Black Teenage Pregnancy

This article examines the relative importance of race and socioeconomic status (SES) in determining whether Black and White teenagers report having ever been pregnant. Data gathered from 1999 to 2006 by the National Center for Health Statistics of the Center for Disease Control and Prevention included 1,580 Black and White females aged 15 to 19 years. Results supported the effects of race and SES, with SES having the stronger effect. However, the effects of race and SES differ when controlling for the state of the economy. No difference between Blacks and Whites was found during better economic times. During 2003-2004, the period of greatest economic stress, race was determined to be the only predictor of teenage pregnancy. In particular, during 2005-2006, the reduction in pregnancy rates for Black minors (15-17) fell below those for White minors within their respective SES categories. Policy implications are discussed in light of these findings.

Teen pregnancy specifically emerged as a term in the early 1960s but did not become popular until the 1970s (Furstenberg, 2007;Wong & Checkland, 1999). Scholars began studying teen mothers and noticed how their lives were affected by poverty, which resulted in targeting concerns toward preventing unwanted births (Campbell, 1968). The premise behind the institutionalization of teenage pregnancy as a social problem was the perceived devastating consequences. Teenage pregnancy cast as a social problem was believed to be the result of lower levels of education, welfare dependency, and low-paying jobs, as well as greater health troubles for these teens and their babies. 1 Alongside establishing teenage pregnancy as a deviant behavior and major public health problem was the concern over high rates of teenage pregnancy for Blacks.  and Furstenberg (2007) point out the unfair equivalence of teenage pregnancy with Black teen pregnancy. Geronimus (2003) even argues that publicizing teen pregnancy and the high rates of teen pregnancy among Blacks assists the dominant group (Whites) in disseminating social control by arguing that teenage childbearing has disastrous consequences.
Studies have also examined the effects of race and class on teenage pregnancy (South & Baumer, 2000). Others have examined the effect of state unemployment rates on Black and White teenage pregnancy over a period of economic growth (Colen, Geronimus, & Phipps, 2006). However, a dearth of literature is found for comparing the changing effects of race and class with differing economic states over time. This article looks at teenage pregnancy rates across differing economies for young Black and White women and distinguishes between pregnancy rates for minor teenagers (15-17) and those of adult teenagers (18)(19). 2 Pregnancy rates for dependents should define teenage pregnancy as a social problem more accurately than those for young, consenting adults. This investigation reveals two particulars. First, the ways in which race and class affect teenage pregnancy are complex and dynamic and respond to changes in the economy uniquely for different groups of teenagers. Second, equating the Black teenager with the problem of teenage pregnancy is a misrepresentation of reality in the 21st century.
The common public image of teen pregnancy as a serious social problem is counter intuitive to two facts. First, rates have been falling for over a decade for all teenagers and especially for minor Black teenagers (Martin et al., 2009), until recently when they have begun to fluctuate (Hamilton, Martin, & Ventura, 2010;Martin et al., 2009). Second, most births to pregnant teenare to adults. In addition, scholars have made convincing arguments that poverty is a cause rather than a result of teenage pregnancy (Furstenberg, 2007;Geronimus, 1991;Wilson, 1987), and any unfortunate outcomes after a teen pregnancy are small, if any, after isolating them from the effects of poverty. 3 Some of the outcomes explored in the literature are the health, cognitive development, and antisocial behavior of the children of teenage mothers. The education and economic circumstances of the teenage mothers and the use of public assistance by the teenage mothers are also explored (Geronimus, 2003). Wilson's The Truly Disadvantaged (1987) is the most commonly cited work in setting forth a sociological explanation for the very high rates of teenage pregnancy among Blacks. However, Ladner's (1971) prior study of Black female adolescents in a St. Louis housing project also provides insight into teenage pregnancy. Wilson provides an explanation that emphasizes structure while acknowledging the contribution of culture. Ladner, starting with interviews with poor Black adolescents, focuses on culture while acknowledging the contribution of structure. Wilson uses a structural framework to explain all forms of social disorganization including teenage pregnancy. Ladner uses a social constructionist framework to explain how Black girls become women. Becoming mothers, even as teenagers, is a salient part of this