Citizenship Status Transitions in the Migrant Life Course: A Typology
Swantje Falcke, Maastricht University
Maarten Vink, Maastricht University
However, other research has shown that the naturalization decision goes beyond this pure dichotomy and is more complex as immigrants vary on naturalization, migration and life course dimensions. Previous studies found that not only whether someone naturalizes but also when they do so, differs (Bloemraad, 2002, 2006). Furthermore, the legal ground varies on a country level (Vink, et al.,2013) as well as on the individual level where migrants may enter a destination country for study, work, family or asylum reasons. Additionally, some naturalized immigrants decide to stay, while others move onward to a new country (Ahrens et al., 2016). Lastly, life course events such as marriage or children have been found to affect naturalization propensities (Bevelander & Veenmann, 2006; Helgertz & Bevelander, 2016; Vink et al. 2013; Yang, 1994).
In other words, in the migrant life course, the question is not only whether someone naturalizes, but also when citizenship is acquired, whether this is done individually or as part of a family context, and also how citizenship affects the subsequent migration trajectory. These different possibilities illustrate that the naturalization decision goes beyond the initial dichotomous divide and call for a systematic account of how the differential outcomes, timing and context of naturalization decisions add up to different types or groups of immigrants.
In this paper, we analyze whether the different aspects of naturalization, migration and life course events add up to different types or groups of immigrant naturalization. We employ the life course approach which emphasizes that life course events are interlinked rather than analyze them as separate events (Giele & Elder, 1998).
Using Dutch population register data from 1995 to 2015, we apply latent class analysis (LCA) to identify typologies. LCA is a latent variable model which identifies unobservable groups within a population by linking response patters of observable variables (Collins & Lanza, 2010). LCA has been successfully implemented in migration research by Bean et al. (2011) for the United States, Rooyackers et al. (2014) for the Netherlands and Luthra et al. (2016) for Polish immigrants in Germany, the Netherlands, United Kingdom and Ireland. We use LCA in this paper to test whether the associated links between different aspects of naturalization, migration and life course can be explained by a latent class structure. The latent classes consist of migrants with similar combinations concerning citizenship acquisition and speed, migration motivation and continuity, as well as partner and children.
In a second step we aim to analyze the compositional variance across those types. We use the results of the latent class analysis to analyze which factors may predict class membership in the immigrant naturalization typologies. The covariates used are dual citizenship policy of the origin country, country development, possible naturalization before the Dutch citizenship reform in 2003, gender, age, and education.
As a third step we employ multi-group LCA and divide the group by cohorts as well as development of the origin country (Collins & Lanza, 2010). The revised Dutch nationality act from 2003 made it more difficult and costly to naturalize; leading to fewer and delayed naturalizations. The effect depends furthermore on the origin country for the migrants, namely whether they are from a developing country (Peters et al., 2016). The explanatory power of this variable may exceed group membership and affect the number of types as well as occurrence within a typology.
In this paper we make use of longitudinal Dutch population register data on first generation immigrants who arrived in the Netherlands between 1995 and 2000, which allows us to follow each cohort for 15 years. We make use of longitudinal data which – in contrast to many previous studies that make use of cross-sectional data – allows us to include both retrospective (what did you do?) and prospective (intentions) information. By using a longitudinal data set, we can track individuals over time.