for example, Bernard, 1994; Murchison, 2010). Since ethnography is dependent on building relationships with participants in the field, the primary “instrument of knowing” employed is the ‘self’ of the researcher (Ortner, 2006, p.42). Utilizing the ‘self’ to access family domain ethnographic work is especially challenging. Unlike an arm's-length methodology such as a mailed questionnaire, the presence of the researcher as she observes, asks questions about personal experiences, perceptions and interpretations can be rather threatening to some participants. Family domain research is sensitive as members naturally coalesce in the processes of preserving and protecting their traditions, secrets and habits. They strive to keep conflictual or dysfunctional behaviors from the scrutiny of outsiders (Daly, 1992).

Reflecting on our respective ethnographic practices when studying Chinese families in Asia, we make two contributions in this article. One, we discuss how the complex interplay between the researchers' characteristics and identities together with participants' receptivity determined how intimately the researchers were allowed into the families; Second, building on the work of Daly (2007), we introduce the framework Radius of positions of different observation points in family domain research and discuss the opportunities and challenges presented to each of us as researchers in different observation points. In addition, we share our observation of guiding principles of intergenerational relationships within Chinese families for the potential benefit of future researchers as they tread the intricate dynamics involved in family domain research.

The Two Cities

While we were involved in separate research projects in Xiamen and Singapore, the cultural similarity between these two cities enable comparison between ethnographic practices. Xiamen (also known as Amoy) is located on the south-east coast of China in Fujian province. Being one of China's first special economic zones, Xiamen has undergone massive economic development since the 1980s and is today a highly urbanized society. The majority of residents in Xiamen speak the Southern Min dialect (known as the Hokkien dialect in Singapore). Singapore, on the other hand, is an island city-state located just south of Malaysia in South-east Asia. Singapore was a British colony between 1819 and 1963. After a transitory union with Malaysia (1963–1965), Singapore finally gained independence in 1965. Singapore's population, which is a result of regional immigration during the British rule, is principally made up of approximately 75 percent Chinese, 14 percent Malay, and 9 per cent from India.

Xiamen and Singapore have many similarities that are of importance when studying family relations and family structure. Many of the early immigrants who came from China to settle in Singapore were from Fujian province and spoke the Chinese dialect, Hokkien. Hence, many Chinese Singaporeans and those in Xiamen share similar ancestral roots. In addition, both cities are predominantly ethnic Chinese cities which have experienced rapid economic development and urbanization, and demographic changes such as a low fertility rate and an ageing population.

wu lun () and xiao () in Transition: Chinese Families in Xiamen and Singapore

The once clear and accepted way of the Confucian teaching known as wu lun (five cardinal relations) stipulated the hierarchical relationships in traditional Chinese society between emperor and minister, father and son, husband and wife, brothers, and friends. These cardinal relationships formed the basis of proper behaviors in interpersonal relationships (Gabrenya & Hwang, 1996). The concept of xiao (filial piety) on the other hand governed intra-family relationships. It obliged parents to instruct children, and children to be unconditionally respectful and obedient, to provide for aged parents, and to conduct themselves so as to bring honor and avoid disgrace to the family name (Ho, 1987). In traditional Chinese societies, the 24 paragons of filial piety ² were exalted as exemplary ways of showing respect, care and honor to one's parents.

In China, both the praxis of wu lun and absolute filial piety came under attack after the Communist Party took power in 1949. Traditional Confucian values were derided as hindrances to the communist ethos. In the Chinese Communist Party's effort to reform the ‘old China’ many of the old practices were abolished. New laws and regulations aimed to reduce the power of the traditional patriarchal families and promote gender equality. Moreover, changes in the economic and demographic structure of China, including the employment of women and the one-child policy, have further undermined structures that supported traditional hierarchical family relationships as well as norms of filial piety (Ho, 1989).

Chinese Singaporeans, on the other hand, are generally second or third generation immigrants from China. It is interesting to observe that in Singapore, once a British colony and now an industrialized globalized city state, traditional Chinese values brought over by the ancestors of the current generation are still very much alive, although not in the original forms. To understand the context of Chinese intergenerational relations in contemporary Singapore one must take into account existing family policy and ideology. The ideal of filial piety is upheld by the Singapore state in its ambition to minimize public welfare. Contrary to China's Communist Party, who sought to combat traditional Confucian family practices, Singapore's political leadership was one of the strongest advocates of the ‘Asian values ideology’ in the late 1970s and 1980s, and, since then, has continued to actively promote traditional Asian family values. Thus, parallel to dramatic economic development, upward social mobility, and a demographic transition characterized by declining birthrates and an ageing population, the family remains the primary unit of support in Singapore. In the absence of sufficient public welfare, elderly family members are often dependent on their adult children for material support and practical care. Adult children on the other hand are not only expected to care for elderly parents and their own children, they also have to balance these familial obligations with pursuing a career outside the home. As young men and women increasingly view the two-income family as vital to upholding a desired living standard, financial support is often substituted for practical and emotional care. Therefore, whereas intergenerational support is still perceived as an expected obligation, the forms of that support are being renegotiated and reinterpreted (see also, Salaff, 1988; Teo, Graham, Yeoh, & Levy, 2003; Mehta & Thang, 2006; Göransson, 2009; Yeoh & Huang, 2010). Understanding how these macro cultures in transition might impact micro family dynamics helps ethnographers better manage ourselves and our intimate interactions with participants in conducting family domain research. As we shall see below, both these contextual factors and the ‘self’ of the researcher interacts to enable and/or limit ethnographic practice, from negotiating entry and building rapport to gathering and analyzing data.

