The Effect of Attitudes on Teenage Premarital Pregnancy and Its

This study examines the influence of self-esteem, locus of control, and attitudes toward women's family roles and school on the probability of teenage premarital pregnancy and, given a pregnancy, whether it is resolved by abortion, having the birth premaritally, or marrying before the birth. The data are drawn from the National Longitudinal Survey of Youth and analyzed using the nested logit method. The evidence suggests that for both whites and blacks the four attitude variables are associated with premarital pregnancy and its resolution in the directions predicted by theory. The Effect of Attitudes on Teenage Premarital Pregnancy and Its Resolution Teenage premarital childbearing is associated with a number of negative consequences for both mother and child such as lower earnings and increased chances of poverty Pofferth, 1987) and leads to substantial public costs for income support, health care, and social services @urt with Levy, 1987). Because many observers believe that the adverse consequences are indeed caused by teenage premarital childbearing, and because teenage premarital childbearing violates mainstream norms, it emerged as a national concern during the 1980s and has remained the focus of heated debate. For these reasons and because the analysis of nonnormative behaviors has been of long-standing sociological interest, understanding the determinants of premarital childbearing can contribute to both public policy debate and social science. Many social, economic, and psychological factors are likely to influence adolescent women's sexual behavior and, consequently, their chances of becoming premaritally pregnant and an unmarried mother. That these factors include attitudes and other psychosocial variables is an intuitively plausible proposition, and one that can be deduced from behavioral theory. Psychosocial variables may mediate the effects of family background and other personal characteristics, exert independent influence on premarital childbearing, or both. Improved knowledge about the relationships among attitudes, family background and other personal characteristics, and premarital childbearing will increase our understanding of the determinants of sexual and marriage behavior and of the mechanisms through which background characteristics influence such behaviors. Using data from the National Longitudinal Survey of Youth (NLSY), this paper provides evidence on the relationship between teenage premarital childbearing and four psychosocial variables: self-esteem, locus of control, attitudes toward women's family roles, and attitudes toward school.' Recognizing that a premarital birth results from a premarital pregnancy followed by a decision to bear the child before marriage, the analysis examines the relationship of such variables to both premarital 2 pregnancy and its resolution. Along the way it explores plausible theoretical arguments why each of the four characteristics would tend to influence behaviors related to premarital childbearing and offers new evidence about the validity of models which predict that these attitudes affect behavior. One strength of this study is its use of the NLSY. In contrast to the demographically skewed samples used in most other studies of this issue, the NLSY is a nationally representative data base containing high-quality, longitudinal data in which the attitudes were measured prospectively for the sample analyzed in this study. Findings derived from it, therefore, will avoid some methodological limitations of many other studies which have examined the strength of these same relationships. Moreover, this analysis is innovative in that it uses an integrated empirical framework based on the nested logit statistical technique. The advantages of this approach will be explained later in the paper. CONCEPTUAL FRAMEWORK The Present Model Figure 1 displays the general model that informs the empirical analysis. The process leading to a premarital birth divides into two major stages: becoming premaritally pregnant and, given a pregnancy, its resolution by abortion, carrying to term and marrying before the birth, or carrying to term and becoming an unwed m ~ t h e r . ~ Attitudes as well as family background and personal characteristics directly affect the likelihood of premarital pregnancy, abortion, and marriage. Additionally, attitudes are partly determined by family background and personal characteristics. This relationship may be one of the mechanisms through which such characteristics transmit their effect on adolescent sexual and marriage beha~ior .~ The model in figure 1 may be further elaborated by distinguishing between the personal attributes of self-esteem and locus of control, and attitudes about family roles and school. An individual's sense of self-esteem and locus of control may be well established by early childhood. A