The Subjective Well-Being of Children in Transnational and Non-Migrant Households: Evidence from Thailand

Abstract

To date, research that includes children’s views on parental migration has been insufficient. Based on the children’s assessment of well-being, we use a case study of Thailand to ask whether children of overseas migrant parents are less or more resilient compared to children of non-migrant parents. We make use of data from the Child Health and Migrant Parents in South-East Asia (CHAMPSEA) Project, one of the few studies that provide space for children, both of migrant and non-migrant parents, to voice their views. Our sample includes 496 children aged 9 to 11 years old. The outcome variable captures children’s subjective well-being as an indicator of whether they are a resilient child. Our multivariate analysis controls for other potential contributing factors, including the children’s individual attributes, carer’s mental health, parents’ education, family functioning as reported by the children, and household economic status in order to investigate the effects of parental migration. Qualitative information from in-depth interviews with selected carers and group interviews with community leaders are also used to explain the results. Our findings highlight the importance of taking into account children’s own perceptions. Compared to children of non-migrant parents, those of migrant parents are more likely to give a positive assessment of their own well-being. The other contributing factors include whether the child sees their family as functioning well. In Thailand, international migration is predominantly undertaken by fathers while childcare remains the responsibility of mothers, and public attitudes towards overseas migration, especially paternal migration, is generally favorable. This may help explain the positive perception of children of migrants towards themselves.

Introduction

In transnational or migrant families, the geographical separation of parents and children marks a departure from the ‘ideal’ family because of the absence of one or both parents who are most frequently considered the best caregivers. Past studies posit that parental absence for a long period of time has negative consequences for children in the dimensions of social, psychological, and academic functioning. The impacts on children can have long-lasting consequences into adulthood (Litchter, 1997).

We argue that in understanding the potential consequences of parental migration for children’s well-being, children’s own perspectives should be considered as central. The United Nations Convention on the Rights of the Child states that “the primary consideration in all actions concerning children must be in their best interest and their views must be taken into account”(Articles 3 and 12, UNICEF, n.d.). While parental absence may be seen by adults as lack of supervision, children might view it as allowing an independent life without restrictive parental control. Many existing studies on migration have presented negative consequences for left-behind children. These studies, however, mainly focus on using standard measures based on the perspectives of a responsible adult, carer, or teacher. Research on the effects of parental migration through the lens of children’s views has been insufficient to date. This is partly due to the limitation of available data on which such research is based.

While there are a number of studies of left-behind children in other regions of the world, such as Latin America and Africa, relatively little research has been done in an Asian context. In this analysis, we use a case study of Thailand to ask whether children of overseas migrant parents are more or less resilient compared to children of non-migrant parents. Our analysis makes use of data from the Child Health and Migrant Parents in South-East Asia (CHAMPSEA) Project, one of the few studies that provide space for children’s voices. We emphasize children’s views on their own well-being, how they see their family, and their general views on parents working abroad. The results highlight both the importance of taking into account children’s own perception and the importance of understanding the in-country context of migration.

This article is an initial attempt in the Thai context to explore children’s perceptions of their own lives, in particular their experiences of living in transnational families. In the first section, we discuss the theoretical perspectives on which this article is based, as well as the concepts of subjective well-being and resilience. The article then discusses in detail the issues of parental overseas migration and left-behind children, both as represented in the wider literature and, specifically, in the context of Thailand. This is followed by sections considering the methodology for the empirical analysis, and the findings and discussion. The final section offers some concluding remarks.

Theoretical Backdrop

Children’s Subjective Well-being and Parental Overseas Migration

Subjective well-being, defined as people’s cognitive and affective evaluations of their lives (Diener, 2000), is increasingly indispensable for positive psychological health, although not a sufficient condition for it (Diener et al., 1998). Subjective well-being can be in terms of cognitive states such as satisfaction with one’s marriage, work, and life, and can be an on-going affect. The concept of subjective well-being allows people to define well-being for themselves, as people have different values, goals, and strengths (Diener et al., 1998). As people throughout the world fulfill more of their basic material needs, it is likely that subjective well-being will become an even more valued goal. Studies in many countries indicate that happiness and life satisfaction have been rated as very important (King and Napa, 1997, as cited in Diener et al., 1998). Subjective well-being is sometimes labelled “happiness.”

In the context of children and parental overseas migration, in many settings (e.g., the Philippines), cultural assumptions about parent-child relationships put considerable authority as well as responsibility on parents in defining their children’s well-being. Children are portrayed as lacking in sense or knowledge, and as innocent, vulnerable, and defenseless. Implicitly, as suggested by Asis (2006), we ignore children’s concerns when we assume that a parent knows what is best for his or her child. This perception may leave children more vulnerable. More research based on children’s subjective well-being or their role as active agents rather than passive subjects is needed. Viewing children from their own perspective may lead to children taking a more active part in shaping responses to the absence of one or both parents.

