Finding the Poster Child: Reference Patterns in OECD Country Reports

In recent years, the Organization for Economic Co-operation and Development (OECD) has become a behemoth in education policy-making. The International Organization (IO) does not only regularly organize PISA (Programme for International Student Assessment), the largest International Student Assessment up to date, but also holds the expertise and manpower to analyses these large complex datasets. Through the evaluation of student achievement levels, the OECD evaluates the effectiveness of education systems and makes recommendations based on these results, and consequently shapes education systems around the globe. Critics of the OECD argue, that these recommendations are inherently ‘Western’, thus spreading a model of education based on western labor market needs, irrespective of the needs of local systems. However, aside from PISA, the OECD also publishes a number of country reports on national education systems based on qualitative evaluations. These reports often contain assessments of the status quo but also recommendations for reforms, as well as references to other, more successful, education systems or programs. This process of highlighting of other countries’ policies can be categorized as ‘policy referencing’. Based on previous criticism toward the western-centrism of the OECD one could suspect that the OECD only uses Western or European countries as model for other education systems, or their ‘poster child’ in PISA, Finland, as a staple country to reference. In this publication I ask, if the OECD does have a ‘favorite’ country to recommend policies from and if there are significantly more references to ‘western’ education systems. To this end I analyze 296 country reports of the OECD on education published between 1961 and 2019 by extracting the reference pattern of countries from the reports. This results in a bipartite network from country reports to reference countries. I test the regional, as well as cultural distribution of this network with an Exponential Random Graph Model (ERGM) to identify reference hubs, ‘poster children’ and general patterns. The results are twofold: There is indeed a preference for references to European countries, as expected, as well as the very frequent appearance of Finland as a model country. However, there are also regional hubs detectable. Especially for reports on non-European countries, geographical neighbors seem to be referred to most often. This is an unexpected result, as it breaks up the simple suspicion that the OECD is only focusing on a Western model of education. In addition to the regional reference pattern, however, there are favorites: Even when controlling for OECD membership, member-states are still the most often referred to countries. While the euro-centrism is expected, the strong regional reference patterns are not. This result indicates, that the OECD is sensitive to local conditions in their recommendations, especially for countries with drastically different cultural traditions, such as African countries. Additionally, there is not one model of education originating from one country, but rather a multitude of aspects from all over the world that are recommended as reforms. The OECD is therefore not as euro-centric as initially assumed.