Josep Oriol Escardíbul, Professor of Applied Economics,
University of Barcelona
The difference in PISA 2012 results between immigrant and non-immigrant students, in the latter’s favour, is only partly due to their immigrant status. Personal, family and school-related factors, and student performance, are related differently with the scores of immigrants and those of non-immigrant students. In turn, performance, especially among non-immigrant students, decreases when the proportion of immigrants at the different education centres exceeds 30%, whereas a lower percentage does not produce this effect.
1. Acquisition of competencies
The Spanish education system has experienced a steady and very significant growth in immigrant students since the end of the 1990s. As can be seen in graph 1, Spain has gone from having practically no immigrants in the classroom to approximately 10% of the student population being of immigrant origin.
In this article we ask what educational results do immigrant students obtain (in comparison with their non-immigrant classmates), and what are the factors that explain these results. These questions are concretized in four objectives. First, we describe the differences in skills levels between students of immigrant origin and non-immigrant students. Secondly, we establish the extent to which these differences are explained by socioeconomic factors or by other factors linked to the condition of being an immigrant. Thirdly, we look at whether the factors related to the acquisition of skills are the same for non-immigrant students and those of immigrant origin. Lastly, we examine whether the concentration of immigrants in schools has a negative impact on the acquisition of skills.
2. The educational performance of non-immigrant students and those of immigrant origin
Given the significant proportion of students of immigrant origin in Spain, it is important to know if their educational results differ from non-immigrant students. To do this we have used data from The Programme for International Student Assessment (PISA) of the OECD, which every three years since 2000, has been evaluating students' competencies (in reading comprehension, mathematics and sciences) every three years. In the last wave of results available, from 2012, 25,000 Spanish students from more than 900 schools were evaluated. The PISA test permits us to know the competency levels of students of 15 years of age along with certain individual characteristics, allowing us to distinguish between non-immigrant students and those of immigrant origin. The latter group can be divided into those born outside of Spain of immigrant parents (referred to as first-generation immigrants), and those born in Spain but of immigrant parents (second-generation immigrants). In 2012, the average score for OECD countries was 494 points in mathematics (in Spain the average was 484), 496 in reading comprehension (488 in Spain) and 501 in sciences (496 in Spain).
In table 1 we see the average scores for both immigrant and non-immigrant students for the three competencies evaluated by PISA 2012. The figures indicate significant differences in the results between non-immigrant and immigrant students. However, in the latter group, we find the results for second-generation immigrants (those born in Spain) are closer to the scores of non-immigrant students than first-generation immigrants. This is particularly the case in mathematics and sciences. Regarding reading comprehension, the scores of first and second generation immigrants are essentially the same.
With further analysis of the above data we can look at the distribution of the results and, more specifically, the proportion of students with very low scores. In this regard, the European Union has set a target for the year 2020, that the percentage of students scoring below proficiency level 2 (on a scale of 1 to 6) on the PISA test should not exceed 15% (see the European Commission's strategy framework, Education and Training 2020). In the case of first-generation immigrants, the percentage of students with low scores is practically double the European target of 15%; in the case of mathematics it is almost triple. These percentages are lower for second-generation immigrants, while for non-immigrant students, the EU targets are already (almost) being met in sciences and reading comprehension, with only scores in mathematics being clearly above the target.
Regarding the results of this evaluation of students' competencies, having been born in Spain seems to be important, particularly with respect to scores in mathematics and science. In the case of students only educated in Spain, these results are indicative of the capacity of the education system to compensate for differences that can originate in the family (although it is clear that some students born outside the country may have initiated schooling when they were already in Spain). The difference in school performance between first- and second- generation immigrants is common and, in addition to the factor mentioned, is explained by the fact that second-generation immigrants do not directly face the obstacles of migration and the difficulties of adaptation to new contexts and a new language (Jensen and Rasmussen, 2011).
3. Immigration and educational performance
In this section we analyse what part of the differences we find between first- and second-generation immigrants is due to the individual and family characteristics of the students (age, sex and socioeconomic and cultural environment of the household, for example), the school (material and professional resources, as well as organizational aspects and interactions with classmates or peer effects) and factors related to educational policies (the degree of comprehensiveness of the system, the existence of external assessments, etc.). We have used a type of analysis that permits us to isolate the effects of each variable on educational performance, independently of the effect of the other variables, and to determine what differences among groups of students are not due to any of these variables and, therefore, can be attributed to the very condition of being an immigrant. Following this strategy, we analyse the relationship of the different factors considered with the scores from the test on mathematical competencies (which was evaluated in depth in PISA 2012).
As can be seen in table 1, non-immigrant students score 55 points higher in mathematics than first-generation immigrant students. However, when we take into account the personal, family and school characteristics included in our model, we see that of this 55 point difference, 35 are explained by these factors (particularly family-related factors, such as socioeconomic origin and culture in the home). Therefore, immigrant status only explains 17 of the 55 point difference between non-immigrant and first-generation immigrant students (graph 3).