Children’s Resilience and Parental Overseas Migration

Resilience, originally referred to as a personality trait, is defined as the ability to bounce back or cope successfully despite substantial adversity (Rutter, 1985). In the past few years, the most common definition of resilience is “positive adaptation despite adversity” (Luthar, 2006), thus implying two distinct dimensions of resilience: significant adversity and positive adaptation. In fact, scholars suggest that resilience is never directly measured, but is indirectly inferred from evidence of the two dimensions. The notion of resilience, once known as invulnerable and invincible, has a sense of recovery and rebounding despite adversity or change (Earvolino-Ramirez, 2007). In their review of the use of the concept of resilience in Aboriginal research, Fleming and Ledogar (2008) suggest two additional concepts. The first one is resilient reintegration, in which a confrontation with adversity leads individuals to a new level of growth. The other is the notion that resilience is an innate quality that needs only to be properly awakened. Although there may be debate about the processes that lead to resilience, it remains a useful concept for understanding certain aspects of children’s lives.

The concept of resilience has received significant attention in various fields of research, especially research on children and youth. In examining children who appeared to be invulnerable to adverse life situations, psychiatric literature indicates that children are highly resilient despite multiple risks, including poverty or other adverse conditions such as parental divorce, alcoholism, or mental illness (Earvolino-Ramirez, 2007). An aspect of resilience in research on children and youth which should be noted is the important observation that a child may demonstrate resilience in one domain, while showing disorder in another domain. In her study, Luthar (2006) notes that some children who encounter significant hardship but demonstrate academic capability at the same time undergo some psychological and emotional difficulties including anxiety and even depression. Thus, resilience in one domain, which is education in this case, may co-exist with adverse outcome in another domain (psychological/emotional disorder) in the same child.

The attributes that identify a child as resilient are distinguished as protective factors as mentioned in several research studies (e.g., Johnson and Wiechelt, 2004; Werner and Smith, 2001; Dyer and McGuinness, 1996). These include positive relationships, sense of personal worth, belief in her or his self-efficacy, sense of humor, and high expectations. In the context of parental migration, being left in the care of non-parents may be considered as an adverse condition for children of overseas migrants.

While the migration of parents potentially leads to an improved household financial situation, it may put children of migrant parents at risk due to lack of parental care and supervision. The mixed results from previous studies (e.g., ECMI/AOS-Manila et al., 2004; Asis, 2006), however, suggest that some children of migrants are resilient; that is, they have an ability to cope with living separately from their parent or parents that is reflected in their positive, rather than negative, well-being. Therefore, understanding the impacts of parental migration upon children will benefit from a consideration of the concept of resilience.

Parental Overseas Migration and Left-behind Children

The understanding of well-being among children of migrant parents in the country of origin has only recently received research attention. Especially in the South-East Asian context, despite being a long-established labor exporting region, research delving into this topic has been scarce. Most studies on left-behind families are in the context of the US as destination and countries of origin for US-bound migration flows. Extending the geographical reach of such studies is important because comparative studies help in the understanding of similarities and differences across contexts. Among the few studies within an Asian context, research on Sri Lanka and the Philippines is most prominent, due partly to their longer-established history of sending migrants abroad (Yeoh et al., 2009).

Previous research has looked at various outcomes of parental migration for children, including physical health, psychological well-being, and schooling. However, in all these dimensions, mixed results appear at best. It is widely noted that migration usually improves life conditions although it may well be damaging in other ways, e.g., higher likelihood of drug abuse, precocious sexual relationships and higher likelihood of dropping out of school (Farrow, 2007). While a study by ECMI/AOS-Manila et al., (2004) finds that children of overseas migrants in the Philippines are taller and have better general health, Cameron and Lim’s (2005) research indicates that nutritional outcomes are worse among children whose parents are both migrants compared to children from non-migrant families. The latter study, however, neither distinguished overseas from internal migration, nor did it separate absence due to marital dissolution from absence due to migration, thus making it difficult to draw any conclusions about the effects of migration.

Using the data set from the project, Jordan and Graham (2012) found that children of migrant fathers and of migrant mothers are less likely to be generally happy when compared to children living with both parents. In addition, their study finds that the migration status of parents is not significantly associated with children’s enjoyment of school, and they conclude that in the school setting, children of migrant parents do not significantly differ from their peers who have both parents currently at home. Using the same data set, rather than looking at the outcomes of general happiness and school enjoyment separately, our study further expands Jordan and Graham’s study by combining the two aspects of children’s subjective well-being, i.e., perceptions on general happiness and school enjoyment, as a proxy for resilience in relation to parental overseas migration.

