Latin American students consistently score low on international tests of cognitive skills. In the PISA 2012 results, students in seven Latin
I. IntroductionIn a recent analysis of on economic growth in Latin America, Hanushek and Woessmann [2012] argue that growth rates have been low despite increases in average levels of schooling because students' cognitive skills are "truly dismal." They show that student scores on mathematics and science tests are considerably lower in Latin America than in other countries with similar levels of schooling, and they conclude that, "the average student seems to get much less learning, as depicted by the test scores, for each year of schooling than the average student in the rest of the world."It is natural to presume that if students in Latin America are not learning as much as students elsewhere, the teachers or the educational system must be at fault. But student learning is a function of activity both within and outside the school system. So the activity in the educational system may not be the whole cause, or even the most important cause of the low scores.Breton [2015] presents evidence that high average test scores in other parts of the world, especially in East Asia, are due in part to the extensive private tutoring that students receive to raise these scores. In addition, it has been known since Coleman [1966] that a student's test scores are substantially determined by the socioeconomic characteristics of the student's family. Since Latin American countries are characterized by an unusually high level of income inequality, with a large share of poor families, the low scores could be due to the low average socioeconomic characteristics of these families.Alternatively, the low scores could be a result of Latin American culture, which assigns less importance to student achievement on standardized tests than countries with Confucianbased cultures. The average level of schooling and the average scores on standardized tests are so uniform across Latin America that there may be an underlying cultural basis for these low scores.In this study we attempt to quantify the causes of the low average test scores in Latin America, using cross-country data from the PISA 2012 evaluation of student test scores. In this analysis we estimate the effects of a series of family and school characteristics on test scores in Latin America and compare them to their estimated effects in Scandinavia.There are many existing cross-country analyses of the effect of family and school characteristics on international test scores, which find that both types of characteristics affect student scores. But these analyses invariably assume that family and school characteristics have the same effect in every country (For example, [Woessmann, 2003]).