Studies worldwide indicate that parents working away from home have negative impacts on their left-behind children in many aspects. Studies by Zhao, Yu, Wang & Glauben (2014) show that parental migration for employment has negative influences on children’s learning, for example, their academic performance is reduced  , or their learning process is interrupted as found by Zhaobao Jia & Tian (2010)  , or children find difficulty in communicating with their teachers and friends in studies by Luo, Gao & Zhang (2011)  . Rural children having absent parents have more negative emotions than those living with parents as indicated in studies of Z. Jia & W. Tian, 2010; Jingzhong & Lu (2011). They feel lonely and bored   . Children with absent parents also show their worry, insecurity and fear in studies of Fan, Su & Gill (2010)  . Moreover, children of migrant parents also have trouble in self-assessment, for example, they are un-self-confident, aloof and isolated in studies of Jingzhong & Lu (2011); Shen & Shen (2014)   . They also have low self-assessment according to studies of Xiajun (2015)  . Wen & Lin, (2012) found that children of migrant parents were less satisfied with their life and learning than those of non-migrant parents, especially children whose mothers worked away  . According to studies of Fan et al., parental migration due to employment does not only cause negative impacts on children’s learning, emotions and self-assessment, but also creates negative behaviors on left-behind children. These studies also show that boys tend to display their negative behaviours higher than girls   .
In Vietnam, it has been an obvious trend that the population in general and parents in particular leave the countryside to bigger cities or more developed countries for employment opportunities. According to the General Statistics Office, in 2015, the number of domestic migrants aged 15 years-old and over was approximately 1.24 million people in which women accounted for 57.7%, and up to 78.4% of them participated in the workforce (the remaining 21.6% are students or unemployed). The majority of these people move to urban areas. Furthermore, by the end of 2015, around 500,000 people in Vietnam were engaged in labor export to different 40 countries and territories  . Many migrant workers are forced to leave their children at home with caregivers, particularly in some families, both husband and wife work away from home. According to Hoang Ba Thinh, Nicola Piper, Vu Ngoc Binh (2012), parental migration can provide positive socio-economic impacts, especially in terms of financial support which helps maintain living and learning conditions of their children at home    .
It can be noted that studies on left-behind people including children and the elderly have been less thorough in Vietnam. According to our literature review, there are a few studies on the impact of working migrant parents to children left-behind. Graham et al. 2011 carried out a survey with 3876 children (aged 3 - 12) on well-being of children in four Southeast Asian countries (including Vietnam) based on opinions of caregivers. However, this study showed that in Vietnam, there was no difference in well-being between children whose parents work away from home and those living with their parents  . Using data from the Young Lives survey in 2007 and 2009 with 7725 children aged 5 - 8 in 4 countries Ethiopia, India, Peru and Vietnam, author Nguyen Viet Cuong supposed that migrant parents helped improve the financial situation of their families, but did not increase their children’s health and cognitive abilities. Parents working away from home were the cause of health and cognitive problems in children in Vietnam as well as in Peru and India. Children who suffered from long-term parental migration tended to be affected more negatively than those with short-term migrant parents. The lack of communication between parents and children as well as parental care was the leading cause of nutrition and learning problems in children  .
Thus, we can see that there have been very few studies on negative impacts of parental migration to left-behind children through the self-assessment of these children. This is definitely a research gap in Vietnam. The purpose of this article is to answer the following questions: 1) how is the current state of the psychological well-being of children whose parents work away from home?; 2) what is the difference in the psychological well-being among groups of children of migrant parents?; and, 3) which individual and family factors could anticipate the left behind children’s total difficulties?
This study involves quantitative cross-sectional research. Samples were selected by simple random sampling method. Total number of surveyed samples include 1119 children aged 9 to 15 (469 children of migrant parents, 650 children of non-migrant parents) and 373 caregivers (132 male and 203 female). The survey was carried out in three sites in northern Vietnam, specifically in Lam Thao district, Phu Tho province; Gia Binh district, Bac Ninh province and Ly Nhan district, Ha Nam province. The site selection is based on the fact that the ratio of local people working away from home in these sites is higher than that of other regions in the north of Vietnam. The questioned children were studying from grade 4 to grade 9 at primary and secondary schools. The study samples are described in detail in Table 1.
Table 1. Description of surveyed samples.
Note: LBC-Left-behind children; Non-LBC-Non-Left behind children.
In this study we used the Strengths and Difficulties Questionnaire (SDQ) for three groups of children including children having parents working far away from home, children living with parents and caregivers of left-behind children. The questionnaire consists of 25 items including 10 items on strength, 14 items on weakness, and one neutral item. Three answer options are available for each item, 0―not true; 1―somewhat true; 2―certainly true. The questionnaire is divided into 5 scales, five questions per scale which are hyperactivity/inattention, emotional symptoms, conduct problems, peer relationship problems, and prosocial behaviour. The scale of prosocial behaviour is positive. The remaining four scales are negative, which are added together to score the total difficulties score. The SDQ questionnaire used in this study is based on the Vietnamese version that has been made adaptable and used by Dang Hoang Minh et al. (2013)  .