On the one hand, the positive effects of migration include higher economic status of migrant families as well as better quality of childcare (ECMI/AOS-Manila et al., 2004; Asis, 2006). On the other hand, several challenges posed by parental migration suggest that the lack of parental presence for a long period of time has lasting negative consequences for children (e.g., Cortina, 2006). A lack of parental monitoring and constant support can contribute to left-behind children being vulnerable to psychological and behavioral difficulties (Robila, 2010). At the same time, parental monitoring can also create stress for children. This is noted by Robila (2010) who found that in Romania, higher monitoring was associated with children’s lower satisfaction in life. It is not clear, however, how monitoring in this case was measured. Past studies also point out different potential impacts on left-behind children. For example, a study in the Philippines suggested that, although migration creates emotional displacement for migrants and their children, it also opens up possibilities for children’s agency and independence (Asis, 2006).

The degree to which the left-behind family, including children, accept and are satisfied with migration determines the psychological well-being of the left-behind family. This is illustrated in Robila’s (2010) study which found that, although their parents are working abroad, Romanian children feel that they provide support and monitor their actions. Higher satisfaction with migration was related to higher parental support, lower levels of conflict, and lower psychological stress. The author argues that the more the children are aware and accepting of migration, the better adjusted they become (Robila, 2010).

Parental Overseas Migration and Left-behind Children in the Thai Context

Thailand has been characterized by both internal and international migration. As in other developing countries, rural-urban labor migration has become a coping strategy for the rural Thai population. Isan or the northeastern region is a prime labor sender to other urban areas, especially Bangkok. Indebtedness, which results from the inability of farming families to recover their investments in agriculture, creates the need for additional sources of income (Porpora and Lim, 1987). By the late 1980s, migrants, of whom the majority were originally from Isan, accounted for about 30 percent of the population in Bangkok (Richter et al., 1997). Seasonal migration is also common among people from the northeast due to surplus labor outside traditional planting and harvesting periods. The seasonal or periodic nature of agricultural work allows rural farmers to seek alternative sources of cash income elsewhere, while maintaining their farming activities at their places of origin. This produces both rural-urban and rural-rural migration streams.

Thailand is also known as a country that exports migrant workers. Since the mid-1970s, Thai laborers sought jobs abroad, mostly in Middle-Eastern countries where there was a huge demand for construction workers. Overseas migration from Thailand to Saudi Arabia has declined since a diplomatic incident in the late 1980s (IOM, 2003). As the demand for workers is mostly in the construction sector, about 85 percent of Thai migrants are male (Sciortino and Punpuing, 2009). Women seeking work abroad is a more recent phenomenon in Thailand. Thus, their numbers remain relatively small. Most of the women migrants are drawn into factory work, especially in South Korea and Taiwan. In 2007, the Ministry of Labor reported about 500,000 Thais working overseas (Sciortino and Punpuing, 2009).

In general, households of international migrants are relatively better off and have more material possessions. Although there are failed migrants, success stories are more noticeable and the improved economic appearance of transnational households becomes a motivation for neighbors and the younger generation. Overseas migrants are treated as role models. Parents expect their children to work overseas and remit money home. As such, they try their best to enable their children to get a job in another country. In the CHAMPSEA study communities, young people were said to look forward to the end of schooling when they would be able to go abroad to work (Group interviews with community leaders, 2008).

Despite migration being a crucial feature of rural livelihoods in Thailand, studies on the implications of parental migration for children’s welfare have been limited. This holds true for both internal and international migration. Available data at the national level in Thailand only show the prevalence of children living separately from parents without distinguishing the causes of separation. Thus, the population of children of international migrant parents is not known. For example, data at the national level, calculated from the Socioeconomic Survey of Thailand in 1986 and 2006, show that the percentage of children under 18 living with grandparents in the absence of both parents increased in the past two decades, from about two percent in 1986 to eight percent in 2006. These figures include children whose parents passed away, had marital dissolution, or migrated for work-related reasons. Unfortunately, the proportion of children whose parents are away due to working somewhere else is not known. To investigate the impacts of parental migration on the well-being of left-behind children, we need more specific data that allow us to distinguish different causes of parental absence.

Bryant’s (2005) report provides the only available estimation of numbers of children left behind in Thailand by international migrant parents. According to the report, the estimated figure of children whose parents are working overseas is half a million. This number implies that approximately 2-3 percent of Thai children have a parent overseas. How this migration affects children remains far from clear. Theoretically, migration may affect children, adolescents, and caregivers through a multiplicity of mechanisms. Public policies seeking to minimize the negative impacts of migration, while maximizing its positive effects should highlight the distinct causal mechanisms by which each group is affected in order to inform the design and choice of policy, monitor its implementation and evaluate its impacts.

The few studies that have addressed the impacts of migration on children left behind in the Thai context are mostly small-scale. Migration studies involving large samples that can compare the children of migrants and non-migrants are very rare. A study by Puapongsakorn and Sangthanapurk (1988) suggested that international migration is related to marital disruption and a rise in child truancy. Another study (Jones and Kittisuksatit, 2003), however, found little evidence that children of overseas migrant parents experienced a higher incidence of social problems.