process. The two authors' viewpoints coincide greatly. In The Declining Significance of Race, Wilson (1978) contrasts the deteriorating conditions among the Black poor with improving conditions of the Black middle class. Therefore, Wilson identifies the more important discriminating factor as social class rather than race. Leigh (2004), in support of Wilson, finds the percentage of the Black population in poverty in selected cities to be related to the percentage of births to Blacks under the age of 20. Ladner (1971) acknowledges the role of poverty in shaping strategies of survival but places a greater emphasis on race than does Wilson. She argues that many cultural traits are products of American social policy and that "the structure of the American social system, through its practices of institutional racism, is designed to create the alleged 'pathology' of the community, to perpetuate the 'social disorganization' model of Black life" ("Introduction" [c. 1995] in Ladner, 1971, p. xxiv). Wilson (1987) provides a historical context for the state of affairs in the Black community. Prior to the Civil Rights Movement and the exodus of professional Blacks from the inner city, all Blacks lived in close proximity, regardless of social class, due to a history of slavery and racism. This common history provided a mechanism whereby race rather than social class was the basis for interaction in the Black community. Although there were lower and middle class sections of the ghetto, the Black middle class tended to serve the lower class within the Black community. Before the mid-1970s (postslavery), single-parent families were not a dominant form, and households were headed by widowed, separated, or divorced adult women rather than by unwed mothers. Welfare dependency was not the norm either. Rather, inner-city communities displayed social organization in the form of "a sense of community, positive neighborhood identification, and explicit norms and sanctions against aberrant behavior" (Wilson, 1987, p. 3).
After the Civil Rights Movement, poverty creates the context for social dislocation even in the presence of antidiscrimination and affirmative action programs (Wilson, 1987). Wilson finds noteworthy the increasing proportion of teen births in general and the more drastic increase among Black, unmarried teenagers. Even when teen pregnancy rates had declined, the proportion of births to unmarried teens increased. Changes in the American economy, which have created astonishing rates of Black unemployment and the exodus of the Black middle class, have contributed to increasing poverty in the inner city. In addition, Colen et al. (2006) provide a compelling argument of how Black teenage pregnancy may also respond to the economic context of the time. Ladner (1971) does not distinguish between different time periods but shows how the meaning of being a poor Black woman arises from poverty. She asserts that institutional racism contributes to poverty for Blacks.
While working on When Work Disappears, Wilson "came to appreciate even more the impact of joblessness on individuals, families, and neighborhoods" (Wilson, 2003(Wilson, , p. 1102. The cost of joblessness outweighs those of poverty because crime, family dissolution, welfare, low levels of social organization, and other problems in the inner-city ghetto neighborhoods are the outcome of the vanishing jobs (Wilson, 2003). Furthermore, few employment opportunities lead to unstable relationships and out-of-wedlock pregnancies and births (Wilson, 2003). Teenage pregnancy is one way poor Black females become women when facing few educational and occupational opportunities to prove one is an adult (Ladner, 1971). Blacks felt the benefits of prosperity in the 1990s, indicated by the reduction of their percentage residing in high-poverty areas (Wilson, 2009). The early 1990s is when the teenage pregnancy rates declined for all groups and at an even greater rate for Blacks. Still, the Black poor remain and the Black teenage pregnancy rates continue to exceed those for Whites.
Although a major contribution of Wilson to a theoretical discussion of teen pregnancy stems from macro-level processes, he also specifies how structure leads to micro-level outcomes. Wilson (2010) emphasizes that cultural traits originate from or are the products of both racial and nonracial structural forces. One might say that the structural factors contributing to poverty indirectly affect teen pregnancy through culture. Wilson (1991) attributes the structural problem of weak labor-force attachment to the cultural problem of the transmission of self and collective beliefs in the neighborhood. As a result people may seriously doubt that they can do or accomplish what is expected, or they may feel confident of their abilities but, nonetheless, give up trying because they believe that their efforts will ultimately be futile due to an environment that is unresponsive, discriminatory, or punitive (Wilson, 1991, p. 10).
This dynamic was present among the adolescents studied by Ladner (1971) as well.