2.3. Research Process
First, we contacted and received the consent of selected schools’ managing boards and homeroom teachers to carry out the survey with their students. Then, within each grade, the researchers chose one group of children whose parents work away from home, and another group of children living with their parents in order to conduct the survey. Each student was received a survey sheet/questionnaires and completed it on their own. The students and caregivers gathered at a meeting hall of each school under the guidance of the research team’s members.
2.4. Processing Techniques
All the data was processed by a Statistical Package for the Social Sciences (SPSS version 20.0).
3.1. Overview on Psychological Well-Being of Children
Table 2 shows that 17.4% and 1.5% of the total difficulties score of self-reported children having parents working far from home are in the borderline and abnormal band, respectively. The total percentage of these two scoring bands is 18.9%. For children living with parents, 16.5% of this group score in the borderline band and 1.8% in the abnormal band, thus, the total is 18.3%. However, as reported by caregivers, the total percentage of the aforementioned two scoring bands is lower with only 15.1%.
As reported by LBC, the proportion of children with hyperactivity/inattention problems was the highest with 12.1% and 1.3%, respectively, scoring in the borderline and abnormal band (the total of these two scoring bands was 13.4%). The next was children with emotional problems with 7.2% placed in borderline banding and 8.6% in abnormal banding, the total was 15.8%. As reported by caregivers, the percentage of LBC having emotional problems was the highest with 7.0% and 13.4% scoring in borderline and abnormal band, respectively (20.4% in total) and followed by those having peer problems with 12.7% and 3.3% (16% for both scoring bands).
3.2. Comparison of SDQ Score between LBC and Non-LBC
One of research questions raised by the survey team is that whether children of parents working away from home have to encounter more psychological well-being problems than those living with parents or not? The results are revealed in Table 3.
Overall, the total difficulties score of LBC is higher than that of non-LBC, 12.55 of mean score when compared with 10.17 of children living with parents.
Table 2. Percentage of children groups’ mental health.
Note: LBC-Left-behind children; Non-LBC-Non-Left behind children.
Table 3. Comparison of mental health of children of migrant parents with those living with their parents.
Note: LBC-Left-behind children; Non-LBC-Non-Left behind children.
There is a statistically significant difference between the two groups, t(958) = 7828, p = 0.00.
In terms of each aspect of mental health, we found a difference in a trend that children of migrant parents were face with more difficulties than those living with parents. Specifically, children whose parents work away from home had higher scores of hyperactivity than those living with parents, the mean scores of these two groups were 3.31 for the former and 2.56 for the latter and t(1046) = 7.310, p = 0.00. For emotional problems, the mean scores of these two groups, respectively, were 3.93 and 3.04, t(1055) = 7.849, p = 0.00; for conduct problems, 1.87 and 1.68, t(1050) = 1.899, p = 0.00 for behaviour problems; and t(1039) = 5.833, p = 0.000 for peer problems.
It is noted that there was no difference between these 2 children groups in scores of prosocial problems (t(1070) = 0.368, p = 0.71).
3.3. Comparison of Mental Health of Children of Migrant Parents
In terms of gender, there is no difference in total difficulties score between girls and boys. However, considering each aspect we found that girls were faced with more emotional problems than boys, the average score they achieved is 4.36 and it is 3.47 for boys and the statistically significant difference was t(441) = 3.929, p = 0.00; while boys had more conduct problems than girls (the average score is 2.08 and 1.68, respectively and t(443) = 2.195, p = 0.02. The group of girls having parents working away from home also reported a higher score of prosocial behavior than those of boy group with the average score of 7.36 and 6.91, respectively, t(448) = 2.527, p = 0.01).
Children aged 12 - 15 reported a higher score of hyperactivity/inattention problems than those aged 9 - 11, the average score is 3.57 and 3.03, respectively and t(443) = 2.916, p = 0.00. There is no statistical difference in remaining aspects of mental health between groups of children having migrant parents.