Past studies in Thailand also looked at outcomes on morbidity and child development and education. A study by Nanthamongkolchai et al. (2006) found no relationship between parental migration and child’s sickness in the past six months and nutritional status. The study found a negative relationship between parental migration and child development and child caring. The authors show that children who lived in ‘migrant’ households were 1.4 times more likely to have lower IQ than their counterparts in non-migrant households. In Jampaklay’s article (2006), mixed impacts of parental absence on children’s educational attainment are noted. While the analysis revealed a negative effect of the long-term absence of mothers and the short-term absence of fathers, it showed a positive effect of the long-term absence of fathers. The results pointed to the importance of the duration of absence as well as whether mother or father is absent. The author argues that while migrants’ remittances may help keep left-behind children in school, remittances might also contribute to children leaving school and migrating for work.

Methodology

Data and Sample

Our analysis employs data from the CHAMPSEA Project, which is a comparative study of four countries: the Philippines, Indonesia, Thailand, and Vietnam, led by researchers at the National University of Singapore and the University of St. Andrews, in collaboration with researchers in academic institutes in the four countries. The details of CHAMPSEA’s methodology can be found elsewhere (Graham and Jordan, 2011; Jordan and Graham, 2012; Graham and Yeoh, this volume). Briefly, the project collected data on about 1,000 index children (ICs) aged 3-5 (young) or 9-11 (older) years old and their households, one child per household, for each country. Three types of structured questionnaires were used in the data collection; i.e., a household questionnaire administered to a responsible adult in the household, a carer questionnaire administered to the main carer of the index child, and an older child questionnaire administered to an index child aged 9-11 years old. In addition, the group interview of key informants was conducted in order to collect community information as well as the opinions on the impacts of overseas migration on their communities. The CHAMPSEA Project provides a unique data set that comprehensively focuses on the impacts of overseas parental migration on the health and well-being of left-behind children in four South-East Asian countries. In this study, we concentrate on the context of Thailand. This is the first study in Thailand specifically investigating the child health impacts of parental overseas migration.

The study sample in this analysis comprise 496 index children aged 9 to 11 years old with complete information on variables included in the analysis. Twenty-five cases (4.8 percent of the total sample) were dropped due to incomplete information. About half of the study children (49.6 percent) had both parents staying with them in their residence. Forty-seven percent of the study children had father-only migrant, 3.0 percent had both parents as migrants, and only two children had mother-only migrant. Most of the children had their mothers as their main caregiver or carer (92 percent).

Study Setting

For CHAMPSEA Thailand, we conducted the fieldwork in two provinces, Udon Thani and Lampang. These provinces were selected based on the study’s two main criteria. The first criterion required the identification of provinces with relatively high number of overseas migrants so as to increase the probability of meeting the survey quotas of eligible households within the time period. The second criterion was that the two sites should be diverse in terms of geographical location to capture the effects of different contexts. Udon Thani, located in the northeastern region, the least prosperous region in Thailand, has been known as the top overseas labor sending region for more than three decades. The withdrawal of American troops in 1976 from the air-base established during the Vietnam War left many local men jobless. Many later found employment as construction workers in the Middle East, which was experiencing a construction boom and had a huge demand for experienced workers. The Thai government facilitated their migration. The success stories spread quickly throughout the province and region, attracting more and more workers in search of their fortunes.

Lampang, although it ranks as the eighth labor sending region in the country, is the top labor sending province in the northern region. The province is known for its craft and artisan skills, which have been passed on from generation to generation. This is probably one of the reasons why its residents of working age are highly sought after by the construction industry, both within Thailand and overseas. As a result, there is a high prevalence of migrants working in other countries and leaving their families behind. The first group of labor migrants went abroad around 1975. Major destinations currently include countries in Africa, Taiwan, South Korea, Brunei, Libya, Qatar, Kuwait, UAE, Saudi Arabia, and Oman.

Measurement of Key Variables

The outcome variable for the modeling captures children’s subjective well-being (i.e., whether they are a resilient child) by combining children’s responses to two questions: (1) whether they enjoy school, and (2) how happy they are in general. We consider a given child who answered “always enjoy school” and “are very happy” as a resilient child. We combined results from two questions to obtain an overview of resilience from the child’s perspective. Although being happy or always enjoying school are not directly mentioned as attributes of being resilient, past studies measuring happiness have used many different indicators, including self-reported life satisfaction (e.g., Biswas-Diener et al., 2005). Previous studies suggest that self-reports of subjective well-being constitute a valid instrument that correlates with measures of self-esteem, optimism, as well as self-efficacy (Diener and Suh, 1999). These are qualities considered as protecting factors in terms of the concept of resilience. Thus, this paper regards these two subjective responses of happiness and school enjoyment as proxies for resilience. At the same time, it should be noted that this ‘combined index’ of subjective well-being is rather narrow as a measure of resilience, because other key dimensions of resilience (beyond enjoying school and general happiness) are not captured in the analysis.