The absence of the middle class eliminated the reinforcement of societal values and norms promoting upward mobility (Wilson, 1987). Later marriage or low rates of remarriage contribute to out-of-wedlock births and femaleheaded households, which Wilson (2003) links to the inability of Black males to support a family rather than to the availability of welfare. Furstenberg (2007) reveals how his Baltimore sample of minor teen mothers, mostly Blacks, claimed the fathers of their babies to be poor prospects for marriage. This theme was also present in Ladner's ("Introduction" [c. 1995] in Ladner, 1971) study. Wilson (1991) points out that persistent joblessness also eliminates the rational planning that accompanies the disciplines and regularities of work. One might apply a rational planning model to teen pregnancy. The decision to use birth control efficiently and consistently is an outcome of rational planning. Motherhood, even teen motherhood, may provide a legitimate alternative to a career in an environment that does not provide adequate educational and career opportunities. Ladner (1995) originally considered social problems such as pregnancy, premarital sex, school dropout, and so on a destructive adaptation. Later, she asserted, "Perhaps a very healthy and successful adaptation, given their limited resources, had been made by all of these girls to a set of very unhealthy environmental conditions" (Ladner, 1995, p. xxxii). The perception of this alternative may lead to a relaxed approach to the proper use of birth control, whether conscious or unconscious. An environment that creates doubt in one's economic sphere may also create doubt in other areas. For instance, regardless of self-perceived abilities, teenagers who feel no control over their reproductive life are not likely to behave in ways to control it. Ladner (1971) finds many Black adolescents to exhibit powerlessness over avoiding pregnancy. Some of these girls even stated that they would not use birth control. Young, Martin, Young, and Ting (2001) find girls who later became pregnant to be more likely to have an external locus of control, low personal efficacy, and lower educational aspirations than teens who did not become pregnant. Hellerstedt, Fee, and McNeely (2001) find future expectations about employment and education to be negatively associated with positive pregnancy feelings. Black adolescents who carried an unwanted pregnancy to term were found to possess the lowest future orientation (strong education and occupational goals) compared with a never-pregnant and abortion group (Freeman and Rickels, 1993). Much of the effect of neighborhood disadvantage may be linked to prochildbearing attitudes and behaviors of peers and to the teenager's own support for teenage premarital childbearing (Ladner, 1971;South & Baumer, 2000). Freeman and Rickels (1993) found little difference between the sexual behavior of Black adolescents in the three groups: never pregnant, abortion, and delivery adolescents, However, the never-pregnant group was almost 3 times as likely to use contraception with their first experience having intercourse. Manlove, Ryan, and Franzetta (2004) find Black adolescent girls to be less likely than White adolescent girls to have ever used contraception compared with never. Black adolescent girls were also found to be less likely than White adolescent girls to always use contraception compared with never or sometimes. However, these differences were not found to be statistically significant after controlling for the two biological parents, parental education, nature of the relationship, sexual history, religious service attendance, cognitive test score, and sex education.
Also consistent with a cultural argument, research suggests that sanctions against teenage childbearing are less severe in Black families and communities (Burton, 1990), and the public perception is that teenage pregnancy is widespread and acceptable in poor minority communities (Mollborn, 2010). Ladner (1971) finds mothers to disapprove of early pregnancy, but the peer group who played a more important role, and at an earlier age than is typical, contributed strongly to teenage pregnancy. Driscoll, Sugland, Manlove, and Papillo (2005) find high educational expectations to decrease the likelihood of White, Latina, and low socioeconomic status (SES) but not Black teenagers from having a baby. They also find Blacks to "be more optimistic than their actual opportunities would warrant," and they surmise that this is the reason that high education expectations were not found to protect low-SES Blacks from teenage pregnancy. Family resources were also found to account for the association between educational expectations and teenage birth for Blacks. No racial/ethnic differences were found in predicting a teenage birth net of other family factors within SES quartiles (Driscoll et al., 2005). South and Baumer (2000) find no link between having children and believing parenthood to interfere with academic ambitions. Ladner (1971) finds definite differences between girls who strived for upward mobility and those who did not. Education was considered key, and role models provided motivation. Upwardly mobile driven girls in Ladner's study were also motivated to avoid men believed to be poor prospects for marriage as well as early pregnancy. Ladner did not notice these girls coming from any one type of family and surmised that in some cases, individual traits contributed to the alternative strategies of upward mobility rather than giving in to a lifestyle of poverty. Mollborn (2010) finds Black adolescents to be about half as likely as Whites to feel embarrassment at the idea of teen pregnancy. Increased parental education and household income were found to increase the likelihood of embarrassment for adolescent girls. Mollborn's individual level analyses were found to garner more predictive power than neighborhood-level analyses. Ladner (1971) speaks of the lack of shame experienced by poor Black adolescents, and if shame was experienced, it did not last long. These adolescents also spoke about teenage pregnancy being common when asked about shame. Having a baby without adequate financial support from the father or other sources was considered a "mistake" and discouraged from being made again. Having sex and having a child moved one from being a girl to being a woman. Mothers may have an initial negative response but recognize that their daughters have now experienced what they have. They are now women to be supported and assisted concerning the pregnancy and the rearing of a child. Children are never stigmatized but valued.
