Ecological Factors Associated With Adolescent Pregnancy: A Review Of The Literature

Adolescence, Fall, 1999 by Jacqueline Corcoran

ABSTRACT

Adolescent pregnancy, the disproportionate number of births to unmarried adolescents, the potential disadvantages for both mothers and their children, and the commensurate costs to society have received the attention of researchers in a variety of disciplines. This article reviews and synthesizes the disparate literature on psychosocial factors associated with adolescent pregnancy using Bronfenbrenner's ecological model. Social influences within the macrosystem, mesosystem, and microsystem are examined. Policy and service delivery recommendations are offered.

Adolescents account for a third of all out-of-wedlock births per year nationwide (United States Department of Health and Human Services, 1995). Associated problems include the increased likelihood of deleterious consequences for both adolescent mothers and their children (e.g., McAnarney & Hendee, 1989) and the cost for society as a whole (Burt, 1986). Reflecting public concern, research on adolescent pregnancy prevention has been ongoing since the 1970s (see Chilman, 1980; PhippsYonas, 1980; Urberg, 1982, for reviews) across a variety of disciplines: child development, psychology, social work, sociology, family science, nursing, and medicine. Synthesizing these disparate approaches is a difficult task (Elster, McAnarney, & Lamb, 1983). However, Bronfenbrenner's (1979) ecological model can be employed to organize the salient contributing factors in adolescent pregnancy. Support for an ecological perspective has been provided indirectly in the early reviews of the literature (Chilman, 1980; Phipps-Yonas, 1980 ; Urberg, 1982) and by more recent multivariate analyses.

Adapted from the physical sciences to explain human behavior, "Bronfenbrenner's (1979) ecological model of human development... conceptualizes ecological space as operating on different levels of systems, each of which is incorporated within the next" (Franklin, 1988, p. 340). At the most basic interactional level is the microsystem, which is the pattern of activities, roles, and interpersonal relations experienced by the individual in a given setting. The mesosystem involves interactions among settings. The next level, the exosystem, includes settings that affect the individual but with which the individual does not interact directly (exosystem variables are not covered here, since there are so many other variables that have a more direct impact on adolescents). The macrosystem includes the cultural variables that influence the individual.

Bronfenbrenner's conceptualization of ecological systems has been used in theoretical formulations, as well as employed in empirical research. Variables have often been chosen to represent system levels depending on the social phenomenon under study, making it difficult to "test" the ecological model. Nevertheless, the model can be seen as a way to organize factors associated with complex social problems so that knowledge building can occur and intervention can be implemented at the appropriate system level.

In the present article, Bronfenbrenner's definitions and relevant literature guided decisions about the psychosocial variables belonging to each system. Studies involving a range of ecological factors were included, even those not typically identified with the ecological systems framework. Studies using a cognitive framework, in which adolescent sexual behavior is viewed from a decision-making perspective, were excluded (e.g., Jorgensen & Adams, 1988; Keith, McCreary, Collins, Smith, & Bernstein, 1991; Marsiglio, 1985; Nathanson & Becker, 1983; Rogel, Zuehlke, Peterson, Tobin-Richards, & Shelton, 1980; Urberg, 1982; White, 1984). Because historical context is an element of the macrosystem, studies published since 1980 were reviewed. Although a case could be built for using data on parental communication to determine which teens are likely to get pregnant, this was seen as an overly specific focus, and more general family functioning was included. Another criterion for exclusion of studies was a focus on sexu al knowledge and attitudes (e.g., Evans, 1987; Morrison, 1989; Nadelson, Notman, & Gillon, 1980), including tests of the effectiveness of sex education (e.g., Ku, Sonenstein, & Pleck, 1989). The studies also needed to have used primarily adolescent (age 19 and under) or high school samples, rather than college-age populations. Additionally, only those studies based in the United States were included. For instance, Oz and Fine's (1988) study in Ontario, Canada, was excluded because Hechtman (1989) reported that the rate of pregnancy and abortion among adolescents in the United States was over double that of adolescents in Canada.

A discussion of ecological factors in each system level follows. The literature is then critiqued, and implications for intervention are discussed. Table 1 presents study designs, samples, and key results.

MACROSYSTEM

At the macrosystem level, socioeconomic status (SES) has been measured several ways, including mother's educational level (e.g., Ralph, Lochman, & Thomas, 1984), father's educational level (e.g., Yamaguchi & Kandel, 1987), parental occupation (Ralph, Lochman, & Thomas, 1984), and various combinations of these. For instance, Robbins, Kaplan, and Martin (1985) relied on mother's education, father's education, father's occupation, and respondent's subjective rating of family social class in measuring SES. Researchers who have found SES to be significant in adolescent pregnancy include Barnett, Papini, and Gbur (1991) and Robbins et al. (1985).