Linear and nonlinear associations between children’s sleep–wake patterns and harsh parenting.

= 298, 46% female, 63% White/European American, 37% Black/African American) and their parents from socioeconomically diverse backgrounds participated in two data collection waves (ages 10 and 11, respectively). Children's sleep was assessed at age 10 utilizing seven nights of actigraphy to derive measures of sleep duration (minutes), sleep quality (efficiency, activity), and chronobiology (sleep midpoint). Parents reported on harsh parenting at ages 10 and 11. Results from latent change score modeling supported a linear dose-response association between sleep activity and change in harsh parenting such that greater sleep activity predicted increases in harsh parenting, and lower sleep activity predicted declines or less steep increases in harsh parenting (β = 0.13). Sleep duration (β = -0.12) and sleep midpoint (β = -0.10) predicted change in harsh parenting in a nonlinear manner. Consistent with a threshold-acceleration model, shorter sleep duration and later sleep midpoint were weakly associated with increases in harsh parenting. After surpassing a critical threshold, beneficial associations emerged such that longer sleep duration and earlier sleep midpoint were progressively linked to declines in harsh parenting. Novel findings build on the literature by considering nonlinear predictors of parenting and applying nonlinear models of sleep to associations beyond developmental outcomes. (PsycInfo Database Record (c) 2026 APA, all rights reserved).