Using Latent Transition Analysis to Test Bidirectional Effects Between Parenting and Children's Emotion Regulation

Maternal parenting practices are known to profoundly influence the development of children’s emotional regulation (ER) development across the first years of life. Although it is established that children’s development influences parenting, few studies have examined the bidirectionality between parenting and ER, particularly across time. Furthermore, no studies have utilized person-centered approaches to delineate bidirectional mechanisms between parenting and children’s development. The focus of present study is to demonstrate the use of a novel application of latent transition analysis to the study of bidirectional effects, specifically, those between profiles of parenting and children’s ER.

Observational data of maternal parenting behaviors captured during the Three-Bag Task were drawn from the Early Head Start Research and Evaluation Project. Four parenting behaviors, Supportiveness, Detachment, Intrusiveness and Negative Regard, were examined using latent profile analyses (akin to latent class analysis, with continuous indicators) across four time points to characterize the sample of 2,876 low-income mothers. Children’s emotion regulation was assessed by an observed with the ER subscale from the Bayley Behavior Rating Scale, and with the Leiter-R at pre-k. Latent profiles were estimated in Mplus 7.2. We subsequently modeled a Latent Transition Analysis (LTA), a longitudinal model with all classes across all time points. Indicators of children’s ER were added to the LTA as concurrent covariates, predictors, and outcomes.

The results indicated four well-differentiated latent profiles of maternal parenting style were classified at child age 14, 24-, 36-months and pre-k: Supportive (the majority at all three time points), Good enough (10-25% of mothers), Harsh, and Detached (each of which comprised of a minority of parents). The Supportive profile was predictive of the highest levels of ER across time, and that the Harsh profile predicted reduced levels of ER at all time points except form 36-months to pre-k. At all time points, the Detached profile predicted the lowest levels of ER. In terms of child effects, higher levels of ER at 14-months predicted movement from the Good Enough to the Supportive profile, but across 24-36 months, reduced levels of ER predicted movement from the Supportive to the Harsh and Good enough profiles. This study highlights the utility of LTA for studying bidirectional effects with person-centered designs. Collectively, the findings suggest that bidirectional effects vary across development, such that children’s ER impacts parenting most prominently in infancy and toddlerhood, and parenting’s influence on ER varies as a function of children’s regulatory abilities.