Methodological Challenges of Including Children in Family Research: Measurement Equivalence, Selection Bias and Social Desirability

The growing acknowledgment of children’s agency has increased the use of child reports in family research. This study investigates three challenges that are related to the inclusion of children in surveys: 1) selection bias in the sample through parent and child refusal; 2) measurement equivalence of parent and child reports; and 3) the effect of parental presence during the interview on child reports. These challenges are investigated with the multi-actor data of the Divorce in Flanders study. The study shows that divorced parents more often refuse child participation than married parents. A parental refusal is also more likely in case of less frequent open communication with the child. The results further indicate that children answer as reliably as their parents on survey questions, but also that children report less frequent parental conflict, a less problematic communication with their mother, and a less open communication with mother and father, than their parents. This suggests that children’s perspectives on family relations are different from their parents. Therefore, including child reports in family research may enhance the understanding of family life. Finally, children tend to report slightly different on family relations when parents are present. It is concluded that processes of social desirability and selection are important to take into account when investigating children’s reports on their family life, as they can bias results and limit generalizability of the findings.