Radius of Observation Positions

According to Daly (2007) the family researcher can choose to focus on different dimensions of family, such as ‘individuals in context,’ ‘single relationship dynamics,’ ‘complex family dynamics,’ ‘families in public spaces,’ ‘interaction of family members and social institutions,’ and ‘families in neighborhoods and communities.’ Borrowing Daly's (2007) ideas on the focus for observation in doing family research, we propose a framework that denotes the extent of intrusiveness by the researcher as ‘radius of observation positions’ in family research (see Figure 1).

This radius of positions is made up of layers of concentric circles. The innermost circle (A) depicts the researcher entering into the natural environment of the families, meaning she intrudes into the intimate space in families. To what extent the researcher is allowed into the intimate and private space of the families she studies is not a unilateral decision but a complex process of continuous negotiations with the participants. Our experience suggests that the ‘self’ of the researcher, including her characteristics and identities, interacts with the perceptions of participants and determines the researcher's position. Access to the private space (A) provides potential for the researcher to witness conflicts, tensions and other intimate interactions within the families. However, these opportunities bring with them complicated relationship challenges between the researcher and the family members; these complications require careful handling. The second concentric circle (B) is a position where the researcher engages with only one or two members of the family. The researcher aims to understand the dynamics of family relationships through the perspective of the one or two members. This brings fewer complications in terms of relationships because of the less intrusive nature of observation. This position poses some limitations on opportunities for observation. The outermost circle (C) is the least intrusive. This position is what Daly (2007) refers to as ‘informal ethnography.’ Researchers observe families from the ‘outside.’ Public places for observing families might include shopping malls, playgrounds, parks, airports, or funeral homes.

Figure 1.

Radius of Positions in Different Focus of Observation in Families Research

In this paper, we discuss how we arrived at different positions of observation, and we compare and contrast these positions of observation: Esther in circle A, and Kristina in circle B. We will not focus on observation from position C as neither of us adopted that position during our field studies. The opportunities and challenges of the different positions will be discussed. In particular, we discuss the challenges and implications of different observation positions in relation to social and cultural context. While both authors studied Chinese families, the construction and meaning of Chinese family differ in different cultural contexts. It is essential for the ethnographer doing family domain research to be sensitive to cultural meanings.

Positions of Focus in Studying Intergenerational Relationships

In this section we will address the opportunities and challenges of our respective positions of observation. Ethnographic fieldwork is a very personal experience. We have therefore deliberately chosen to present our ethnographic accounts below in first person.

Kristina: Intergenerational Obligations of Families in Singapore

My research on the renegotiation of intergenerational obligations in Singapore focused on the experiences of the middle generation, those caught between obligations to elderly parents and their own children. In this regard, my focus of observation falls into what Daly (2007) calls “individuals in context” and “single relationship dynamics,” (p. 134) or position B on the radius of positions. “Individuals in context” refers to questions of role construction – in the case of my research, the role of adult children and the intergenerational responsibilities attached to this role in a specific cultural setting. “Single relationship dynamics” likewise is a relevant description of my focus of observation, as I paid special attention to the ways in which the middle generation renegotiates familial relationships and responsibilities in everyday life.

