摘要：Objectives. We examined whether food insecurity was different for children in cohabiting or repartnered families versus those in single-mother or married-parent (biological) families. Methods. We compared probabilities of child food insecurity (CFI) across different family structures in 4 national data sets: the Early Childhood Longitudinal Study—Birth Cohort (ECLS-B), the Fragile Families and Child Wellbeing Study (FFCWS), the Early Childhood Longitudinal Study—Kindergarten Cohort (ECLS-K), and the Panel Study of Income Dynamics—Child Development Supplement (PSID-CDS). Results. Unadjusted probabilities of CFI in cohabiting or repartnered families were generally higher than in married-biological-parent families and often statistically indistinguishable from those of single-mother families. However, after adjustment for sociodemographic factors, most differences between family types were attenuated and most were no longer statistically significant. Conclusions. Although children whose biological parents are cohabiting or whose biological mothers have repartnered have risks for food insecurity comparable to those in single-mother families, the probability of CFI does not differ by family structure when household income, family size, and maternal race, ethnicity, education, and age were held at mean levels. In 2012, 10% of US households had food-insecure children, meaning that access to adequate food for these children was limited by their households’ lack of money and other resources. 1 Food insecurity poses a serious risk to the health and well-being of children; it has been linked to behavioral problems, developmental risk, poor health in infants and toddlers, 2,3 and negative academic, social, and psychological outcomes in older children and adolescents. 4,5 Traditionally, households headed by single mothers have had the highest rates of child food insecurity (CFI) whereas married-couple households have had the lowest rates: 18.7 versus 6.3%, according to the most recent data from the United States Department of Agriculture (USDA). 1 However, federal reports do not provide data on CFI in households characterized by other family structures, which are of increasing prevalence and interest. The most common of these family structures is cohabitation. Today, one fifth of all children in the United States are born to cohabiting, but not married, parents. 6–8 There is also little information on CFI in repartnered families, where only 1 of the 2 adults heading the household is a biological parent of the children in the household. Although there are few consistent estimates of the prevalence of these types of families in the United States, US Census Bureau data suggest that between 10% and 20% of children currently live in repartnered families and that more than one third of children will experience this type of living arrangement. 9,10 National reports do not provide estimates of CFI for this group; rather, families in which 1 biological parent has remarried are currently grouped with families in which the biological parents of the child are married to each other. 1 There is good reason to believe that the prevalence of CFI in cohabiting or repartnered families may be very different from its prevalence in married-biological-parent families. Most studies find that cohabiting unions are less stable and that these families have fewer resources than married-parent families, 11–13 although findings on child well-being in cohabitating families are mixed. Regarding repartnered families, new partners may contribute resources, thereby improving food security, 14 but previous research suggests that stepparents may underinvest in nonbiological children, because they may be providing resources to their prior biological children in other households or because they are less committed to nonbiological children. 15–18 Additionally, the instability that often accompanies repartnering may be harmful for a child’s well-being. 19,20 Economic models for the dynamics of food insecurity 21,22 suggest that decisions about food consumption are driven in part by families’ past and future resources and their ability to maintain consistent consumption over time, implying that stability and consistency may be as important for children’s food security as absolute level of resources. Thus, although single-mother families may have the fewest resources, they may not necessarily have a higher risk of food insecurity than these other nontraditional family types (cohabiting parents and repartnered parents), because of the potential instability of these family structures. A handful of previous studies have examined food insecurity across different family structures; however, these studies are dated and have relied on limited measures of food insecurity. 11,23–26 The USDA’s 18-item food security module (FSM) is considered the best measure of household food security. Previous studies, however, have often used measures of food insecurity based on 3 or fewer questions, making them of questionable validity. As a consequence, most studies have not been able to identify CFI, which involves limited access to adequate food specifically among children. Separately examining CFI is important, as parents often act to protect children from food insecurity by reducing their own food intake, 1 implying that general household measures could indicate food insecurity when children themselves may not be food insecure. One recent study of family change, which used the full 18-item FSM, found that transition into a maternal union was associated with lower household food insecurity. However, this study did not investigate CFI, nor did it report on rates of food insecurity by different family structures. 27 We investigated 2 complementary research questions: (1) How do rates of CFI for children in cohabiting and repartnered homes compare with those for children living with married biological parents or single mothers? (2) Do any differences in the rates of food insecurity among children in different family structures persist after adjustment for sociodemographic factors typically associated with both family structure and food insecurity? Our study makes a number of concrete contributions. First, we used a highly reliable and valid measure of CFI: the 8 child-referenced items from the USDA’s FSM. Second, our first research question was used to generate comprehensive and contemporary epidemiological evidence about potential differences (or similarities) in rates of CFI in different family types on the basis of analyses of 4 national data sets, an important contribution given limitations in current federal reporting. Last, our adjusted models (which examined differences in CFI between families that were average in all other regards) have the greatest potential to inform policies and programs that aim to eliminate CFI.
