Italy is characterized by a rich linguistic landscape. In addition to a multitude of dialects (Tuscan, Neapolitan, Sardinian, Sicilian etc.), in some areas there are several linguistic minorities: French in Valle d’Aosta, Slovene in Friuli-Venezia Giulia, German and Ladin in Trentino Alto Adige, Albanian and Croatian in South Italy.

According to the Council of Europe (1992) «“regional or minority languages” means languages that are: i) traditionally used within a given territory of a State by nationals of that State who form a group numerically smaller than the rest of the State’s population; and ii) different from the official language(s) of the State».

At international level, the European Charter for Regional and Minority Languages (ECRML) and the Framework Convention for the Protection of National Minorities (FCNM) promote the protection of national minorities’ languages spoken in the Member States of the Council of Europe also concerning education (Council of Europe, 1992; 1995).

In this context, it is necessary and useful to examine whether and to what extent achievement gaps among students of different language groups occur.

This essay attempts to analyse, through data deriving from National Institute for the Evaluation of the Education System (INVALSI), the academic performance of some linguistic minorities in Italy. The attention is focused on students belonging to Slovenian and German minority languages groups compared to Italian students in, respectively, Friuli-Venezia Giulia and the Autonomous Province of Bolzano, specific contexts of our country characterized by a multilingual social reality.

In order to investigate the skills and competences acquired, in Italy, every year, Italian and Mathematics standardized tests are administered to all four grades students (grades 2 and 5 of Primary School, grade 8 of Lower Secondary School and grade 10 of Upper Secondary School). Therefore, our databases refer to the whole students’ population for each grade involved.

For a correct comparison, we consider 2016-2017 results in Mathematics of the four grades investigated for the Slovenian students and only of the grade 8 and grade 10 for German language students (Slovenian schools administer a specific test in their mother tongue, while German language students take only Math tests).

The guiding research questions of this presentation are:

- are there any differences among Italian students and students belonging to minority languages schools?

- if so, which competences show more evidence?

- are there statistically significant factors that influence these differences?

- using INVALSI data series through a longitudinal or a cross-sectional analysis, are there the same significant factors?

As a preliminary analysis, descriptive statistics are conducted observing the average performances at the INVALSI tests assessment (Math) of Italian, Slovenian and German language schools. The variables considered are: - gender; - citizenship (an index based on the student-father-mother country of birth); - socio-economic-cultural indicator (ESCS); - language spoken at home (taken from the INVALSI student questionnaire); - type of school (Lyceums, Technical schools and Vocational schools). An analysis of variance (ANOVA) was carried on to confirm the significance of the difference found (Tabachnick e Fidell, 2013; de Smith, 2015). The following step is to observe if the emerged evidences are confirmed using data of the previous assessments (from 2014-15 to 2015-16), through a cross-sectional design. That is, we provide, for each school year, a data time series of differences between school test scores and the respective annual benchmark, based on the fact that INVALSI tests are not directly comparable across years (at least until last school year 2016-17). Finally, another step has a longitudinal design which allows us to carry on an analysis on the same cohort of students (the 2016-17 cohort), matching their test scores obtained at the INVALSI standardized tests administered in their previous career (Xian, 2015). This has been possible by using the SIDI code (‘Popcorn code’ for the German minority) which identify uniquely each student along his career: in our study this code can be used to associate grade 10 test scores in 2016-17 to grade 8 test scores in 2014-15.

According to the main hypotheses on which this research is based, performances depend on several aspects. Therefore, it’s possible to identify three dimensions: i) individual dimension, including gender – generally, males and females achieve different scores, for instance males perform higher than females in Financial literacy and vice versa in Reading (Contini et al., 2017; Dalla Villa et al., 2017; Ricci, 2017) – social background (Campodifiori et al., 2010) and language spoken at home – according to Rosa (2013) and Martini and Ricci (2009) lower performances are associated with a higher use of another language spoken at home (different to that one used at school); ii) school dimension, for example the type of school attended - as highlighted every year by INVALSI national Report (INVALSI, 2017); iii) dimension about characteristics of test: items are grouped by arguments (‘Statistics’, ‘Numbers’, ‘Shapes and figures’, ‘Relationships and functions’) and by cognitive process involved (‘Problem solving’, ‘Knowing’ and ‘Arguing’). The attention on this dimension is justified by evidences that confirm a connection between some of these aspects and a good command of linguistic competences.

Campodifiori E., Figura E., Papini M., Ricci R. (2010) Un indicatore di status socio-economico-culturale degli allievi della quinta primaria in Italia. INVALSI Working Paper n. 2. Contini D., Di Tommaso M.L., Mendolia S. (2017) The Gender Gap in Mathematics Achievement: Evidences from Italian data. Economics of Education Review, 58, pp.32-42. Council of Europe (1992) European Charter for Regional and Minority Languages. Retrieved November 2017, from: https://www.coe.int/en/web/conventions/full-list/-/conventions/rms/0900001680695175. Council of Europe (1995) Framework Convention for the Protection of National Minorities and Explanatory Report. Retrieved November 2017, from: https://rm.coe.int/16800c10cf. Dalla Villa V., Fiorini L., Russo F. La competenza in Lettura dei quindicenni. In Servizio provinciale di valutazione per l’istruzione e la formazione in lingua italiana (ed) (2017) PISA 2015. Risultati dell’Alto Adige. De Smith M.J. (2015), STATSREF: Statistical Analysis Handbook. The Winchelsea Press, Winchelsea, UK. Martini A., Ricci R. (2009) Effetti di variabili individuali e di variabili scolastiche sulla comprensione della lettura: analisi multilivello dei dati PISA 2006 dell’Alto Adige. In: siniscalco M.T., Meraner R. (eds.) (2009) Il livello di competenza dei quindicenni in scienze, lettura e matematica. Pisa 2006. Risultati dell’Alto Adige. Retrieved January 2018, from: http://www.schule.suedtirol.it/pi/themen/documents/pisa2006/PISA2006italiano.pdf. Ricci R. (2017) La competenza in Financial Literacy dei quindicenni. In Servizio provinciale di valutazione per l’istruzione e la formazione in lingua italiana (ed) (2017) PISA 2015. Risultati dell’Alto Adige. Rosa A. (2013) Il valore aggiunto come misura di efficacia scolastica. Un’indagine empirica nella scuola secondaria di I grado. Edizioni Nuova Cultura, Roma. Tabachnick B. G., Fidell L. S. (2013) Using Multivariate Statistics, 6th Edition. Pearson. Xian Liu (2015), Methods and Applications of Longitudinal Data Analysis. Academic Press.