1. Deborah Cobb-Clark (U Melbourne) Mathias Sinning (ANU) Steven Stillman (U Otago)

2. What Questions Do We Try to Answer? 1. How do reading, math and science skills for 15- year-old migrant children compare to those of native children across OECD countries? 2. How does this vary with the child’s age at arrival and the language spoken in their home? 3. To what extent do differences in parental education and socioeconomic status explain migrant-native differences in test scores? 4. Do different institutional arrangements mitigate or exacerbate differences in achievement?

3. Data: PISA 2009 OECD Programme for International Student Assessment (PISA) is an internationally standardised assessment administered to 15-year-olds in schools. Surveyed between 4,500 and 10,000 15-year-old students from at least 150 different schools in 67 countries. Each student completed an assessment covering reading literacy, mathematical literacy, and scientific literacy, with the 2009 focus on reading literacy. Students also completed a survey questionnaire Principals completed a questionnaire about the school Nearly identical data were collected in all OECD countries.

4. Data: Defining Migration Status Native-born: those born in the test country who have no foreign-born parents First-generation migrant: those not born in the test country who have at least one foreign-born parent Second-generation migrant: those born in the test country who have at least one foreign-born parent Drop small number of children w/o enough info to classify or who are foreign-born but have no foreign-born parents First-generation migrants further classified into 3 groups based on their age at arrival: up to 4, 5–10, and 11–15 Migrants are further stratified based on whether the primary language spoken at home is the same as the language in which the test was given

5. Nativity Distribution by Country

6. Mean Reading Score by Nativity and Country

7. Baseline Regression Model Pooled cross-country regression T isc = standardised reading, math or science score of student i in school s of country c A isc = age in months and gender E isc = highest parent education and both foreign-born X isc = a vector of household socioeconomic status vars α isc = country fixed effects

8. Overall Migrant Reading Score Gap in OECD Countries

9. Full Regression Model Pooled cross-country regression I c = variables describing immigration policy and the education system in each country We interact the full set of country-level variables with our population indicators (M), with native-born youth as the omitted category Hence, the coefficients in the vector β5 are interpreted as the differential impact that each country-level characteristic has on test scores for different migrant students relative to the impact each characteristic has on test scores for native-born students.

10. Relationship between Country-Level Policies and Migrant Reading Test Score Gap in OECD Countries

11. Conclusions (1) Achievement gaps are wider for those migrant youths who arrive at older ages and for those who do not speak the test language at home. Differences in parental education and SES explain the worse performance for migrant children, except those who migrated at older ages and do not speak the test language at home. Educational systems do not work equally well for native-born and migrant students, or indeed for all groups of migrant students.

12. Conclusions (2) For example, Earlier school starting ages appear to reduce the relative achievement gap for some migrant students while leaving unaffected or exacerbating the achievement gap of others. Other arrangements, such as tracking on ability, are beneficial for migrant students when implemented in a limited way, but become detrimental when implemented across the board. Finally, what works for native-born students does not always work for students with a migration background. In particular, migrant students’ achievement relative to their native-born peers falls as proportionately more funding is devoted to educational spending generally and teachers’ salaries in particular, but improves when examination is a component of the process for evaluating teachers.