In line with Wilson's work, race and SES are expected to be important precursors to teenage pregnancy with SES having a stronger impact. A large part of his theoretical justification for the effect of race had to do with the disproportionate number of poor Blacks compared with poor Whites. South and Baumer (2000) find the effect of being Black to disappear in models that included neighborhood variables. Arline Geronimus describes a dynamic that she calls "weathering" whereby low-SES Blacks having children at an earlier age is advantageous for lower income Blacks (Geronimus, 1992(Geronimus, , 1996(Geronimus, , 2001Geronimus, Bound, & Waidmann, 1999). The teenage years are optimal for childbearing for three reasons: first, because infant mortality rates are low for teenage mothers compared with mothers in their 20s; second, early childbearing reduces the risk of being widowed or orphaned; and third, early childbearing will result in a greater likelihood that extended family will be able to care for the young and aged. Health, of course, is a commodity better afforded by those with higher SES. Therefore, the following assertions are proposed: Hypothesis 1: Black teenagers have a greater likelihood of reporting having ever been pregnant than do White teenagers (the race effect). Hypothesis 2: Low-SES teenagers have a greater likelihood of reporting having ever been pregnant than do not-low-SES teenagers (the social class effect). Hypothesis 3: SES has a stronger effect than race on reporting having ever been pregnant.
Single-parent homes are a prominent form of Black lower income families primarily because of the unavailability of men who are able to support a family. Although single-parent homes have a relationship with social class and race, it is important to control for because of its possible independent effect. Having a single parent may assist in early sexual debut (Bonell et al., 2006) and early childbearing (Scott & Perry, 1990a;Scott & Perry, 1990b). The effect of being raised in a single-parent home or being the offspring of a teenage parent, although a well accepted precursor to teenage pregnancy, is probably an indirect one. That is, this phenomenon accompanies a variety of unobservable variables that may more directly affect the likelihood of teenage pregnancy. A singleparent home may provide the context for an environment ripe for the emergence of teenage pregnancy. Teasing out the unobservable variables is beyond the scope of this article; instead, single-parent home is a proxy for the variety of factors, which may or may not exist in any given case. 4

Hypothesis 4: Teenagers from single-parent homes
have a greater likelihood of reporting having ever been pregnant than do teenagers from two-parent homes.
In addition, whether a home is headed by a male or female may differ. Not only do men and women have access to different resources, but they also possess different parenting styles and influence on their offspring. Therefore, the female-headed household is also controlled for.
Hypothesis 5: Teenagers from female-headed homes are expected to have different rates of reporting having ever been pregnant than teenagers from male-headed homes.
Because the economy as reflected in employment opportunities has assisted in setting the stage for the dynamics that make low socioeconomic and Black teenagers to be more vulnerable to having ever been pregnant, the effects of race and SES may be expected to change over time. Two patterns of childbearing in response to economic conditions have been discussed in the literature. Colen et al. (2006) describe these patterns. Procyclical fertility is when the likelihood of bearing children increases during periods of economic growth and decreases during periods of economic recession. In the context of a nuclear family, the husband plays the role of breadwinner and the wife takes the major responsibility of caring for the children. During times of economic growth, greater resources are available to better afford children. This description does not apply to all cultural groups, especially those where women are heavily relied on to contribute financially as in the Black community. Countercyclical fertility is when the likelihood of bearing children decreases during periods of economic growth and increases during periods of economic recession. Here, women who are relied on to work outside the home are able to take advantage of employment prosperity by postponing childbearing. Colen et al. found support for countercyclical fertility among Black minor and adult teens after controlling for welfare policy, incarceration rates, and statelevel restrictions regarding abortion. They found teen pregnancy rates to increase for both minor and adult teens when compared with an increase in state unemployment rates. No significant effects of state unemployment rates on teen pregnancy rates were found by Colen et al. for White teenagers.