In terms of migration time, there is a difference in the total difficulties score between groups of children having parents working away from home F(319) = 5.408, p = 0.005. Particularly, children whose parents worked away from home for less than 2 years reported higher total difficulties score than those having absent parents for from 2 to 5 years (the average difference score between two groups was 2.44; p = 0.001). Considering each aspect of the difficulty, on notable point is that there is difference between those children groups in hyperactivity, F(341) = 5.921, p = 0.003, in which children of parents working away from home for less than 2 years reported higher score in hyperactivity problems than those having parents working away for 2 - 5 years (the average score of difference between two groups was 0.84, p = 00.1); conduct problems F(339) = 10.396, p = 0.000, in which children whose parents working away from home for less than two years had higher score in conduct problems than those having parents working away for 2 - 5 years (the average score of difference between 2 groups was 1.00, p = 0.00). However, there is no statistical difference in aspects of emotional problems, F(338) = 2.412, p = 0.09; peer problems F(335) = 2.645, p = 0.07. It is also noted that there is a difference in these children groups in terms of prosocial behaviour, F(342) = 7.788, p = 0.00.
The data in Table 4 shows that there is a difference in the total difficulties score between groups of children whose parents migrate for employment (F(397) = 4.743, p = 0.00). Considering every aspect, there is a statistically significant difference in hyperactivity problems between these children groups F(429) = 3.699, p = 0.02, in which the group of children with both parents working far away from home reported higher score of hyperactivity problems than those having migrant father (the average difference score is 0.56, p = 0.00); emotional problem F(424) = 4.124, p = 0.01, in which the group children whose both parents work away from home reported higher scores of emotional problems than those with only fathers migrating for employment (the average difference score between 2 groups is 0.71, p = 0.00).
In terms of residential area of surveyed participants, one notable point is that there is statistical significance in the total difficulties scores in children whose parents work away from home (F(412) = 12.709, p = 0.00). Considering each aspect, we found that there is statistical significance difference between children groups in hyperactivity problems (F(446) = 11.380, p = 0.00). Children group in Phu Tho province reported a higher score of hyperactivity problems than those in Bac Ninh province (the average difference score between two groups is 0.66,
Table 4. Comparison of mental health of LBC (N = 464).
Note: Significance level *-p < 0.05; **-p < 0.01; *** -p < 0.001.
p = 0.00). Emotional problems: F(442) = 7.904, p = 0.00, children group in Phu Tho also reported a higher score of emotional problems than those in Bac Ninh (the average difference score between two groups is 0.91, p = 0.00). Conduct problems: F(444) = 9.879, p = 0.00, children group in Ha Nam province reported a higher score of conduct problems than those in Phu Tho (the average difference score between two groups = 0.63, p = 0.00). The data in Table 4 also indicates a statistical significance difference in prosocial problems between children groups in different localities F(449) = 11.478, p = 0.00, in which there is a difference between children groups in Phu Tho and Ha Nam (the average difference score between two groups is 0.99, p = 0.00).
The results in the Table 4 also reveal that in terms of parent-children relationship, there is no statistical significance in expression of difficulties of children having migrant parents.
3.4. Some Factors Predicting the Total Difficulties of LBC
One of our research questions is which individual and family factors could anticipate the children’s total difficulties?
In order to model the impact on children’s total difficulties, an analysis of multiple linear regression was carried out by establishing Enter’s models (at the same time putting the variables into the model). Multiple linear regression model is as follows:
where “Yt.d”―is the indicator of children’s total difficulties; “β0”―is a constant; “β1”―is the regression coefficients; “Xi” is the independent variables included in the model; “k” is independent variables of the model.
We set up the regression model with two variable groups at individual and family levels: Model 1 includes only independent variables belonging to personal characteristics of children, specifically including age; self-esteem; educational stages/level of school; self-assessment of academic performance; self-assessment of life success. Model 2 has the variables which had been included in the Model 1 and additional independent variables of the children’s family, inclusive of the number of siblings living with them; length of parental migration; father or mother who works away from home; frequency of parents’ visits and frequency of communication between children and their parents. In total, 5 variables under the group of children’s personal characteristics and 5 variables of family group were analyzed in the models. Prior to the multiple linear regression, checking for assumption violations was carried out which indicated that the assumptions was basically met.
Application of multiple linear regression analysis has provided us the following results (Table 5).
Testing of the appropriateness of the model showed that the F-statistic in both models was at a very high significance level (p < 0.001). Therefore, it can be confirmed that the database is appropriate for these two models. In Model 1, 5 va-
Table 5. Linear regression model predicting the total difficulties of LBC.
Note: Significance level *-p < 0.05; **-p < 0.01; ***-p < 0.001.
riables of children’s individual characteristics can explain 26.6% of variation of indicators of total difficulties in children with migrant parents. In Model 2, when the variable group of family factors (5 variables) is added, the adjusted R- squared increased to 28.3% compared with Model 1 which means that the addition of independent variables of children’s family factors in Model 2 strengthen the explanation for the influence of the independent variables to the total difficulties of children of migrant parents.
Testing statistical assumption for regression coefficients in the models reveals that some hypotheses of this research on the effects of variables is confirmed (β ≠ 0). However, the number of variables that can affect the statistical significance in each model is the same in these models, 2 variables in Model 1 and 3 variables in Model 2. We believe that children’s self-assessment of their school performance and self-esteem is the key factor affecting their total difficulties. It means that, if left-behind children’s self-esteem and success in their school performance can reduces the total difficulties of left-behind children in this study.
