When we write of ‘effects’ we acknowledge that some portion of the associations may involve
reciprocal causation.
Moreover, as noted above, selection effects regarding pre-divorce characteristics of parents and
children cannot be ruled out (Fomby and Cherlin 2007).
Longitudinal data would provide considerable analytical leverage on these issues. Lastly,
our results regarding the impact of parental role behaviour and parental conflict are
based on adolescent reports.
Many other studies examine wellbeing and family relationships of children as reported
by parents.