"Immigrant status", in our approximation, is composed of diverse elements that we cannot precisely measure or differentiate, but based on the data are only present in the case of students of immigrant origin. Some of these elements may be related to psychological traits that shape a determined orientation on the part of the immigrant student or his/her family toward the educational institution in the receiving country. They may also result (at least potentially) in discriminatory treatment (conscious or unconscious) on the part of the teaching staff and other students, or in a tendency for immigrant students to relate with other immigrant students, and, therefore, so-called "peer effects" for immigrant students may differ from those of non-immigrant students.
In the case of second-generation immigrants, starting from a "gross" difference in scores of 34 points less than non-immigrant students, we arrive at a "net" difference (subtracting the effect of the factors we have considered) of 13 points. These results show that second-generation immigrant students have fewer differences with non-immigrant students than first-generation immigrant students. This is not unexpected given what we have explained previously; however, it is noteworthy that once we have discounted the effect of the series of factors considered, the variable related to immigrant status has a similar impact on both groups of immigrants.
4. Factors with a differential effect on the mathematical competencies of non-immigrant and immigrant students
In this section we develop a similar analysis as in the previous section, although, in this case we look at non-immigrant and immigrant students separately (this is a common practice in international studies; for example, Dronkers and Van der Velden, 2013). The objective of the study is to analyse, for the Spanish case, whether individual, family and school-related factors are associated in a different way with students' results for each group considered.
In table 2 we see some of the relevant characteristics of the students participating in the PISA study. As can be seen in the table, there are practically the same proportion of males and females in each group of students. However, there is a significantly lower proportion of students of immigrant origin that were enrolled in early childhood education for more than one year than among non-immigrant students (65.5% versus 88.2%, respectively). In addition, the percentage of immigrant students that have repeated a year of school (54.9%) is much higher than that found for non-immigrant students (30%), as is the percentage with high rates of absenteeism (37% versus 26.9%, respectively).
Regarding the family, we find the level of education of the parents of non-immigrant and immigrant students is very similar (approximately 11 years of schooling). In contrast, the availability of books in the home, interpreted as an indirect indicator of family cultural resources, reveals a disadvantage for immigrant students: only 14.5% live in homes with more than 100 books, while the figure for non-immigrant students is 47.5%. In terms of school characteristics, immigrant students attend schools with a much higher proportion of the student body also being immigrants - 24.3%, in comparison to only 8.1% for non-immigrant students. Also higher is the percentage of immigrant students (8-point difference) in schools described by their administration as suffering discipline problems. However, the difference in the years of schooling of the parents is very small, and the student-teacher ratio is slightly favourable for students of immigrant origin (11 to 1) in comparison to non-immigrant students (13 to 1).
In short, with the exception of parents' educational level and student-teacher ratios, students of immigrant origin share certain characteristics that place them in a worse position than their non-immigrant classmates in terms of reaching the same educational outcomes.
When we analyse the association among different factors and mathematical competencies, the results show that the majority of the individual variables impact in the same direction for both non-immigrant and immigrant students. Thus, being a female is negatively related to scores for both groups. Also in both groups, students that repeated a year of school, have problems with absenteeism and that began to use information and communication technologies late tend to have lower scores. However, we find a positive relationship between having attended pre-school for more than one year and the acquisition of competencies (although only in non-immigrant students). For immigrant students, the time residing in Spain is positively correlated to PISA scores (a result also found in Zinovyeva et al., 2014).
Regarding family-related variables, the socioeconomic and cultural level of the household (defined through a single indicator) is clearly and positively related to results for both groups of students. However, if we breakdown this indicator into its various components (years of schooling of parents, their employment and occupational status, as well as the educational and cultural resources available to the student in the home), only this latter factor has a positive impact on the results for all students; in addition, father's occupation and a greater level of home educational resources available are also favourably associated with results, although only for non-immigrant students.
Regarding school-related variables, the lack of significance of the variables related to type of school (public or private) in the analysis for both groups of students stands out. Thus, the results for independent private schools and publicly subsidised private schools are not significantly different from those found in state schools, once we control for other significant variables.
In short, the empirical evidence shows that the majority of variables are associated in a similar manner with the results for both immigrant and non-immigrant students. Among the differences that we have found that could have an impact on education policy, are the particular sensitivity of immigrant students to the variation in the average number of years of parents' schooling and that attending pre-school for more than one year only has positive effects on the performance of non-immigrant students. In the first case it appears that the social capital of classmates (measured by the years of parents' schooling) is more important among immigrant students; in the second case, the lack of a relationship of pre-school education and results for immigrant students could be explained by the characteristics of the pre-schools they attended (Rovira et al., 2013).
5. Concentration of immigrants in schools and the acquisition of competencies
An element associated with immigration that some studies include as an influential factor on educational performance is the concentration of immigrant students (or ethnic minorities) in specific schools. Studies reveal uneven effects, although the majority of analyses conclude that the concentration of immigrant students has negative effects, particularly for non-immigrant students (Jensen and Rasmussen, 2011). The positive cases correspond to countries that attract immigrants with higher levels of qualifications, a factor that Schnepf (2007) defines as "immigrant capital".
