Evaluating a bilingual education program in Spain: the impact beyond foreign language learning

We evaluate a program that introduced bilingual education in English and Spanish in primary education in some public schools of the Madrid region in 2004. Under this program students not only study English as a foreign language but also some subjects (at least Science, History and Geography) are taught in English. Spanish and Mathematics are taught only in Spanish. The first class receiving full treatment finished Primary education in June 2010 and they took the standardized test for all 6th grade students in Madrid on the skills considered 'indispensable' at that age. This test is our measure of the outcome of primary education to evaluate the program. We have to face a double self-selection problem. One is caused by schools who decide to apply for the program, and a second one caused by students when choosing school. We take several routes to control for these selection problems. The main route to control for self-selected schools is to take advantage of the test being conducted in the same schools before and after the program was implemented in 6th grade. To control for students self-selection we combine the use of several observable characteristics (like parents? education and occupation) with the fact that most students were already enrolled at the different schools before the program was announced. Our results indicate that there is a clear negative effect on learning the subject taught in English for children whose parents have less than upper secondary education, and no clear effect for anyone on mathematical and reading skills, which were taught in Spanish.