Estimating the Effects of the Duration and Timing of Poverty Experiences on Maltreated Infants' Developmental Outcomes

Background: That poverty affects children's development is well-established. Prior research has indicated that duration of poverty is a significant predictor of cognitive development and behavioral problems. Research on poverty experienced in the first year or two of life, so-called 'early poverty' has been mixed. Earlier studies indicate early poverty is a predictor of subsequent cognitive and academic problems, independent of the duration of poverty. More recent studies have not replicated these effects. A key limitation is that most studies have only academic or cognitively-related outcomes, such as school-readiness. None have simultaneously controlled for maltreatment, despite the high risk children in poverty experience for maltreatment. The aim of this project is to evaluate how duration and timing of poverty affect infants' developmental achievement at a five year follow-up and whether effects occur uniformly across all domains of development in the context of maltreatment.

Methods: Using data from the infant sub-sample of the National Survey of Child and Adolescent Well-Being, the first nationally-representative study of children involved with child welfare services, six standardized measures of development in the domains of cognitive and reasoning ability, language, adaptive behavior, and behavioral problems were evaluated as outcomes at 66 month follow-up. This yielded an achieved sample of 1,197 infants. Missing data was controlled for by using a full-information maximum likelihood estimator. All analyses were completed with weights and complex sampling design variables. Multivariate linear regression models were constructed that controlled for demographics and maltreatment experiences while poverty was dummy coded as early/not early and as a linear measure of total time in poverty using data from baseline and 18, 36, and 66 month follow-ups.

Results: Overall, participants spent a mean of 71% of their lives in poverty during the first five to six years of life. Multivariate linear regression modeling of developmental scores found no support for the hypothesis that poverty experienced in the first two years of life (early poverty) having a uniquely detrimental effect compared to total time in poverty. The exception was the Woodcock-Johnson test of applied problems (cognitive/reasoning domain), where early poverty was associated with a modest 3.6 point drop in standardized scores. Time in poverty was a significant predictor of Pre-School Language screener (9.4 standardized point drop) and Woodcock-Johnson test of passage comprehension (6.0 standardized point drop). Both instruments evaluated the language domain. Based on the patterns of significance across all six models, development clearly did not occur in parallel nor were all domains of uniformly affected.

Implications: Poverty appears to most directly and significantly affect language development. This is of concern to child welfare and education practitioners and policy-makers because problems in language development tend to predict subsequent academic problems. Little support was found for the hypothesis that early poverty is particularly detrimental. Overall, findings vary significantly from prior research. This may suggest that some effects previously attributed to poverty may be related to maltreatment. Further research is necessary to identify which developmental risk factors are affecting development in this high-risk group.