Reading achievement growth across countries

International surveys of students like PIRLS and PISA provide comparisons of learning achievement across participating countries. Although country rankings usually receive considerable attention, other results are far less discussed. While ranking position of a country is clearly important when assessing the overall potential of its students, this is less informative for education policy as these differences might be strongly affected by other characteristics than just the performance of school systems. In national assessments, school mean score is rarely taken as a measure of school efforts and the so-called value added scores are often considered as much more reliable indicators of school performance. We provide similar estimates to value-added scores but at the country level. We compare reading achievement in primary schools as measured by PIRLS to reading achievement of 15-year-olds in the PISA survey. We discuss various issues related to comparability of the results of these two surveys. We adjust achievement growth estimates for differences in the distribution of student background characteristics and for the difference in testing age across countries and surveys. We also provide a direct test of how changing a pool of test items would affect country rankings in respect to achievement growth captured by our methods. Finally, we compare achievement growth at different student performance levels using quantile regression approach. Generally, the results demonstrate remarkable consistency suggesting noticeable differences across countries in the reading achievement growth between primary and secondary education levels. The quantile regression evidence demonstrates also in which countries the lowor high-achieving students are progressing relatively slower or faster. These results can be used to assess the effectiveness of different school systems in their work with students between the age of 10 and 15.