The topic of transnational partner choice has gained attention over the past decades, but most research focuses on the migrants
and their descendants in Europe. Very little is known about the mechanisms underlying marriage migration decisions in the
sending contexts. With reference to migration and partner choice theories, we develop hypotheses on (1) education-related
selectivity, (2) family and community network effects, and (3) the role of consanguineous and arranged marriages for family
formation migrations from Turkey. Drawing on data of the 2000 Families Study, we apply multivariate regression models which
consider (married) stayers and marriage migrants from the same regions of origin. Findings indicate the expected inverted
u-curved effect of education and strong positive influences of prior migrants in the family network. Previous marriage migrants
exert a stronger influence than other migrants in the family. Gender differences are low, but while family networks hardly
diminish the selective impact of education for women, we find a negative interaction effect for men. Unlike the family, community
network effects are low and inconclusive. Moreover, marriage migration is strongly associated with consanguineous marriage,
and less with marriage arrangement. The implications of these findings are discussed in the light of existing theoretical
and empirical research.