Negotiating entry into observation position B

By focusing on a specific generation and not the family as a whole, I neither depended on building rapport with all family members, nor on gaining the consent of family members who were not part of the research. Consent was obtained by each individual who agreed to participate. They were informed of the purpose of the study and granted anonymity. Many interviews were conducted with one participant at a time, and away from other family members. For this reason, I did not face the dilemma of managing intra-family tensions or conflicts, a challenge which Esther experienced. Similar to Esther's strategy, however, I built my network of participants with the help of key liaison persons. Important key liaison persons were my Singaporean host family, Alan and Carole (pseudonyms used throughout). This young married couple had been living and working in my hometown in Sweden for a period of time. I got to know them through my aunt, who worked as assistant nurse when Carole gave birth in Sweden. By coincidence, Alan and Carole were moving back to Singapore at the time of my fieldwork, and they invited me to live with them for five months. Not only did this give me a chance to participate in the daily life of a Singaporean family, it also helped me to build a network of participants whom I could interview and spend time with on a regular basis during fieldwork. Over time, I managed to establish other clusters of participants, each cluster emanating from one or a few key liaison persons. ³ Later in this article, Esther highlights the importance ascribed to personal connections or guanxi in Chinese society, past and present. The role of personal connections, along with expectations on reciprocity and mutual obligations, is something that I also noted while conducting research in Singapore. While I was primarily interested in how sentiments of reciprocity characterize intergenerational exchanges, the same logic can be ascribed to all forms of social relationship, including the one between me and my host family. Coming from abroad, I was considered a guest and a subject of hospitality. Carole and Alan's generous offer of letting me stay in their family and take part in their activities can be seen as an act of indirect reciprocity. It reflects their appreciation to my aunt, who had invited them to her home several times during their stay in Sweden. At the same time, the Chinese family is a unit with distinct boundaries and a strong sense of preserving ‘face.’ I clearly recall how Carole, at an early stage in the development of my research, dismissed my hope of finding participants by ‘knocking on doors’ in their neighborhood. “They will not entertain you, they will probably just shut the door in your face,” she said. Certainly the urban environment also plays a part here. Life in high-rise buildings is more anonymous than in small communities, and entering a home occurs only if you are invited to do so. In this context, working through already existing networks was the most viable solution.

Implications of researcher's self

Key liaison persons are crucial in gaining access to research participants, but they are no guarantee for establishing the solid rapport needed to conduct ethnographic fieldwork. The researcher's professional role intersects with a number of other roles, such as friend, guest, foreigner, and intruder. These roles are constructed and negotiated in interaction with research participants (Jordan, 2006). How participants perceive the researcher will affect her possibilities of building rapport as well as the ways participants respond to being studied. Another important aspect here is the researcher's identity and background; gender, ethnicity, culture and age may, in various ways, facilitate or obstruct the researcher's possibilities of entering family domain. My fieldwork experience demonstrates some of these implications.

The fact that I had a different ethnic and national identity was a challenge; however, because Singapore is highly exposed to other cultures, it was easier to manage than the language barrier discussed subsequently. Not only is the population multicultural in its composition, but Singapore is in many ways a westernized society with a strong presence of foreign workers and multinational companies. English tends to be the preferred language of communication among college or university graduates and white-collar workers, such as my host family. In this context my position as a foreign researcher did not constitute a major problem. On the contrary, I felt that many of my research participants accepted and included me precisely because I was from a different cultural background. I was often invited to join in festivities and celebrations, and participants willingly answered my questions about their traditions. Thus, cultural difference between researcher and participants is not necessarily an obstacle; it may in fact work to the former's advantage. Given that I was not expected to be familiar with Chinese family customs, participants did not seem to find my inquisitiveness strange. I took on the role of a learner or novice, a role that was also assigned to me by the participants. The learner role has a number of advantages. A learner, by definition, is someone who needs to be educated and in that sense is assumed to be incompetent. She can convincingly present herself as a “non-judgmental observer” and may thus appear less intimidating to participants (Jordan, 2006, p. 174). My relatively young age (twenty-five) at the time of fieldwork and the fact that I was still a doctoral student further legitimized my role as a learner, as did the obvious fact that I was a ‘guest’ in a foreign country. On the downside, however, the very same role sometimes made me feel controlled and overprotected. In particular, my host family expressed concern if I was out late or had not informed them about my whereabouts, something that I initially found hard to get used to. I saw myself as capable and independent, not someone who had to be looked after. This clash of ideas, however, exposed our different constructions of adulthood and family life. Whereas Sweden has one of the highest percentages of single-person households worldwide, Singaporeans live with their parents as long as they remain unmarried and they often continue to do so after marriage. Singapore's public housing policies, which strongly discourage single-person households, play an important role here, but the norm of living with parents is also deeply embedded in the culture: moving away from parents without being married is seen as abandoning them.

Language differences limited the inclusion of elderly family members

As a way of managing communication in a multilingual society, English has been Singapore's official first language since 1979, but the extent to which Singaporeans use English in daily life varies greatly. ⁴ Older generations, many of whom did not have the opportunity to study English as children, tend to speak various Chinese dialects or Mandarin. Younger generations are effectively fluent in English, although the actual use of English often corresponds to educational background and occupation. Prior to fieldwork, I had not intended to focus on a particular age cohort, but once I entered the field I quickly realized that the language barrier made it difficult to involve elderly Singaporeans who did not speak English. The option of working through a translator involves a number of problems. Not only does it prevent the spontaneity and intimacy of a person-to-person conversation, it also increases the risk of misinterpretations and misunderstandings. Therefore, I deliberately chose to concentrate on participants fluent in English. Since language use in Singapore overlaps with both age and class, the group of participants was predominantly made up of middle-class individuals in their 30s and 40s. My original hope to investigate familial obligations across several generations, and from each generation's perspective, had to be narrowed down, in part due to this practical reason. I adjusted my strategy by studying the middle generation and trying to access the ways in which members of this particular generation perceive and renegotiate their responsibilities to older and younger dependants.