The question may also be raised about the reliability of responses from children aged 9-11. While the question is valid in its own right, it implies an attitude which assumes that children are passive and incompetent actors. The age range of 9-11 is considered middle childhood. In fact, during this age range, children develop a sense of competence, forming ideas about their abilities, especially the development of attitudes towards school, academic achievement, and aspiration for the future, which can have major implications for their success as adults (Garcia Coll and Szalacha, 2004: 82). From a child rights’ perspective, society needs to recognize children as valued contributors on the issues that affect them and should be listened to as experts of their own lives, regardless of age. While a child is regarded as a person of a lower level of maturity, the notion implies their need for adult protection, love and nurturing, which is different from ignoring their voice. Child development research suggests that 9-11 years or middle childhood is an important time for children to gain a sense of responsibility along with their growing independence (CDC, 2013). In addition, it is also suggested that older children (aged 9-11) begin to think more logically about events. While they do not think abstractly the way adults do, they begin to understand more complex ideas and their insights can be amazingly deep. Although older kids still rarely see the whole picture and do not always put together their conclusions in a logical or adult way, they start to look at causes for events and ask more challenging questions (PBS, 2013). In the context of the CHAMPSEA Project, previous analysis (Jordan and Graham, 2012) demonstrates very similar results for child’s and adult’s reports on particular topics. This article focuses on children’s insights using responses from children of 9-11 years old, and as this previous work suggests, the index children are quite capable of providing insights comparable to those of adults in their households.

As will be discussed in the next section, the results from the two questions on happiness and school enjoyment are not perfectly correlated. While general happiness may sufficiently imply children’s perceived well-being, it might not reflect their well-being in school life, which is an important sphere in the lives of those in middle childhood. Hence, a combination of general happiness and enjoyment at school provides a more comprehensive measure of children’s subjective well-being, and thus resilience, compared to using each measure alone.

Our main independent variable is whether a child has one or both parents working abroad. The variable is coded 1 if one or both parents are abroad and 0 otherwise (both parents are usually resident at the same address as the index child). Our multivariate analysis controls for other potential contributing factors, including the children’s individual attributes (i.e., age, sex, presence of siblings, school performance), their views towards parental migration, carer’s mental health, parents’ education, family functioning reported by the children, and the household’s economic status. We anticipate that children who have a positive view on their parents’ migration will be better able to adjust to the parents’ absence and thus, more resilient.

Carer’s mental health could also be closely related to children’s well-being and parents’ education could affect children’s resilience. CHAMPSEA measures the mental health status of carers using the Self Reporting Questionnaire-20 (SRQ-20), an effective and low-cost screening measure of mental health developed and recommended by the World Health Organization (WHO) as a valid and adaptable method for evaluating mental distress (WHO, 1994). The SRQ-20 is a self-administered measure of psychological distress, containing 20 items, marked dichotomously (yes= 1, no= 0) over a 30-day recall period. Thus, the maximum score is 20. Previous literature (e.g., Harpham et al., 2003; Richardson et al., 2010) recommended a cut-off of 7/8 (i.e., 7 and below = probable non-case; 8 and above = probable case).

We categorize parents’ education into two groups: (1) primary level or less and (2) secondary level and more. About two-thirds of parents (64 percent of fathers and 67 percent of mothers) have no more than primary level education. Family functioning, measured by the family APGAR (adaptability, partnership, growth, affection, and resolve), as reported by index children, is another variable which might be related to children’s resilience. The family APGAR is a short and suitable screening tool to assess family functioning, proposed by Smikestein (1978) and widely validated and used, including in Asian contexts (Smilkstein et al., 1982; Preechawong, et.al., 2007; Panganiban-Corales and Medina, 2011; Kim et al., 2012). We use the family APGAR score as a continuous variable. Children who see their family as functioning well may be more likely to be resilient in the face of their parental absence. Lastly, household economic status, as measured by the asset-based household wealth index, is another external factor which is potentially associated with children’s resilience.

Findings and Discussion

Table 1 shows the percentage distribution of 496 children aged 9 to 11 according to their general happiness and school enjoyment, the two questions on which the dependent variable is based. The majority of the children are resilient, both in terms of their perceived general happiness and enjoyment at school: 94 percent considered themselves happy or very happy, while only six percent reported less happy (neither happy nor unhappy, unhappy, or very unhappy). From this minority group, most answered neither happy nor unhappy, and only three and two children reported ‘unhappy’ and ‘very unhappy,’ respectively. Most of the children also enjoy school with about two-thirds reporting that they always enjoy school and 17 percent reporting that they almost always enjoy school. Most of the remaining 17 percent answered that they enjoy school some of the time, and only three children and one child answered ‘hardly ever’ and ‘never,’ respectively. By and large, children in our study are resilient, which might reflect children at this age in general.

TABLE 1.

