Higher education dropout among migrant students in Australia: the role of cultural and linguistic proximity

About one-third of Australia’s population is foreign-born. Yet, evidence on how migrant background shapes higher education persistence remains limited by crude classifications. Existing research typically focuses on Non-English Speaking Background (NESB) students but does not differentiate among students from specific countries. Segmented assimilation theory predicts that immigrant students’ educational trajectories vary systematically by country of origin rather than following a single, uniform path. How dropout risk varies across individual origin countries—and why—remains largely unknown in the Australian context. We examine this question among 1.5-generation migrants, students born abroad who arrived in Australia as children, using linked administrative data on the entire population of domestic students commencing undergraduate studies at Australian universities between 2012 and 2015 (n=613,647). We document dropout differences by world region and by individual country of birth, covering 99 origin countries, and test whether cultural and linguistic proximity to Australia accounts for these differences.

Compared with Australian-born students, most 1.5-generation migrant groups exhibit significantly lower dropout rates. However, dropout rates vary considerably within regions. Contrary to what student departure theory suggests, greater cultural and linguistic distance from Australia is associated with a lower, not higher, risk of dropout: students from the most culturally distant countries have a 9.6% predicted probability of dropping out, compared with 13.4% for the culturally closest, and those speaking the most linguistically distant languages at home have a 9.3% probability, compared with 12.2% for English speakers. The immigrant paradox in our data thus strengthens with distance.

Our study also demonstrates the value of large-scale data for uncovering diversity within migrant populations that broader classifications obscure. From a policy perspective, our findings suggest that support policies targeted solely on the basis of migrant or non-English-speaking background may be less efficient than policies directed toward specific disadvantaged subgroups.