Opportunities provided by this position of observation

An ethical dilemma in all forms of ethnographic fieldwork is the double role of researcher and friend (Powdermaker, 1966). The ethnographer's primary research tool during fieldwork is her interpersonal skills and abilities to build rapport with research participants. Over the course of fieldwork the roles of researcher and confidante become blurred and harder to separate. This may be a unique feature of ethnography, but nevertheless the researcher has to carefully consider the effect her presence may have on participants. Entering a closed and private domain such as the family further heightens this ethical dilemma as the information shared with the researcher is often very personal in nature. As already mentioned, the risk of getting involved in intra-family power structures was eased by my position of observation (position B), as my focus was on individual members and single relationship dynamics, rather than on the family as a whole. Participants knew that I could not leak information to other family members and I, on the other hand, did not run the risk of taking sides. In that regard, my role as an outsider was an advantage. Participants often confided in me precisely because I was not a native family member nor was I going to be a permanent part of their lives. The fact that my sole purpose was scholarly and that they were guaranteed confidentiality and anonymity made me a ‘safe’ confidante.

One telling example is Angela, a female in her late thirties who became a frequent respondent over the course of my fieldwork. At an early stage of my work she informed me that she was not getting along with her father, who had been abusive to her in the past, and that she hardly kept in touch with him. Later on she confided that her parents were, in fact, divorced. She also admitted that she did not provide any kind of support to her father, thereby rejecting the cultural norm of ‘repaying’ ones' parents by providing financial/material support and practical care. During this conversation she made clear that she kept both the divorce and the fact that she did not provide any kind of assistance to her father a secret from her friends, because she did not want them to judge her. Her concern on this account is logical considering that filial piety, a teaching with roots in Confucian teaching, remains a moral imperative in contemporary Singapore. Filial piety historically defined children's moral duties to their parents and, although it is less ritualized today, it still plays a central role in family life. Adult children are expected to provide financial support and practical care for their elderly parents. In fact, the flow of resources from adult child to parent normally begins once the child takes up paid employment, often years ahead of the parents' actual retirement. Monetary contributions are perceived as an act of reciprocity or even repayment, not determined by actual need. The size of these monetary contributions depends on the child's income and housing arrangements (i.e., whether child resides with parents or independently), but in either case the amount given is quite a substantial portion of the salary (Göransson, 2009). While there is strong moral pressure to live up to the ideal of being a filial child, it is not easily fulfilled. Members of the middle generation struggle to meet the needs of both their elderly parents and their own children, which has given rise to the term ‘sandwich generation,’ (squeezed between double responsibilities). As already mentioned, the ideal of filial piety is also reinforced by the Singapore state as a means to minimize public welfare. Current family policies favor multi-generation households, and children's responsibility to maintain elderly parents who are unable to support themselves is prescribed by law since 1995 (Maintenance of Parents Act, 1996). In light of these explicit and implicit expectations on children to be filial to their parents, Angela's decision to keep her problematic family situation to herself is understandable. My role as a researcher without any links to her family and friends enabled her to express her bitterness more openly, which might not have been the case if I had engaged her family members in my research, as in position A. That I was also a cultural outsider, who did not share the Chinese ideal of filial piety and parental support, probably made me appear an even safer confidante who would not judge her.

Challenges of this position of observation

Understanding family culture. While my focus on individuals in context and single relationship dynamics worked to my advantage in certain respects, there were also a number of challenges involved. When I initiated my fieldwork I had no previous experience of conducting research in Singapore or even in South-east Asia. My knowledge was purely theoretical and my preconceptions colored by what I had read and heard. Lacking firsthand cultural knowledge I had to discover everything from scratch – from societal context (e.g., welfare structure, family policy) to the traditions and routines of everyday life. Approaching the private domain of family without being acquainted with obvious codes of behavior was a great challenge, and consequently the process of ethnographic understanding was slow. My fieldwork stretched over a period of more than one year. This time frame was necessary considering that approximately the first three months was spent establishing a network of participants as well as becoming familiar with the society and culture. Ethnographic understanding can be described as a process of constant interpretation and reinterpretation. With time, our understanding gets more complex and coherent, but it is a process that lacks an absolute endpoint (Agar, 1986). In this process, the researcher needs to reflect on how her personal experience of family “affects values, priorities, and interpretations in the research endeavor” (Daly & Dienhart, 1998, p. 101). The ethnographic examples rendered above, such as clashing ideas of adulthood and of children's obligations to aged parents, surfaced as a result of such different experiences of family. My initial difficulties in understanding the role of intergenerational exchanges of money and services was largely a result of my own notion of family life, where these forms of exchanges are not nearly as pronounced and where the state, rather than the family, is responsible for the care of the elderly.

