A large body of research has found students’ educational outcomes to be related to contextual factors, such as school climate, teacher quality and students’ opportunity to learn (e.g. Kyriakides and Creemers, 2008). However, relying on cross-sectional data entails threats to causal interpretations, such as the problems of reverse causality and omitted variables. Results presented in this symposium were obtained using data from Trends in International Mathematics and Science Study (TIMSS). Through focusing on change over time, some of the mentioned threats to valid causal inference may be avoided. With six cycles of the TIMSS the IEA provides data for cross-time comparisons on country level. Through the alignment of populations and variables between the cycles the data from these studies provide unique opportunities for education analysts to conduct research on trends using repeated observations at the system/population level.

The first contribution investigates whether teacher professional development is directly or indirectly (through instructional quality) related to increased mathematics achievement in grade eight in Sweden and Norway from 2011 to 2015. It applies a two-level (students and classes) two-group (Norway and Sweden) mediation model with a dummy variable for time to investigate whether professional development and instructional quality may mediate the relation between the time and achievement. The second contributions explores teacher quality and school climate and changes in mathematics achievement from 2011 and 2015. Using a structural equation modeling approach, the analysis is done at the country-level with fixed effects for countries and time. In the third contribution content changes in mathematics between 2011 and 2015 for all participating countries are analyzed establishing a causal link between the changing OTL (Opportunity to Learn) and the changing achievement gaps between students of different social demographic and ethnic background.

The discussion will be guided by the overarching question whether and under which conditions trend analysis at the country level may be considered longitudinal and what are the resulting implications for policy and practice. The symposium provides a platform for discussions on related topics such as:

• Benefits and add-ons from trend analysis for research, policy and practice
• Limitations, challenges and approaches (methodological, dissemination of results, etc.)
• Future developments and recommendations

Kyriakides, L., & Creemers, B. P. (2008). Using a multidimensional approach to measure the impact of classroom-level factors upon student achievement: A study testing the validity of the dynamic model. School Effectiveness and School Improvement, 19(2), 183-205.