Students' competencies are influenced by a host of factors, including the individual school's effectiveness. Measuring this contribution is extremely difficult. One way of circumventing the problem is by focusing on changes in competencies over time, i.e. value-added measures. Using the results of an INVALSI survey of high schools, this paper implements these measures for Italy, in an attempt to identify a general pattern of value-added among schools. Purging the sample of measurement errors - which require the exclusion of schools with too few students tested - and taking into account the selection bias implied by the non-compulsory nature of schools' participation in the survey, we find that the positive gap in favour of general programs (licei) when looking at the level of competencies tends to vanish (in maths and science) when focusing on value-added measures. By contrast, in the maths and science field schools located in the Southern regions are characterized not only by a lower starting level of competencies but also by a lower value-added. For maths at least, there is also a general tendency for teachers' turnover to have a negative effect on student improvements.
Published in 2010 in: Giornale degli economisti e annali di economia, v. 69, 2, pp. 81-114