Lack of multiple perspectives. My decision to concentrate my observations on individual family members did restrict the possibility of gaining solid insights into the complexity of family dynamics. I had no way to verify my interviewees' narratives with the rest of the family, or to investigate if different family members held different ideas of their obligations. Instead I conducted multiple interviews as well as many informal conversations with each participant; these generated deep and rich ethnographic data. I often asked participants to share their life stories. These stories are not only interesting for providing a biographical context to an individual's ideas and attitudes, they are also reflections of, and embedded in, a cultural structure (Crapanzano, 1985). The personal narratives I collected during fieldwork illustrate how the life of a single individual captures the vast transformation of Singapore since the 1970s; how the emergence of a conspicuous generation gap is negotiated in everyday life; and how the middle generation, which represents a link between the past and the future, balances its obligations to elderly parents and their offspring.

Esther: Intergenerational Childrearing in Xiamen

In my research I was principally interested in understanding how grandparents and parents in Xiamen jointly raise grandchildren. In order to have access to data on intergenerational dynamics my goal was to enter into position A within actual family environments. In fact, I had to negotiate entry to become temporarily part of the family system. Since I was a stranger to both the communities and families in Xiamen, I had to tap into the networks of three key liaison persons. It was my intention to work with a few families in Xiamen with grandparents residing under one roof and involved in different ways in providing childcare and raising the grandchildren jointly with their adult children.

Process of negotiating entry into the private space of families

Based on field visits prior to prolonged immersion in the site, I developed friendships with three key liaison persons: a professor from the local university, a clinical psychologist, and the principal of a preschool. After spelling out the sampling criteria I requested their help in the recruitment of potential research participants. Using key liaison persons was an effective way of recruitment in the Chinese community, as compared to putting advertisements in local newspapers. As Yang (1995) explained, both the traditional and contemporary Chinese are well known for their strong reliance on guanxi (interpersonal relations) as the base for social behavior. Having trusted and respected persons acting as the key liaison persons in the field may lend some trustworthiness to the research project. Introduction by these key liaison persons contributed to my access into position A within the five families that participated in my study.

Obtaining consent usually took a relational style of gradual engagement. The key liaison persons would make the initial contact by a phone call and explain briefly to the family my intention to recruit them as my research participants. When the family indicated a tentatively positive answer, the key liaison person would then take me for a home visit and make a formal introduction. The purpose and intention of the study was explained to the anchor person of the family, usually the adult son or daughter, who was usually the one granting consent. Once the anchor person agreed to participate in the research, he/she would introduce me to other family members including the grandparents and children. Of the seven families with which initial contacts were made, one felt it was too sensitive a topic to discuss and declined to be included; another was not appropriate for this study. A total of five families were included in the study.

To thank the participants for allowing me access into their families, I offered myself as an English tutor free of charge to their children/grandchildren. This strategy of reciprocity was also used by Fong (2004) in her ethnographic work in Dalian, a northern city in China. Learning English was valued by the parents in Xiamen as it gave their children an edge in the education system. The ages of the six children in the five families ranged from five to ten years of age. I designed different curricula for the children according to their age and language competency levels. I visited each family once a week, tutored the child, and then stayed to chat with the adults at home. The families often invited me over for dinner. This was the way they expressed their appreciation for my efforts in tutoring their children. I generally accepted these invitations as they were opportunities for observation and participation at position A. I also learned from my key liaison person that it was keqi (polite) to bring fruit or dessert to share with the hosts during these dinners. At the end of the field work, two sets of parents were overjoyed when reporting to me their children's improved English results. I knew that credit was also due to the children for learning effectively and I was delighted to be deemed useful by them.

Implications of researcher's ‘self’

During the introduction meetings with the families, I provided a narrative of myself to the participants. It was a conscious effort on my part to capitalize on my Chinese ancestry. Although a Chinese by ethnicity, I was born and raised in Singapore, a much more westernized society. I described myself as a huaqiao (overseas Chinese) who had a keen interest in Chinese culture and society. I also elaborated on my ancestral roots in the neighboring city, Shantou , from which my father immigrated to Singapore. This narrative seemed to be welcomed by the participants; they perceived that my purpose was to learn from them. It is interesting to note that Kristina emphasized her different ethnicity from the Chinese Singaporeans, and this was beneficial for her in gaining entry. In my case, aligning my similar ancestral background worked well in creating common ground with the participants. In my self narrative, I also disclosed to participants my professional background as a clinical social worker in Singapore. When I explained my research goal, most of the participants agreed that it was an area that needed attention.