General Happiness and Enjoyment at School among Children Aged 9 to 11

	Percentage
General happiness
Very happy	45.0
Happy	49.0
* Neither happy or unhappy/unhappy/very unhappy	6.1
	(n=496)
Enjoyment at school
Always enjoy school	66.7
Almost always enjoy school	16.7
** Some of the time/hardly ever/never	16.5
	(n=496)

^*From this group, 3 answered unhappy and 2 very unhappy

^**From this group, 3 answered hardly ever and 1 never enjoyed school

The results of combining these two questions, as a proxy for the subjective well-being of children, are shown in the next crosstabulation table. The results are consistent with what one might expect, namely that children who enjoy school are also generally happy. Children who reported that they always enjoy school are more likely to report that they are very happy. The percentages responding ‘very happy’ among children who reported that they enjoyed school ‘always,’ ‘almost always,’ and ‘some of the time/hardly ever/never’ are 54 percent, 24 percent, and 32 percent, respectively. This shows that general happiness is not perfectly correlated with enjoying school as the percentage of children who said they were very happy is lower among those who almost always enjoyed school compared to those who did not enjoy school. Although the percentage of children who are not happy is also smallest among children who always enjoy school, a Chi-squared test reveals that the difference between the two measures (of happiness and school enjoyment) is statistically significant (Chi²=29.15, p<0.001).

Table 2 shows that children who always enjoy school are not always happy, and those who say that they are very happy do not always enjoy school. To get a stronger sense of subjective well-being, we created a proxy variable to indicate whether or not a child is resilient by taking into account both happiness in general and enjoyment at school. The children who answered very happy and always enjoy school are considered “a resilient child.” The children are classified into two groups, resilient or otherwise. Our dependent variable is thus an indicator of a resilient child who is coping well with both school and life more generally, while the other group represents less resilient children who are not necessarily unhappy but have lower subjective assessments of their own well-being. The data show that children who are very happy in general and also always enjoy school account for a little more than one-third (35.7 percent) of the sample.

TABLE 2.

Percentage of Children by General Happiness and Enjoyment at School

	Enjoyment at school
General happiness	Always	Almost always	Others	Overall
Very happy	53.5	24.1	31.7	45.0
Happy	42.0	66.3	59.8	49.0
Others	4.5	9.6	8.5	6.1
N	331	83	82	496

Chi²=29.15, p<0.001

Table 3 reports crosstabulations of the ‘resilient child’ indicator variable across all independent variables, categorized into child’s individual characteristics and household’s characteristics, including migration status of parents. We also tested for statistically significant relationships between the dependent variable (being a resilient child) and each independent variable using Chi-squared test for categorical variables and a t-test for interval variables. The proportions of resilient children are not much different across child’s characteristics and household’s characteristics, suggesting that other determining factors for children’s resilience may not be captured by the independent variables. Our bivariate results, however, show that being a resilient child is significantly related to the migration status of parents. The proportion of resilient children is evidently higher among children of migrant parents, compared to the children of non-migrant parents (41 percent compared to 31 percent). The other variable – besides the migration status of parents – significantly related to being a resilient child is the child’s positive attitude towards fathers working abroad. The proportion of resilient children is higher among children who think that it is very good if a father works abroad (38 percent), while it is only 26 percent among those who expressed a more negative attitude. The proportion of resilient children among those with positive perceptions towards a mother working abroad is also higher, but the difference is not significant.

TABLE 3.

Resilient Child by Selected Characteristics of the Child, Carers and Household

	Whether a child is “a resilient child”
Characteristics	Yes	No	Total
Dependent variable: “Resilient child”?	35.7	64.3	496
Gender
Female	37.5	62.6	251
Male	33.9	66.1	245
Age
9	36.1	63.9	166
10	35.7	64.3	171
11	35.2	64.8	159
No. of siblings
No sibling	35.6	64.4	160
Have at least one sibling	35.7	64.3	336
School performance (child report)
Average/below average	33.6	66.4	351
Above average	40.7	59.3	145
School performance (carer report)
Average/below average	35.3	64.7	431
Above average	38.5	61.5	65
IC works to support family?
No	35.5	64.5	437
Yes	37.3	62.7	59
Migration status of parents*
Non-migrant	30.5	69.5	246
Migrant	40.8	59.2	250
Mother ever moved for work since IC’s birth
Ever	30.4	69.6	69
Never	36.5	63.5	427
Carer’s mental health is poor
Yes	29.8	70.3	121
No	37.6	62.4	375
APGAR mean score (child report)	14.0 (3.2)	14.2 (2.9)	496
Father’s education
Primary or less	37.8	62.2	315
Secondary+	32.0	68.0	181
Mother’s education
Primary or less	37.5	62.5	331
Secondary+	32.1	67.9	165
Household economic status (wealth index)
Low	36.7	63.3	188
Medium	32.6	67.4	181
High	38.6	61.4	127
Father working abroad is very good*
No	26.3	73.7	99
Yes	38.0	62.0	397
Mother working abroad is very good
No	32.7	67.3	272
Yes	39.3	60.7	224

