Questioning children about family relations within a multi-actor perspective: selectivity bias and social desirability according to the participation and presence of parents

Sofie Vanassche, Katholieke Universiteit Leuven
Nele Havermans, Katholieke Universiteit Leuven
Koen Matthijs, Katholieke Universiteit Leuven

The focus of this study is on surveying children about the relationship with their parents and the level of parental conflict. Previously, most findings in family studies only reflect the vision of adults, whereas children were deemed as being incapable of formulating their own views and attitudes. This passive conception of the child has changed. The growing awareness of children’s agency has several methodological implications. Children are more susceptible to issues such as suggestive questioning, social desirability and interviewer effects. Furthermore, parents often need to give permission to interview the child which may result in a selective group of children. Moreover, parents’ presence during the child’s interview cannot always be prevented and this can increase the likelihood of social desirability. To study the methodological issues above, data from both parents and children on the same relationships are required. Multi-actor data are considered to be valuable for increasing the reliability and validity of relationship characteristic measurement. Multi-actor data allow to tackle problems such as shared method variance in studying the relationship between family relationships and individual wellbeing measures. In the present study, we compare the perspective of parents and children on these relationships, and explore different strategies to combine information from different actors. Moreover, we examine selection bias relating to multi-actor response rates (participating parent-child dyads and triads) and to the permission to participate granted by parents. We use the database Divorce in Flanders, containing information on 1168 marriages (intact or dissolved) with a randomly selected child between 10 and 17 years old. Preliminary results for the parent-child dyads show that children report lower scores with regard to the quality of the parent-child relationship and a higher conflict frequency between their parents. The likelihood of social desirability and item non-response increase when the parent is present during the interview.

  See paper

Presented in Session 43: Bias in demographic surveys