Language similarity facilitated access to private space of participating families

In terms of language, Putonghua (which is known as Mandarin in the West and is the lingua franca in China) is my mother tongue; therefore, I enjoyed the advantage of speaking and writing fluent Chinese. Furthermore, I am familiar with the Minnan dialect spoken in Xiamen and southern Fujian province as my ancestors migrated from Shantou, a region where a variant of Minnan, Caozhou, is spoken. These two dialects are mutually intelligible. Despite some cultural differences, the language fluency enhanced my ability to participate in the field. However, I found myself occasionally having to clarify certain colloquial points with my key liaisons. Also, more effort had to be put in while communicating with the elders as most of them spoke Putonghua with the accent of their dialect tongues. For instance, one grandfather spoke with a heavy Shandong accent. I had to politely clarify his words or to ask him to repeat so that I could understand him better. Through the six months in the field I also picked up many expressions in Putonghua which were new to me.

Opportunities provided by this position of observation

One of the greatest opportunities afforded by the position of observing families in their natural environment was the opportunity to build relationship and trust with all family members. The process of rapport building, however, demanded effort on my part and willingness on the part of the participants. Although consent was granted by the anchoring family member, entry had to be continually negotiated. The grandparents were most accessible as they were usually at home. I would sometimes just drop by to chat with them, walk with them to pick up their grandchildren from school, or have lunch together with them. It was through these informal contacts that rapport was built over time. My relationship with one family, who were migrants and lived in a slum, experienced a breakthrough when I responded to their invitation for lunch in their quarters. Initially I really struggled within myself. In my field diary I reflected on my fear of having lunch in the slum. Besides my concern about the hygiene standards, I knew cooking a meal for me might exhaust a large portion of their food budget. But when I overcame my inner hurdles and went to have lunch with the family, they were visibly happy. From then on, I often ate with them and would take supplies like rice, oil and other food items as gifts. By the final phase of my field work, friendships with most of the participants had become rather strong and I had to mentally prepare both myself and the participants for termination. Many of the mothers and grandmothers phoned during the last week of my stay to express their bushede (separation anxiety) and urged me to return to Xiamen to visit them in the future.

Challenges of this position of observation

Dealing with unanticipated disclosure. As the rapport with different family members strengthened, many participants began to treat me as a confidante and friend. Often grandmothers and mothers would ventilate to me their grievances about other family members. Two grandmothers often wept silently when they related their stories of being taken for granted and exploited by their adult children. Following is an excerpt of grandmother Jiang's unanticipated disclosure:

They [son and daughter-in-law] were exploiting me when I first came to help. To be honest, I felt like a domestic maid. In actual fact my life here is exactly like a domestic maid. I should not feel this way because I am helping my own son and grandson. I stay with them, do everything for them, and they do nothing to help me. Another mother-in-law would have walked out.

Grandmother Bai also shed tears almost every time I visited her in the afternoons:

She [adult daughter] often raises her voice at me, it makes me very angry inside. I am doing everything for her. If she had to engage a domestic helper, it would cost her one thousand yuan [per month]. I do everything free of charge for her. I care for the child, cook and work from morning to night, do laundry and housework… I told her I do not blame her. I understand she has her own struggles. I do all this out of love and I take pity on her. If I didn't love her, I could have led my own life after retirement.

Initially I was rather surprised at the level of disclosure these grandmothers were willing to make. This deep disclosure could be attributed in part to participants' perceptions of the objectivity of the “stranger,” which, as Simmel noted, gave rise to the “most surprising openness – confidences which sometimes have the character of a confessional which would be carefully withheld from a more closely related person” (as cited in Lee, 1993, p. 113). Also, the grandmothers might not have large social networks. Having someone visiting them whom they deemed to be trustworthy, yet who would only be in the community for a short time and therefore posed a low risk for disclosure, provided them with an outlet for expression of personal feelings about sensitive issues in the families. I had to balance empathy with neutrality in order to avoid any strained relationship with other family members who were also my research participants. These unanticipated disclosures were invaluable. They alerted me to sensitive areas of intergenerational dynamics to which I needed to pay during my participant observation at position A of the radius of position. I knew, however, that they needed to be handled with great care. This was especially sensitive because Chinese is a shame avoidance culture (Bedford & Hwang, 2003) where jia chou bu ke wai yang (don't wash your dirty linen in public) is upheld among Chinese to preserve family pride. I could not afford to let my perceptions of the adult children be tainted by these disclosures; instead, I needed to maintain my respect for them and try to listen to their perspective too. At the same time, I had to be careful not be too ‘objective’ lest the grandmothers felt I did not understand them. I adopted the strategy of active listening and used empathic responses generously so as to help the discloser feel understood and validated. Ethically, I had to calm the intense emotions and ensure that the grandmothers were not overcome by negative emotions. In most cases, they expressed a sense of release after talking and regained their composure after some time. In a sense this position of observing complex intergenerational dynamics had an unintended therapeutic effect, even though I did not treat the time as a counseling session.