^*Chi-squared test is significant at <’0.05 level

None of the other variables are significantly related to being a resilient child. Some marked differences, however, may be noted. A higher proportion of girls than boys are resilient (37 percent compared to 34 percent). More children who reported above average school performance are resilient compared to those who did not (41 percent compared to 34 percent). This is also true for school performance as reported by main carers. More children whose main carer reported above average school performance are resilient, compared to their counterparts whose carers reported less positive school performance (38 percent compared to 35 percent), although the difference is smaller than when using school performance from child reports. A higher percentage of children whose mothers have never migrated for work since their birth were found to be resilient (37 percent), compared to children of ever-migrant mothers (30 percent). Similarly, more children whose carers do not have poor mental health were found to be resilient (38 percent), compared to children whose carers have mental health problems (30 percent). Surprisingly, a higher percentage of resilient children have fathers with less than secondary education (38 percent), compared to children of fathers with at least secondary education (32 percent); child resilience and mothers’ education show a similar difference with comparable percentages of 37 percent and 32 percent, respectively. Further, using the wealth index as a measure of household economic status, bivariate analysis indicates the highest percentage of resilient children among those in high wealth households (39 percent) but the lowest percentage of resilient children in the medium wealth group (33 percent). This raises interesting questions about the relationship between children’s subjective well-being/resilience and the socio-economic characteristics of their households, since resilience is not always associated with achievement of higher socio-economic status. However, as noted, the bivariate differences described are not necessarily indicative of significant relationships.

Our multivariate analysis fits a logistic model predicting ‘resilient child.’ The dependent variable is a summary indicator of whether the child considers him/herself generally happy and enjoying school. Prior to the multivariate analysis, we produced a correlation matrix to explore the correlations between each pair of independent variables in order to identify any problems posed by multicollinearity. The highest correlation coefficient was 0.36 (between whether mother is the carer, and whether the mother ever moved away for work since the child was born). As no pairs of independent variables are highly correlated, we include all independent variables in the multivariate model.

The results from the multivariate model are presented in Table 4. We find that, compared to children of non-migrant parents, children of migrant parent(s) are 52 percent more likely to be resilient (very happy in general and always enjoy school). The effect accounts for children’s individual and other household characteristics and is significant at the 0.05 level. This finding demonstrates an important positive impact of parental/paternal migration on children’s perceptions of their own well-being. However, it is also important to highlight the sensitivity of such a finding to the outcome measure used, as well as the geographical context. Jordan and Graham (2012) analyzed CHAMPSEA data for Indonesia, Philippines and Vietnam (not Thailand) to examine the self-reported well-being of children in different types of transnational households (father migrant, mother migrant and both parents migrant) and found that children of migrant fathers and children of migrant mothers were less likely to report being happy or very happy compared to children living with both parents, whereas there were no significant differences in reported school enjoyment between these groups. Combining the two outcome measures for this analysis of the Thai data, we find in contrast that left-behind children in Thailand are more likely to be resilient in terms of their subjective well-being. The only other factor contributing significantly to the positive well-being, or resilience, of the children is if the child sees their family as supportive and functioning well (from the APGAR score). Children who see their family as functioning well are 19 percent more likely to perceive themselves as very happy and always enjoy school. In other words, those who perceive their family is supportive are more likely to be resilient.

TABLE 4.

Odds Ratios of Independent Variables from a Logit Model Predicting Being a Resilient Child

Independent variables	Odds ratio	S.E.	95% confidence interval
Male	0.92	0.19	0.61-1.38
Age	0.99	0.12	0.78-1.26
School performance above average (carer report)	1.02	0.31	0.56-1.86
School performance above average (child report)	1.34	0.30	0.86-2.08
Work to support family	1.10	0.34	0.60-2.03
Father working overseas is very good	1.43	0.41	0.81-2.50
Mother working overseas is very good	1.25	0.28	0.81-1.93
One or both parents is/are migrant(s)	1.52	0.31 *	1.02-2.25
Mother never moved to work since the child was born	1.56	0.47	0.86-2.81
Carer is mentally healthy	1.59	0.38	0.99-2.55
APGAR (child report)	1.19	0.04 ***	1.11-1.27
Have sibling	0.99	0.21	0.65-1.50
Father has secondary+ education	0.77	0.17	0.50-1.17
Mother has secondary+ education	0.78	0.17	0.50-1.20
Medium economic status	0.77	0.18	0.49-1.22
Rich household economic status	1.11	0.28	0.67-1.83
Log likelihood		−297.86
N		496

The composition of the Thai sample also lends support to the importance of recognizing the context in which international migration occurs. In Thailand, overseas migration of migrant mothers is not as common as that of fathers and definitely less common than in the other South-East Asian study countries (i.e., Philippines, Indonesia, and Vietnam). As noted earlier, the vast majority of children of migrant parents are those of migrant fathers (93 percent). Moreover, the majority of the children have mothers as their main carer, regardless of whether or not they are from migrant (92 percent) or non-migrant families (95 percent). Due to the very small number of mother-migrant families included in the sample, we are not able to test whether our results hold across types of migrant families (father migrant, mother migrant, and both parents migrants). While Thai women are relatively autonomous in terms of mobility, they remain outstandingly dedicated to their motherly role as child carers and are seen as irreplaceable by fathers or other family members.