If parental migration, for such as going to a city or working abroad for employment opportunities is considered as a livelihood strategy that helps rural workers get a job and improve their families’ living conditions, a potential cost for the psychological well-being of these families’ children due to their separation from parents needs to be taken into account  . The study of the psychological well-being of children with parents working away from home in Northern Vietnam is one of the first studies on separation between children and their absent parents. The self-report and caregiver-report SDQ versions were used in this study. Results of a number of previous studies on the impact of parental migration for employment on children's psychological well-being remain very different.
Comparing the results of our study with that of a research by Amstadter et al. with 1368 adolescents in two Central Vietnam provinces of Da Nang and Khanh Hoa, using the same SDQ scale, the mean scores of this study is higher (15.1% as rated by parents, compared to 9.1%)  ; This percentage is also higher than that in a study by Dang Hoang Minh et al. (2013) carried out with 1314 parents of children aged 6 - 16 years, 15.1% compared to 13.18%  . This is lower than that of Ngo Thanh Hoi et al. who also used the SDQ scale in children aged 10 - 16 years in some urban districts in Hanoi (19.46%). However, overall, our research results fall in the range of 10% to 20% which is the threshold of mental health problems in Asian children (cited according to  ).
The compared total difficulties score reported by children having parents working away from home in our study is higher than that of children in a study of Dang Hoang Minh et al. who also used a SDQ scale for 591 adolescents(18.9% vs. 10.73%). The reason for this difference is that we carried out surveys with children of migrant parents in only three provinces, while the study by Dang Hoang Minh et al. were implemented at 60 sites in 10 representative provinces of Vietnam.
Research by Graham et al. (2011) showed that there was no difference in the psychological well-being of Vietnamese children of migrant parents and children of non-migrant parents  . Conversely, research by Nguyen Viet Cuong (2015) showed that, lack of long-term parental care caused nutrition, psychological well-being and school problems in the children with parents migrating  . Our study demonstrated that children of migrant parents had a higher total difficulties score than that of children with non-migrant parents (Table 4). In addition, the psychological well-being problems tend to increase in children whose mothers or both parents migrating for employment.
Parents working away from home (in city or abroad) and leaving their children to their relatives (grandparents in mother’s or father’s side) mean a separation event to these left-behind children. According to scholars of attachment theory, the insecurity and distance of these children with their caregivers since childhood is one the factors which are able to predict psychological well-being problems such as anxiety, depression, hyperactivity and behaviour disorders  . Researches by authors in the field of mental health also revealed the same results. The separating children from their parents for various reasons are likely to affect their psychological well-being   .
In Northern Vietnam, with the heavy influence of Confucianism, women in the family are involved in housework, caring and educating their children   . When a woman leaves home to work in the city or abroad, her partner has to take responsibility for every housework and child care. As a result, the husband plays two roles of either earning money or doing housework and childcare, which makes him stressed. Rapid changes in men’s roles in families having their wives working away from home can lead to violent reactions, including violence against family members  . Children are able to become victims for men to give vent to their frustrations and tensions due to their separation from wives and their dual role in the families. This partially explains a higher total difficulties score in children of mothers migrating for employment than those of fathers in the same situation.
In this study, the proportion of children of migrant parents under borderline and abnormal categories is higher than that of children living with their parents. In order to confirm this, further research on this topic is needed for other children’s groups since our study was conducted in three residential areas having a higher number of export workers than other areas in Northern Vietnam. Children left behind in a community where many parents are away for employment do not showed more psychological health problems than children living with their parents. Graham and his colleagues pointed out that left-behind children in the Philippines did not have more psychological health problems than children living with their parents. When there are many children left in the community, they themselves will feel that they are not the exception or different from the others, jealousy and ridicule between children do not occur. It is also a factor protecting children from many psychological health problems  .
According to SDQ score scale, the proportion of children of migrant parents under borderline and abnormal categories is higher than that of children living with their parents and also higher than other children groups in some studies in Vietnam.
Comparison between LBC and Non-LBC groups indicated a statistically significant difference in the total difficulties score between the two groups and each aspect of the psychological well-being.
Self-esteem and self-assessment of academic performance is the key factors affecting total difficulties of left-behind children. It means that, if left-behind children’s self-esteem and success in their school performance increases, the total difficulties of these children can be reduced.
The results of the study show that parental migration due to employment have a negative influence on their children's psychological well-being. Therefore, further research on this issue is needed so that solutions are available to soften negative impacts of parental migration for employment on their children.
This research is funded by Vietnam National University, Hanoi (VNU) under project number QG.15.43.