Managing participants' expectations. As a social worker with many years of clinical practice experience, skills such as establishing a trusting relationship, listening with empathy and skillful use of questions have become an ingrained part of my personhood. It was natural for me to engage my research participants as I would clients in social work practice. The dilemma, however, lay in achieving a balance between building relationships and managing the expectations of participants for me to dispense advice and solutions. In social work practice there is a thin line between engagement, assessment and intervention. Sometimes, these processes are so intertwined that it is hard to tease out the different segments. However, I was aware that my role was that of a researcher, and hence I had to pull back from the tendency to cross into intervention. The fact that I did not hide my previous experience as a social worker with children and families in Singapore worked both for and against relationship building. My research participants viewed my experience as some form of ‘expertise’ that was lacking in Xiamen. Hence, it was quite natural for them to confide problems relating to child rearing, behavior and learning issues. These chats helped me to understand their concerns and priorities in childrearing. I was happy to discuss these issues and would listen attentively and empathically, however, I had to be careful not to turn these chats into therapy sessions and to consciously refrain from giving advice, unless they expressed a strong wish to hear my views. Even then, I would express my lack of understanding of the cultural context and suggest they always take what I said with a ‘pinch of salt.’ In one family I referred the child to a local therapist for intervention since I could see that the child had behavioral and learning difficulties, but I did not feel it was suitable for me to play the role of a social worker to the child. Another danger in the social worker's role was that it might cause participants always to frame issues in a ‘pathological orientation’ when talking to me. I had to be careful not to ‘see ghosts at every corner.’

Being triangulated by certain family members. Prolonged intimate interactions with all the members within a family system over a sustained period of time meant that I had to handle situations where family members subtly looked for support for their viewpoints or feelings from me. One example was the frequent indiscriminate criticism of their son-in-law by grandparents Tian. This pair of grandparents stayed with their adult daughter periodically to help with childcare and other housework. Grandparents Tian also regularly interfered and protected their granddaughter from punishment when their son-in-law disciplined his ten-year-old daughter:

He [son-in-law] is worried that we will protect her [granddaughter]… it is not unreasonable for us [grandparents] to side with her. He hits the child so severely, don't you think we have to do something? Children are like a piece of blank paper, we adults can paint them green or red, and it is up to us. We should not blame the child. (Grandmother Tian)

It was also not uncommon for the grandparents to blame their son-in-law for setting a bad example to his daughter, and hence they aligned themselves with the child against her father. When the grandparents related these incidents to me, I felt rather awkward and uncomfortable, especially when they sometimes did it in front of the adult daughter or son-in-law. I could not help but feel that these grandparents were rallying support from me. In order not to be seen as siding with the grandparents, I initially maintained an awkward silence. However, I understood that I needed to attend to the grandparents' concerns. Hence, I coped with my discomfort by sorting out my thoughts and emotions in my field diaries and mentally rehearsed the best possible ways to avoid the trap of triangulation. This echoed Daly and Dienhart's (1998) view that the researcher's reactions of agreement, support, and interest or indifference give our participants relationship cues that communicate acceptance or rejection, withdrawal or support. Hence, the strategies I utilized to avoid being triangulated were very similar to those employed to handle unanticipated disclosure. I would listen attentively and empathically without taking sides as far as possible. I also had to challenge my own possible bias towards the son-in-law owing to the influence of grandparents Tian. To distance myself from their negative opinions, I consciously made connections with him during some home visits, taking the initiative to talk to him and let him feel that I was not against him. I tried to give him equal ‘air time’ and attempted to hear his perspective regarding these complaints by his parents-in-law.

Discussion

Complex Processes that Determined Position of Observation

Our respective accounts of ethnographic practice in Chinese families in Asia highlight how we arrived at each particular position of observation and the opportunities and challenges these positions brought. While both our practices involved certain common processes (for example, using key liaison persons for negotiating entry, and language as a facilitating or inhibiting factor of entering into families), our experiences revealed more complex processes that governed the position each researcher ultimately adopted. In this paper we conceptualize and discuss these experiences in terms of the different foci of observation depicted in figure 1. Our discussion centers on position A (complex family dynamics) and position B (single relationship/individual in context) because these are the positions our experiences stem from. While we preferred to anchor in a certain position of observation in order to achieve the intended research objectives, in actual practice, the final decision was not determined unilaterally by the researcher. Instead, it depended on complex interaction processes between the characteristics of each researcher and her potential participants. For instance, because of a language barrier, Kristina pragmatically modified her original intention to enter into position A and chose position B. That is, she concentrated on one generation instead of involving multiple generations. Compromising one's original research strategies is sometimes necessary, but never an easy decision. As ethnographers we have to continuously reflect on, be aware of, and be sensitive to participants' responses. We also have to be aware of practical issues in the site before settling for a position of observation that is realistic. One also has to bear in mind the opportunities and challenges of the final position adopted. Observation from position A has the potential of providing insights into complex family dynamics, but it also requires the researcher to manage the aspects of intrusion that it brings. In Esther's case, for example, unanticipated disclosures emerged as a significant challenge that required careful attention. In Kristina's research, it was a prudent decision to observe from position B so as to avoid the complications of using interpreters. The challenges in position B were less pronounced as she did not involve the whole family in her research. However, on the other hand, the research did not include data from multiple generations.

