Long-Term Effects of an Intervention on Adolescents’ Emotional Insecurity: Modeling Patterns of Change Across Time

• In Event: Poster Session 13
  In Poster Session: PS 13 Topic - Prevention and Interventions

Abstract

Given that destructive interparental conflict poses a threat to adolescents’ emotional security, the Family Communication Project (FCP) was a four-week intervention designed to increase adolescents’ security in the interparental relationship by teaching adolescents and parents how to engage in constructive, rather than destructive, conflict (Cummings & Schatz, 2012; Davies et al., 2019). Although the intervention did not have a direct effect on emotional security at earlier follow-ups, recent findings indicated that being in the intervention conditions predicted decreased emotional insecurity at the three-year follow-up (Brickman, 2020). However, changes in emotional insecurity over time have not yet been examined. Looking to prior studies involving longitudinal evaluations of similarly-structured interventions, some studies have failed to identify any growth in outcomes over time; others have indicated linear change over time; still others identified curvilinear changes over time (see Hawkins et al., 2001; Slesnick et al., 2013; Trudeau et al., 2007). Therefore, the current study sought to answer the following questions: First, What is the appropriate change function to specify when examining the effect of FCP on emotional insecurity over time? Second, If emotional insecurity does change, do patterns of change differ between intervention conditions?

Participants were 225 parents with an adolescent between the ages of 11 and 17 (Mage at pretest = 13.23 years; 112 females). At five measurement occasions (pretest, posttest, 6-month follow-up, 1-year follow-up, and 3-year follow-up) adolescents reported their emotional security using the Security in the Interparental System scale, in which higher scores reflected higher insecurity (Davies et al., 2002). Families were randomly assigned to the parent-adolescent (PA; n = 75), parent-only (PO; n = 75), or control condition (C; n = 75).

For the first question, three nested multilevel models were fit using SAS PROC MIXED: a no-growth, a linear growth, and a quadratic growth model (see Table 1). Missing data was handled using FIML. Likelihood ratio tests indicated the quadratic model fit the data significantly better than the alternative models. The quadratic model explained about 19% of the within-person variability compared to the no-growth model, and an additional 5% of the within-person variability compared to the linear model. For the second question, the quadratic model results (see Table 2) indicated a significant difference between the PO and control in the estimated average difference in curvatures (change in rate of change), p = .045. Relative to the control, those in the PO group (on average) tended to initially decrease in emotional insecurity more steeply, at a decreasing rate, until reaching a trough, and then very gradually increasing in emotional insecurity again; the control group only exhibited a slight initial decrease before hitting a trough and increasing more rapidly again.

Although missing data and small effect sizes were among the limitations of this study, the current findings do provide evidence that (1) changes in emotional insecurity as a function of this intervention may not necessarily follow a linear pattern and (2) decreases in emotional insecurity may not necessarily be permanent. Future evaluations of FCP should more closely examine this possibility.

Authors

• Sarah Hoegler, University of Notre Dame
  Presenting Author

• E. Mark Cummings, University of Notre Dame
  Non-Presenting Author