Adolescent Childbearing, Poverty, and Siblings: Taking New Direction From New Literature* This article critically appraises sibling-comparison method by pointing to following limitations: (1) it assumes within family homogeneity; (2) it neglects naturally occurring heterogeneity across siblings; (3) it overlooks unique effects of a teen's for her family and siblings; (4) it underappreciates selectivity of sisters of teenage mothers who delay childbearing; and (5) it ignores differences in outcomes by sisters' birth order and age spacing. Directions for future research stemming from this new literature are discussed. Key Words: adolescent childbearing, adolescent pregnancy, differential parenting, parental investment, poverty, siblings. One of most innovative lines of research to come out of adolescent literature is use of within-family estimates to gauge socioeconomic costs of teenage childbearing. This is often referred to as the new literature on teenage childbearing (Hoffman, 1998) and was fueled by a 1992 article by Arline Geronimus and Sanders Korenman entitled, The Socioeconomic Consequences of Teen Childbearing Reconsidered. In this article, Geronimus and Korenman compared socioeconomic indicators of sister pairs who differed in age at first childbearing, with one sister of each pair having given birth at age 19 or younger and other sister having had her first child at age 20 or older. Results showed that although teenage sister was less likely than her non-teenage sister to have any postsecondary schooling or to be married as an adult, family income and welfare dependency rates of both sisters were not significantly different and high school graduation rates were actually higher for teen mothers (although not statistically significant). The authors concluded that much of economic adversity associated with teenage motherhood stems not from itself but, rather, from disadvantage that preceded it. The results of Geronimus and Korenman ( 1992) study have been intensely debated in a series of articles by Saul Hoffman, Frank Furstenberg and colleagues (Furstenberg, 1991, 1992; Geronimus, 1991, 1992; Geronimus & Korenman, 1993b; Hoffman, Foster, & Furstenberg, 1993a, 1993b). Using similar procedures but analyzing a more recent version of one of data sets employed by Geronimus and Korenman, Hoffman and colleagues reported that accounting for unobserved family heterogeneity reduced but did not eliminate estimated consequences of early childbearing. Hoffman recently summarized this debate noting that case for weak or even positive effects of teenage is far from definitive (Hoffman, 1998). The sister comparison method is unique because it is first to attempt to disentangle independent costs of teenage from pre-existing family background factors that precede it. The use of kinship data, particularly sibling comparisons, as a means of reducing effects of heterogeneity on estimates of early is becoming increasingly common. For example, cousin comparisons have been used to gauge effects of young maternal age on children's socioemotional, cognitive, and health outcomes (Geronimus & Korenman, 1993a; Geronimus, Korenman, & Hillemeier, 1994; Rosenzweig & Wolpin, 1995). Results of these studies indicate that, when controlling for biases from family background, differences in outcomes of cousins (the children of sisters) were not overly significant or disappeared entirely. This article does not dispute specific findings reported by Geronimus and Korenman (1992) or Hoffman et al. (1993a). Rather, purpose of this article is to call attention to several limitations inherent in sister comparison approach. These limitations are significant because estimates that attempt to control for heterogeneity through use of sibling-based comparisons have become increasingly prevalent with little recognition paid to shortcomings of approach. …
Read full abstract