What works for high risk students? Looking for clues to improve their academic outcomes

What works for high-risk students? Looking for clues to improve their academic outcomes. 

Gudrun Vanlaar (Supervisors prof. B. De Fraine & prof. J. Van Damme)

Educational research shows that students with a low socioeconomic status (SES), an immigrant background, and/or special educational needs are more vulnerable for academic underachievement. This doctoral project focuses on the question how the gap between high- and low-risk students can be narrowed. We try to gain insight in how to improve classes and schools in such a way that all students, including high-risk students, can benefit from it.

Study 1 applied repeated measures multilevel regression models to investigate the effect of class practices that have previously demonstrated to be effective on students’ achievement and learning gains in reading comprehension: discovery learning, cooperative learning, teaching learning strategies, effective class management, well-organised and attractive instruction, and differentiated instruction. Data was retrieved from the Flemish, longitudinal SiBO-project in which about 6000 students were followed from kindergarten till grade 7. Most class practices turned out to have a similar effect for both low- and high-risk fifth graders. However, ‘discovery learning’ and ‘well-organised and attractive instruction’ appeared to be more beneficial for low-risk than for high-risk students. Group composition in terms of social and ethnic background turned out to have no significant effect on learning gains in reading comprehension.

Study 2 used the same data and method as Study 1, to test effects on mathematics achievement and learning gains. We investigated whether ‘active learning’, ‘cooperative learning environment’, ‘attractive instruction’, ‘class management’, and ‘structured instruction’ contribute to narrowing the achievement gap between high- and low-risk students, while controlling for student characteristics and socioeconomic class composition. We focus on two groups of high-risk students: low-SES students and foreign-language speaking students. Results indicate that for both high- and low-risk students, mathematics achievement at the end of grade 5 is positively associated with all tested class practices. Concerning the learning gains however, we found that foreign-language speaking students profit even more from ‘active learning and metacognitive training’, and ‘a cooperative learning environment’. No significant interaction effects were found between the students’ SES and the use of class practices.

Study 3 tests the impact of teacher and school factors of the Dynamic model of educational effectiveness (DMEE) on mathematics and science achievement and identifies factors with equalising qualities in terms of helping initially low-achieving classes to catch up with their better-achieving peers. Data were retrieved from a large-scale, longitudinal project conducted in 6 European countries (Belgium/Flanders, Cyprus, Germany, Greece, Ireland, and Slovenia). Tests in mathematics and science were administered to all grade 4 students (N=10,742) at the beginning and end of the school year. Two-level regression models were applied. Interaction effects between the factor and the group composition were estimated, while controlling for prior achievement. Our results confirm the importance of most tested teacher factors (except for modelling and qualitative structuring) and all tested school factors of DMEE for effective math and science education. The majority of these factors appear to make an even greater difference for low-achieving student groups.

Study 4 uses the same data as Study 3 and investigates the impact of the teacher factors included in the DMEE on student achievement in mathematics, but this time, differential effects were tested applying multilevel structural equation models. Our results provide further validity to the DMEE at classroom level, and suggest that most factors have differential effects for low- and high-achieving classes. We compare the methods used in Study 3 and 4.

Study 5 evaluates the effectiveness of segregated special education on math learning by exploiting two matching designs: (1) propensity score matching and (2) many-to-one matching based on the school-track advice at the end of kindergarten. Mathematics achievement in grade 1 and 2 of students who transferred to a special school after kindergarten (age 5-6) is compared to the math achievement of equally at-risk peers who promoted to the first grade in mainstream education. Our results indicate that students who transferred to special education achieved, on average, a lower attainment in mathematics in the following two years, than students equally at risk but following mainstream education.