The Effect of Immigrant Status on Changes in Student Achievement and Relation with Teachers in Secondary School: A Multivariate Growth Curve Model
Bieke De Fraine, Dickson Nkafu & Jan Van Damme
Students in Flanders (Belgium) are scoring very high in international researches like TIMSS and PISA. On the other hand, Flanders is not performing good when it comes to social equity. Especially immigrant students with a Turkish or Moroccan background tend to have very low scores. This underachievement of immigrant students in the high performing education system of Flanders, is a significant problem.
This paper will therefore focus on differences between Flemish students and immigrant students. Two major explanations will be explored: the relation of the students with his/her teachers and the school context. Because the school is considered as having an effect on students, this study pertains to the domain of educational effectiveness research.
In recent years, educational effectiveness researchers have acknowledged the need for more longitudinal studies in order to study changes in student outcomes over time as an important criterion for assessing school effects. The current study will therefore employ growth curve models that indicate how the student outcomes change over time.
The current project will focus on differences between Flemish students and immigrant students regarding their growth in two outcomes: language achievement and teacher-student relation. How are these two outcomes changing in secondary school? Is the gap between Flemish and immigrant students widening or narrowing over time? This project also looks at differences between schools in the changes in the two outcomes. Are some schools more effective than others in raising their students’ language achievement? And are some schools more effective than others in lowering the decline of the student-teacher relation over time?
Data were gathered through repeated assessment of about 3000 students in 50 secondary schools in Flanders. Student language achievement was assessed five times (start of Grade 7, end of Grade 7, end of Grade 8, end of Grade 10 and end of Grade 12). At four of these measurement occasions, student also filled out a questionnaire containing 10 items on their relation with the teachers. These longitudinal data were analysed by a multivariate growth curve model. The inclusion of a school level in the model makes it possible to estimate the impact of the school on the changes in the two outcomes. Student variables (immigrant status and gender) are added to the model in order to explain differences between students and to estimate the school’s value-added. Also school variables (school composition, denomination, …) are added to the model to explain differences in effectiveness.
The average growth curves indicate an increase of student language achievement, but a strong decrease of the relation with teachers in secondary school. Flemish students are growing faster in their achievement as compared to immigrant students. Also Flemish students show a smaller decline in the relation with the teachers as compared to immigrant students. It is expected that a steeper decline in the relation with the teacher correlates with a smaller increase in language achievement. It is also hypothesized that this correlation is stronger for immigrant students, indicating that their underachievement might be countered by improving the teacher-student relation.