International assessments of student achievement and public confidence in education: evidence from a cross-national study

Abstract

One of the overarching goals of international large-scale assessments (ILSA) is to inform public discourse about the quality of education in different countries. To fulfil this function, the Organisation for Economic Co-operation and Development (OECD), for example, raises awareness of the Program for International Student Assessment (PISA) results through different forms of traditional and social media (e.g. press releases and other activities under the slogan PISA Day). Scholars have responded to the rapid growth of ILSA by examining public discourse through newspaper articles, policy documents, and other outlets. However, we know very little about whether and to what extent the general public is actually affected by PISA results. In order to address this gap, this study uses data regarding public trust in education from the 2011 wave of the International Social Survey Program (ISSP). Drawing on survey data from 30 countries and Hierarchical Linear Models (HLM), the study shows that PISA rankings have a significant effect on public perceptions. We find that in high performing countries the general public expresses higher levels of confidence in the education system. We discuss these patterns in the context of growing politicisation of education policy making and the use of ILSA as evidence.

Keywords:

• Public opinion
• trust
• OECD, PISA
• education system
• cross-national

Notes

1. But parents of school-age children, for example, may be more interested than others in the performance of the education system.

2. We also examine bivariate correlations between country-level variation in confidence in the education system and performance in PISA 2009. Across all three PISA domains, we found that countries that perform well on PISA also show high levels of agreement among the general public with respect to confidence in education system. Perhaps this is related to interpretation work (by the media and policy makers) that is possible in countries that are ranked in the middle or the bottom of the PISA rankings table, but not in countries that perform well.

3. In other model specifications (available upon request) we found that income has no effect on confidence in education system. For example, we found that the bivariate coefficient is small and non-significant (b=0.002, p=0.674).

4. This coefficient is easily calculated by subtracting the coefficient for the third category (below OECD average) from the coefficient for the second category (no difference from the OECD average). For example: (–0.552) – (–0.263) = 0.289.