Data and analytical approach
The GSCF is a representative sample of 2000 children and their mothers, fathers, village heads, teachers, and school masters in 100 villages from 20 rural counties in Gansu Province. The first wave of the GSCF data was collected in 2000 when the children were aged 9-12 years. At that time, almost all the surveyed children were enrolled in school. These children were surveyed again in 2004. However, by then about 13 percent of the children who had participated in the 2000 survey had dropped out of school. As the focus of this study is children’s engagement at school, we limited our analytical sample to those who were still in school in 2004. After eliminating those with missing data, 1589 cases are included in our analysis.
To make full use of the rich information on children’s engagement, their perceptions of their school experience and home environment, we first ran factor analysis to identify factors and then create scales based on the results of factor analysis. These scales are then used in multivariate analysis.
Figure 2.1 presents our conceptual model. Here, two scales of children’s engagement—feelings of disengagement and motivation at schoolwork—are used as outcome measures. As disengagement and motivation reflect children’s own beliefs and feelings, it makes sense to investigate the school and home influence by using children’s own perceptions of their school experience and home environment. We consider scales of children’s self-evaluation, their evaluation of teachers and school, and their evaluation of parenting at home as the main factors that may be closely associated with children’s engagement measures, while
Figure 2.1 Conceptual model.
controlling for child, teacher, and family characteristics. Considering that the school environment may have considerable impact on children’s engagement, we use school fixed effects in all models in our multivariate analysis.
Measures
Engagement measures
Based on factor analysis, two scales of children’s engagement are created: disengagement and motivation. Disengagement is a scale created using three items, which are children’s answers to the statements, “Most of the time I do not want to go to school,” “I often feel bored at school,” and “I often feel lonely at school.” The scale of motivation is created using six items: “Schooling is very important for my future,” “I believe I can do well in my school work,” “So long as I work hard I can achieve better grades,” “I am happy in school,” “I enjoy participating in group activities at school,” and “I enjoy learning new things at school.” For all the above items, children chose from four possible answers, including completely disagree, disagree, agree, or completely agree. The scales reflect the sum of these items. The disengagement scale ranges from 3 to 12, with higher values on the scale indicating more disengagement. The motivation scale ranges from 8 to 28, with higher values indicating higher motivation.
Children’s perception of school experience
The scale of children’s self-evaluation includes five items concerning whether the child considers that they were making effects in school work, with three possible answers, from seldom to always; and how they see their academic standing in class compared with other children (with five possible answers, from poor to excellent). The scale is standardized. Evaluation of teachers and school is a scale that includes three items of children’s perception of their school experiences: “The teaching quality at our school is good,” “Our teachers care for students,” and “The teachers are fair to students.” These items have four possible answers, including completely disagree, disagree, agree, and completely agree. The scale is standardized.
Children’s perception of parenting style
Three scales of children’s perception of their home environment regarding parenting style are created based on factor analysis results. Encouraging parenting includes nine items that children rate to determine the degree to which parents care about their school work and provide encouragement when they are having problems. Communicative parenting includes eight items that measure children’s perception of whether their parents discuss with them what is on their minds, and children’s willingness to communicate with parents. The punitive parenting scale of five items has children rate whether their parents use physical or verbal punishments. All items used in these three scales have three possible answers: never, sometimes, and often. All three scales are standardized.
The descriptive statistics of all items used in constructing the scales are presented in Appendices 2.1 and 2.2.
Control variables
For children’s characteristics, we measured children’s previous school achievement by using children’s math and Chinese grades from the previous semester. They are reported by children’s homeroom teacher and are on a scale of 100, with 60 as the cut-off point for failing the class. In multivariate analysis, we also include children’s age and gender. For teachers’ characteristics, we included two measures. A teacher is considered local if they are from the same village as the child. This could indicate that they may have a better understanding of and stronger connection with the children and their families. We also include whether the teacher is female. Family background measures family wealth (logged),father’s and mother’s education, and total number of children in the family.
