Social Capital and Mathematics Achievement of Fourth and Fifth Grade Children in Segregated Primary Schools

ABSTRACT

Scholars have consistently demonstrated that the socioeconomic composition of the pupil body is related to academic achievement. The effect of ethnic/immigrant concentration, on the other hand, is more controversial, as some have found no impact of the ethnic/immigrant composition when other aspects were taken into account. Social capital theory claims that it is possible to compensate for a disadvantaged background or school composition when pupils benefit from being integrated in specific social structures. This article tests whether social capital is positively related to the mathematics achievement of pupils in the fourth and fifth grades of Flemish primary schools in which most of the pupils have a low socioeconomic and/or an ethnic/immigrant background (i.e. segregated schools).

Notes

¹ Brothers and sisters were counted as one family member by the Flemish Department of Education.

² We used the information of children in fifth grade classes included in this study that were present in all waves and children in included fifth grade classes for which we do not have data for all the waves but who, according to the school administration lists, should have been present in all three waves. Both groups of children have missing values. The level of missingness is reported in Table 2. We have 376 children in included fifth grade classes that are in the school administration lists of all waves. We compare their last wave mathematics scores with the scores of children in the last wave (42) that were not included in this group of 376. There were insignificant differences (p > .05) between both groups (tested with an independent t-test with 42 versus 366 students). Six children who, according to the school administration lists, should have been present in all three waves had missing information on almost every variable. They are excluded from analysis.

³ We also tested alternative covariance structures (first-order autoregressive and Toeplitz structures) with a model that includes only a random linear time effect and a fixed quadratic time effect. We did not find any significant improvement, so we stuck to the unstructured covariance structures.

Additional information

Notes on contributors

Kenneth Hemmerechts

Kenneth Hemmerechts is a doctoral researcher at the University of Brussels in Etterbeek (Brussels) (VUB) where he is currently working in the field of the sociology of education. Previously, he has worked at different universities in Belgium on a variety of topics, including: migration, criminal recidivism, fraud, police capacity, employment and trade unionism, genocide and social theory. [email protected]

Nohemi Jocabeth Echeverria Vicente

Nohemi Jocabeth Echeverria Vicente is currently pursuing a PhD in Political Science at the Vrije Universiteit Brussel (VUB), as part of a program in collaboration with the Organization of the American States (OAS). Her research focuses on peace education in violent contexts. Nohemi holds a Master in Politics with specialization in International Relations from the Jawaharlal Nehru University (JNU), India and a Bachelor in International Relations from the Universidad Nacional Autónoma de México (UNAM), Mexico. [email protected]

Orhan Agirdag

Orhan Agirdag (PhD in Sociology) is a tenure track professor at the KU Leuven and the University of Amsterdam. His main research interests include inequalities in education, teacher education, multilingualism and religiosity. Currently, Orhan Agirdag also holds the chair of scientific research at the Netherlands Initiative for Education Research (NRO). [email protected]

Dimokritos Kavadias

Dimokritos Kavadias is professor at the Political science department of the Vrije Universiteit Brussel. He currently teaches methodology courses and political psychology. His current research activities focus on political socialisation, political psychology, civic education, and educational policy. [email protected]