Longitudinal models that assess change in a behavior over time or relations between behaviors over time rest on the assumption that we are measuring the same underlying construct across multiple time points. The current talk and discussion will center around the challenges and implications of testing longitudinal measurement invariance within a Structural Equation Modeling framework, using harsh parenting and data from the Fragile Families and Child Wellbeing Study (N=4,898 families followed since birth) as an example. Discussion questions posed to the group will surround issues of (1) the importance of testing longitudinal measurement invariance within large-scale survey data that employ abbreviated measures of constructs, (2) "rules of thumb" to help researchers balance the quantitative rigor of measurement invariance with feasibility, and (3) whether formal tests of measurement invariance are more important for some longitudinal models of change (e.g., latent growth curves).

Arianna Gard is a 4th year Developmental Psychology PhD student. Her work examines biopsychosocial models of psychopathology, with a focus on the interplay between parenting, corticolimbic function, and genome-wide genetic liability. She currently collects data for the Michigan Twin Neurogenetic Study (PI: Luke Hyde) and the Adolescent Wellbeing and Brain Development Project (PI: Christopher Monk).