In the case of Spain, a series of studies using PISA data have analysed the effect of the concentration of immigrants on students' scores, establishing diverse thresholds to see if there exists a differential effect of this concentration (the ultimate object of study). Graph 4 shows that the concentration of immigrants in schools is negatively associated with the scores of non-immigrant students in tests of mathematical competencies when the levels of concentration are high (at least 30% and especially at 40% or more), and with those of immigrant students (although in this case, only if the concentration is above 40%). The empirical evidence shows that, in Spain, the concentration of immigrants is negatively associated with students' results, particularly non-immigrant students, although the threshold for concentration to have an impact has increased over time.
6. Immigrants and skills: what matters most?
Our objective in this article has been to examine the impact of immigration on the skills acquisition. We have done this through four differentiated approaches.
In the first, we have quantified the difference in scores on the PISA test between non-immigrant students and students of immigrant origin. In the second, we have seen how the lowest results among the latter group are to a great extent related to socioeconomic and cultural origin and the specific conditions of schooling, which places them in a disadvantaged starting position in comparison to non-immigrant students. However, our analysis reveals that not all of the difference is explained by the variables in the model, leaving a proportion of the score directly related to immigrant status. Specifically, 17 points on the PISA test in the case of first generation immigrants (almost one third of the total difference with non-immigrant students) and 13 points in the case of second-generation immigrants (accounting for 38% of the difference).
In our third approach we have examined up to what point the factors that are related to the acquisition of competencies are different for non-immigrant and immigrant students. The results show that the factors that impact on their education are similar. In the education sphere, the main differences are the particular sensitivity of immigrant students to the educational level of their parents and the fact that pre-school attendance is only related positively with results among non-immigrant students.
In the last of our approaches we show how the concentration of immigrant students in schools negatively affects the acquisition of competencies. However, this association occurs at a certain threshold, which in 2012 stood at approximately 30% of the student body was of immigrant status for non-immigrant students and 40% for immigrants. This threshold has increased in recent years, which suggests reasons for optimism in light of the gradual improvement in processes of integrating immigrant students in Spain's education system.
The results of our research lead us to suggest certain policy actions. We would advise a more balanced distribution of immigrant students among schools to improve the performance of both non-immigrant and immigrant students. In addition, compensatory educational measures should be initiated as soon as possible, particularly among immigrant students, to correct deficits in the educational quality of pre-school education.
Jorge Calero, Chair Professor of Applied Economics
Josep Oriol Escardíbul, Professor of Applied Economics
University of Barcelona
Aparicio, R., and A. Portes (2014): Crecer en España. La integración de los hijos de inmigrantes. Barcelona: Obra Social ”La Caixa”.
— and J.O. Escardíbul (2014): Recursos escolares y resultados de la educación, Madrid: Fundación Europea Sociedad y Educación.
— and J.O. Escardíbul (2013): El rendimiento del alumnado de origen inmigrante en PISA-2012. PISA 2012. Informe español. Volumen II. Análisis secundario, Madrid: Ministerio de Educación, Cultura y Deporte–Instituto Nacional de Evaluación Educativa.
Calero, J. and J.O. Escardíbul (2014): Recursos escolares y resultados de la educación. Madrid. Fundación Europea Sociedad y Educación.
Dronkers, J., and R. Van der Velden (2013): «Positive but also negative effects of ethnic diversity in schools on educational achievement? An empirical test with cross-national PISA-data», in M. Windzio: Integration and inequality in educational institutions, London: Springer
Jensen, P., and A.W. Rasmussen (2011): «The effect of immigrant concentration in schools on native and immigrant children’s reading and math skills», Economics of Education Review, 30.
MECD (2013): Datos y cifras curso escolar 2013/2014, Madrid: Ministerio de Educación, Cultura y Deporte.
Rovira, M., E. Saurí and X. Bonal (2013): Rendibilitat social dels serveis públics educatius. Una proposta d’anàlisi per a Sant Feliu de Llobregat, Institut de Ciències Polítiques i Socials, ICPS-Working Paper 319.
Schnepf, S.V. (2007): «Immigrants’ educational disadvantage: an examination across ten countries and three surveys», Journal of Population Economics, 20(3).
Zinovyeva, N., F. Felgueroso and P. Vázquez (2014): «Immigration and students’ achievement in Spain: evidence from PISA», SERIEs, 5(1).
Europe, a land of hospitality for all?
Do we show the same solidarity to all refugees? In this series of seminars, co-organised with the Catalan Sociology Association, we will analyse the role of Europe as a land of hospitality.
Inequality of opportunity in educational performance in Spain and Europe
What lies behind educational inequalities? Factors beyond students’ control (such as gender, background, or parents’ financial or cultural status) explain 32% of the differences in their academic performance.
Evolution of wage gap between native-born and migrant youth in Spain
In what way does our background affect wages? This report analyses the evolution of wage trajectories among native and immigrant young people between 2007 and 2015.
Vocational Training in Catalonia: academic performance, dual VET, and gender
While women continue to opt for more socially-oriented courses, this report points out that the occupations most necessary in 2030 will be those linked to the digital economy.
Population changes can occur due to variations in the population caused by natural changes and migratory movements.