Examining Longitudinal Bidirectional Associations Between Parental Monitoring and Child Conduct Problems Using Autoregressive Latent Trajectory (ALT) Models

Introduction: A multitude of studies indicate that low parental monitoring is associated with child conduct problems. Consistent with transactional models (Pardini, 2008), recent studies have documented a bidirectional association between monitoring and conduct problems, as higher levels of monitoring predict fewer problem behaviors, and more problem behaviors predict lower parental monitoring in subsequent years (Laird et al., 2003; Williams & Steinberg, 2011). However, other studies have argued that these associations do not reflect bidirectionality, but instead are driven by the concurrent correlations and stability in these behaviors over time (Kiesner et al., 2009). Thus, the goal of the current study was to examine the longitudinal bidirectional relationship between parental monitoring and child and adolescent conduct problems.

Methods: Data for the current study (N = 753) were collected as part of the Fast Track project, a multisite longitudinal study aimed at the prevention of conduct problems. Half of the sample was Caucasian (49.7%) and 57.7% were male. In keeping with work by Stattin and Kerr (2000), parental monitoring was measured in grades 4-5 and 7-11 as child- and parent-reported solicitation and control from the Supervision Questionnaire (Loeber et al., 1998). Child-reported child conduct problems were assessed with the Things That You Have Done questionnaire in grades 4-5 and with the Self-Reported Delinquency measure in grades 7-11. Autoregressive latent trajectory models were conducted in Mplusversion 6.1 (Muthén & Muthén, 2010) to evaluate bidirectional associations between parental monitoring and child conduct problems. These models are particularly useful given that cross-lagged effects may be difficult to detect when stability in trajectories is high (Kiesner et al., 2009).

Results: Significant cross-lagged effects were identified for child-reported parental solicitation, indicating that higher levels of parental solicitation and child conduct problems in one year predicted lower levels of conduct problems and parental solicitation in the next year, respectively, over and above what would be expected based on the underlying trajectories. No cross-lagged effects were found for parental control.

Conclusions: Consistent with other studies (Laird et al., 2003; Pardini et al., 2008), these bidirectional findings may be interpreted as evidence that increased parental solicitation in one year deters children from conduct problems in the next year, or that parents disengage from solicitation as children engage in more conduct problems. These bidirectional associations suggest that short-term interventions to improve parental solicitation might be effective in deterring child conduct problems in the next year.