In this context, where most parental migrants are fathers and mothers remain prominent carers, general views (including those of children) on migrant families are positive. Children do not see the absence of fathers negatively or as problematic as long as mothers are around. Descriptive data from our analysis show that 80 percent of children say that a father working abroad is good or very good, while less than half of them (45 percent) think the same about mothers working abroad. This is also true from the adult’s perspective. Descriptive data from the responsible adult in the household survey indicate a high acceptance of fathers working abroad, with high percentages viewing it as good or very good for the child (69 percent) and for the family (64 percent). However, there is a much lower acceptance of mothers working abroad, with 21 percent regarding it as good or very good for children, and 30 percent as good or very good for the family. These gendered norms are also evident in the qualitative responses from community leaders’ during group interviews, as well as in in-depth interviews with left-behind mothers. As one of our field interviewers noted in the qualitative study:

… Interviewees perceived working abroad as more positive than negative. It is the advantage of migration that was firstly reported, especially in terms of income generating resulting from remittance. Clearly, remittance from migrants is recognized as a main support for the living of left-behinds including children of migrants. Only a few caregivers expressed concerns on emotional issue of children and their reactions when their parents went away or came back… (CHAMPSEA-Thailand’s qualitative fieldwork report)

… In general, households of international migrants are relatively better off and more sufficient in terms of material possessions. Although there are failed migrants, success stories are more evident. The improved economic appearance of migrant households became a motivation for their neighbors. Migrants are treated as role models. Parents expect their children to work overseas and remit. They would try their best for their children to get a job in other countries. Young people look forward to the end of school life and wait for their chance to go abroad… (Report from group interviews with village key informants)

The positive views of transnational households are also related to the remittances sent home by migrant parents. In the case of the Thai sample, most of those working abroad may be considered as successful. Based on household data, all children of migrant fathers lived in transnational households that received remittances from absent fathers. It was also revealed in the group interviews with community leaders that remittances from overseas migrants contribute substantively to community infrastructure improvement such as temple-, school-, and health facility-related projects. This can help to create a positive perception of migrants, as well as of migrant families and their children. Feeling proud of what their migrant parents (in our case, fathers) have done for the community can also contribute to a positive sense of well-being among children.

The degree to which absent parents maintain their relationships with left-behind children might also play a role in preventing negative impacts from parent-child separation. The role of left-behind parents in keeping intact relationships between migrants and their non-migrant family members may also be important. In another paper (Jampaklay et al., 2011) using the same data set and focusing on transnational fathering, we point out that migrant fathers do keep close contact with their family back home. The paper also made clear that left-behind mothers play a crucial role in reinforcing transnational relationships between fathers and their non-migrant children.

Concluding Remarks

Systematic research that includes children’s views on parental migration has been insufficient to date. Using children’s own views, our analysis investigates data from CHAMPSEA Thailand, the first comprehensive study that assesses the impact of parental overseas migration on children’s health and well-being in the context of Thailand. The research project is one of the few studies that provide space for children, both of migrant and non-migrant parents, to voice their views on their own well-being, their family, and how they perceive parents’ working abroad. In this analysis, we combine children’s own assessments of their general happiness and enjoyment in school as a proxy for their subjective well-being and resilience. Compared to children of non-migrant parents, those of migrant parents are more likely to be resilient. Although this evidence points to a positive impact of parental migration on the well-being of left-behind children, it must be interpreted with caution. As mentioned, our measurement of children’s resilience is only a proxy, based on self-reports from 9 to 11-year-old children on two aspects of well-being, general happiness and school enjoyment. Thus, our outcome may not fully capture the essence of children’s resilience. In addition, in Thailand, international migrants are largely fathers, while the main carer is almost always the mother. Thus, our findings indicate that children benefit most when they see their family as functioning well, and when their fathers work overseas and their mothers are the main carers.

Our study underscores the importance of taking into account children’s own perceptions in researching the well-being of those left-behind. Comparative studies that take into account the context are also important so we know to what extent our understanding can be generalized. The findings presented in this article refer to migration in the context of Thailand, where international migration is predominantly undertaken by fathers and childcare remains in mothers’ hands. The results corroborate traditional gendered norms with the role of fathers as breadwinners. At the same time, the negative impact of mothers’ migration during their children’s lifetime signals the crucial role of the mother as the main caregiver, which in Thai culture, is not easily replaceable by the extended family. In this father-migrant oriented context, attitudes towards international migration, especially paternal migration, are generally positive. This may help explain the positive perception of children towards the migration of their fathers as well as their own well-being.