Association for Public Policy Analysis & Management

Numerous studies have linked children’s difficult temperament with increased behavioral problems at school entry. However, no prior studies have examined whether these associations differ by socioeconomic status (SES) such that the association is stronger for more socioeconomically disadvantaged children. If this difference exists, it would mean that children with difficult temperaments in infancy—already at risk for poorer developmental outcomes—would be at an even greater disadvantage if they were also socioeconomically disadvantaged. We posit that one reason difficult temperament might predict behavior problems more strongly for children in low SES families is if having a difficult infant undermines parenting quality among low SES parents more than higher SES parents, who have greater emotional, social, and financial resources with which to cope with parenting a difficult infant.

In this study, we address both questions, drawing data from the National Longitudinal Survey of Youth 1979 (NLSY), a nationally representative survey of youth ages 14 to 21 in 1979 who were re-interviewed regularly thereafter. Although numerous studies have linked difficult temperament with increased behavior problems, these studies have typically compared unrelated children who differ in temperament. We take a more causal approach by comparing the behavioral outcomes of siblings with different temperaments. These sibling fixed-effects models control for unobserved family characteristics that do not vary between siblings and that typically bias the association between temperament and child outcomes in OLS models.

Mothers reported on children’s temperament in the first interview following each child’s birth (prior to age two). Child behavior was assessed using the internalizing and externalizing subscales of the Behavior Problems Index (Zill & Peterson, 1986), taken when children were age 5 (mean=61.88 months). Maternal education, our measure of SES, was taken from the year of each child’s birth. Parenting was measured using the emotional support and cognitive stimulation subscales of the short form of the Home Observation for Measurement of the Environment scale. We also examine whether mothers reported having spanked the child in the past week. All parenting measures were assessed between 0 and 2 years old and again between 3 and 5 years old.

Within families, more difficult temperaments predicted greater internalizing and externalizing behavior problems (Table 1). These effects were stronger for children of less educated mothers in the domain of externalizing, but not internalizing, behavior problems (Table 2). Moreover, within families, parents were less cognitively stimulating with and were more likely to spank more difficult infants, but only during infancy (0 – 2 years). The effect on spanking, but not cognitive stimulation, was stronger for less educated mothers (not shown).

These findings provide the first piece of evidence that the effect of difficult temperament on externalizing behavior problems is stronger for low-SES children, likely placing them at an even greater risk for poor developmental outcomes than their peers; moreover, this stronger effect may be due in part to increased rates of spanking among low SES parents of difficult infants. Future analyses will test formally for mediation of the effect on externalizing behavior through maternal spanking.