Differences in the value added of primary and lower secondary schools in Norway
A quantitative study
Lots of research show that pupils’ family background, such as parents' education, have implications for how well they are doing at school. There are differences in the composition of pupils between schools in Norway. Even though a school has weak unadjusted results averaged at the school level, it does not necessarily contribute little to the pupils' learning. In order to identify how much schools contribute to pupils learning one must control for differences between schools in terms of pupils composition.
By estimating indicators of school contribution, more specifically value-added indicators and cross-sectional indicators, this study identifies and analyzes the contribution of schools and municipalities to students’ achievement in Norway. The indicators of school contribution can be interpreted as the expected outcome of a school had its student body been average along all student characteristics included in the analysis. Therefore, the indicators can be used to compare schools with different compositions of pupils.
This study estimates and presents indicators of school contribution for lower primary school, higher primary school and lower secondary school. We find that the indicators provide important additional information on schools when compared to unadjusted results. However, it is important to be aware of the limitations. The indicators for lower primary and higher primary school reflect the contribution of schools to pupils’ basic skills in English, reading and math. In much the same way, the indicators for lower secondary school reflect the contribution of schools in the subject the pupils are examined in. Besides this, the indicators do not reflect how much the school contributes in other subjects and they do not reflect the schools ability to ensure other areas of its mission.
We find that compared to unadjusted results averaged at the school level the differences in the indicators of schools contribution are significantly smaller. This means there is less difference in how much schools contribute to pupils learning in Norway, than suggested by unadjusted results. A large part of observed differences in unadjusted school results can be explained by the composition of pupils.
Moreover, the report highlights the importance of taking into account uncertainty when interpreting the indicators of school contribution. We find that given a 95 percent confidence interval, the majority of schools do not differ from the country average, which we have set to 3.4 points. It is very important to be aware of the uncertainty in the indicators when they are being used for policy development purposes or school development purposes. Measures of uncertainty must be considered when the indicators are interpreted.
This study shows that it is possible to estimate robust indicators which identify which schools and municipalities have high and low indicators of contribution. However, as with all research, further research on indicators of school contribution must strive to develop the best possible indicators. This will always be work in progress. In addition, it will be up to further research to answer why some schools contribute more than others and how schools and municipalities should organize in order to contribute to pupils learning in the best way possible.