Research suggests that second language proficiency is key in students' study progress. Prior research primarily focused on the statistical relation between proficiency and success among university students. This approach has two limitations. First, it is unclear how this statistical relation varies across different educational levels. Second, it is unclear whether these relations are informative about the causal relation between proficiency and success, which makes it difficult to develop

more »

... ention policies based on these studies. To address these shortcomings, we analyze a large historical registry dataset, which contains information on study success and language proficiency in examinees of the Dutch as a second language state exam. First, we map out the statistical dependency between language proficiency and examinees' study success in various educational levels. Next, using novel graphical approaches to causal inference, we estimate the effect of a hypothetical intervention: would an increase in language proficiency affect study success rates?