the delinquency level of their children. Here, we notice a substitution effect from one type of externalising behaviour to another one. It suggests an association between
the different behaviours of parents (which cannot be tested with the present data),
relating to a more general, parental role or socialisation model.
However, the effects of family type on externalising behaviour do not decrease
when controlling for alcohol use of the parents.
Hence, the role model variables cannot be seen as real mediators between family type and delinquency or alcohol use.