Dropout rates from higher education programmes are high and constitute a problem for both the individual and society as a whole. To effectively develop measures to combat dropout, the reasons why students drop out of higher education need to be understood. Building on Tinto's integration model, this paper tests the extent to which students' social and academic integration increases dropout intentions and whether the effect differs by students' academic family background. A sample of German students in their first year of university studies was presented hypothetical scenarios with varying degrees of academic and social integration (vignettes) and evaluated the likelihood of dropping out under the described conditions. This factorial survey design improves upon previous studies that were unable to separate the ambiguous causal ordering of the relationship. Subsequent multilevel analyses corroborate Tinto's integration model by revealing that all subdimensions of academic and social integration predict dropout intention, but also show that not all subdimensions are equally important. Cross-level interactions unveil that the effect of academic and social integration largely does not depend on students' academic family background.