Position of Observation Embedded within Cultural Context

The idea of a radius of observation positions is a useful framework for family domain research as it allows conceptualization of the degree of intrusiveness in different positions as well as consideration of the opportunities and challenges each position entails. In addition, we propose that this framework has to be understood within the larger social and cultural context. The ethnographic illustrations rendered in this paper (for example closely knit families, shame avoidance culture, multigenerational cohabitation (Goh & Kuczynski, 2010), filial piety and intergenerational obligations), are characteristic of the Chinese families we studied. Our fieldwork experiences demonstrated the ways in which we had to adjust our strategies to the family culture of our research participants. In this process we also had to continuously reflect on, and cope with, the question of how our own family values influenced access negotiation and rapport building with participants. In this respect, being tuned in with the ‘self’ of the researcher is crucial in effective ethnographic practice. At the same time, when entering and collecting data in the families, researchers should not be overly reliant on impressions formed by reading published literature. We have to be aware that family cultures are never static. In rapidly changing societies, such as Singapore and Xiamen, intergenerational relations are being renegotiated and reinterpreted. Despite the notion of yi he wei gui (harmony at all costs), a notion reported in extant literature as a quality of Chinese families, our ethnographic work reveals that contradictions and dialectical dynamics are commonplace. It is our belief that as researchers we need to have a ‘third eye’ for how dialectics within families could arise from cultural transitions on the macro level. Such understandings aid us in making sense of observations, coping with emotionally charged episodes, and handling our own inner turmoil as we engage our participants while in the field. The ethnographic data presented in this article demonstrate some of the implications such cultural transitions may have on family life, specifically, declining status of the elderly and the struggle to live up to traditional expectations in a changing society.

Declining power and status of older persons in Xiamen

According to Chinese tradition, older persons occupied powerful positions in the family. Today, the weakened position of the grandparents in Xiamen can be seen in Esther's experience of the unexpected revelation by grandparents of their perceptions of exploitation by their adult children. These grandparents, brought up in the ‘old China’ era were coping with discrepancies between the traditional ideal of venerated grandparents and the starkly different reality they were experiencing (Goh, 2009). An understanding that grandparents were trying to reconcile their ideals with actual practice provided emotional space for Esther to be reflexive. It helped her to avoid being ‘triangulated’ by the grandparents' expressions of intense emotion, to retain a certain level of objectivity in her relationships with the seemingly ‘ungrateful’ adult children, and, at the same time, maintain a position of observation at A.

Guilty feelings of an ‘unfilial daughter’ in Singapore

Kristina's informant, Angela, though a well-educated, modern, English-speaking Chinese woman in her thirties, was still very much bound by the perceived filial responsibility of the Singapore society. While she justified her unwillingness to provide for her father by the abuse he inflicted on her as a child, it was apparent that she was carrying a sense of guilt for not doing so. Her action in keeping this a secret from all her friends, based on the fear that she might be judged, shows that the lack of filial action did not equate with freedom from self-imposed cultural filial expectations. An understanding of the contradiction between the seemingly modern exterior and the traditional core values that Angela had to reconcile allowed Kristina to access the rich intergenerational dynamics, even while adopting position B, a position which did not involve direct interaction with multiple generations. Adopting position B in a situation of family tensions was thus potentially less intrusive and less ethically complicated than entering the family unit and involving the other family members.

Conclusion

By comparing and contrasting our experiences of doing ethnographic fieldwork on Chinese intergenerational relations we have tried to highlight opportunities and challenges of this endeavor. As researchers interested in family domain research we have to be self-reflexive about how our own characteristics (including ethnicity, language abilities, cultural and professional backgrounds) influence us and affect interactions with our participants. The complex interplay between the characteristics of the researcher and participants' perception and receptivity to her will determine the ‘radius of position’ allowed in the families. Reflexivity is vital to doing family research, especially when one observes and interacts with families in the natural environment. Reflexivity on the meaning of family dynamics, how these dynamics/conflicts impact us as individuals, what kind of emotions they stir up in us, and how these emotions influence our research practice are critical issues for constant reflection while conducting family domain studies. As a way of theorizing different forms and levels of ethnographic family research we have proposed the framework of a radius of observation positions. This idea may serve as a methodological model for researchers who are planning and conducting research on and in families. The practical challenges and ethical dilemmas researchers face in family domain studies are sometimes heightened or eased depending on our specific position of observation. The framework presented in this paper may aid in critical reflection on the implications of different research practice in family domain research. It provides a tool for discussing the extent of intrusion and intimacy to which researchers may venture in the private family domain, as well as a tool for designing viable fieldwork strategies.