Squaring the circle: Is triangulation of child abuse reports helpful?

Child maltreatment measurement has been a longstanding issue, with discrepancies across administrative records, parent-reports, and self-reports. One proposed solution is "triangulation," or integrating data from multiple reporters and sources. However, it remains unclear how best to operationalize this concept.

This study examines the concept of "triangulation" by employing different analytic methods to determine whether these methods reveal a common underlying construct of physical abuse and whether they predict adult depression.

Data come from the Lehigh Longitudinal Study, a 40+ year prospective study that began in the 1970s with children ages 18 months to 6 years of age. Data were collected in early childhood, middle childhood, adolescence, and adulthood (ages 36 and 46, on average).

We applied five analytic approaches - network analysis, ordinary least squares (OLS) regression, structural equation modeling (SEM), latent profile analysis (LPA), and a cumulative index regression - to assess the relationships among multiple reporters of childhood physical abuse and adult depression.

SEM best modeled the latent construct of physical abuse and significantly predicted adult depression, with adult self-reports playing a particularly strong role. Network analysis also highlighted strong intercorrelations among self-reports and meaningful links with depression.

SEM and network analysis were the most informative for triangulation and prediction of adult depression. Adult self-reports of abuse were most related and most predictive of adult depression.