Hypothesis 6: The effect of race on having ever been pregnant as a teenager is expected to differ, depending on unemployment rates over time.
This dynamic is also expected to differ depending on SES. Some SES groups may be more vulnerable to ebbs and flows in the economy.
Hypothesis 7: The effect of SES on having ever been pregnant as a teenager is expected to differ, depending on unemployment rates over time.
Because minor and adult teens may respond differently to the economic opportunities due to a difference in dependence status, they are examined separately in determining the likelihood of saying "yes" to having ever been pregnant over time.
Hypothesis 8: The rates of having ever been pregnant as a minor are expected to differ when compared with the rates for all teenagers, depending on unemployment rates over time.

Description of the National Health and Nutrition Examination Survey (NHANES)
NHANES is a national survey conducted by the National Center for Health Statistics of the Centers for Disease Control and Prevention (2008). A complex, multistage sampling design is used to collect the data, which is adjusted to be representative of the U.S. noninstitutionalized civilian population. Data collection methods consisted of standardized in-home interviews as well as standardized physical examinations and laboratory studies conducted in Mobile Examination Centers (MEC). We used data collected from 1999 to 2006, during which teenage pregnancy status was available on the public-use data file for 15-to 19-year-olds. Of the 1,580 Black and White female teens aged 15 to 19 years, 1,318 (83.4% of those eligible) met the criteria for inclusion in our study after we excluded those for whom pregnancy status was unavailable (n = 110) and those with missing values due to nonresponse for ratio of family income to poverty (n = 110), education level of household reference person (n = 59), and marital status of household reference person (n = 137).

Determination of Pregnancy Status
The main outcome variable, ever pregnant (yes/no), was determined by the self-response answer to the question, "Have you ever been pregnant?" in addition to pregnancy status at the time of the MEC exam, determined by self-report or a positive urine pregnancy test. A pregnancy included a current pregnancy and previous live births, miscarriages, stillbirths, tubal pregnancies, and abortions. In our sample, 11.51% of the teens experienced a pregnancy.

Selected Characteristics
Sociodemographic information was collected through interviews conducted in the participants' homes. Age (in years) was determined at the time of the screening interview. Race (non-Hispanic Black/non-Hispanic White) was determined by self-reported responses to questions regarding race and Hispanic origin. SES was classified as low or not low. SES was determined by the education level of the household reference person and the ratio of family income to poverty. Poverty income ratio values below 1.00 are below the official poverty threshold determined by the U.S. Census Bureau, whereas values of 1.00 or greater indicate family income above the poverty level. Low SES was defined as the household reference person having less than a high school education or the poverty income ratio being less than 1.5. Not-low SES was defined as the household reference person having more than a high school education and the poverty income ratio of 1.5 or greater.
Parental status (single parent/not single parent) was determined by the marital status of the household reference person, which is the first household member listed on the Screener household member roster who is 18 years of age or older and who owns or rents the residence where members of the household reside. A single parent was defined as the household reference person being widowed, divorced, separated, or never married, and otherwise as married or living with partner. Sex of the head of household was used as well.
Of the single-parent households in this study, 87.4% were headed by females.

Economic Context
Four time periods reflect different states of the economy. The Bureau of Labor Statistics (2010) reports the unemployment rate as a percentage of the labor force 16 years of age and older for the 8 targeted years. In 1999-2000 the unemployment rate was low (4.2% and 4.0%, respectively), reflecting a healthy economy. Following this time period (2001)(2002), the unemployment rate began to rise, and this trend was aggravated in the aftermath of the events of September 11, 2001. The unemployment rate rose to 4.7% in 2001 and then to 5.8% in 2002. The third period (2003)(2004) reflects the peak of unemployment in 2003 (6.0%) and the beginning of recovery (5.5% unemployment rate) in 2004. The unemployment rate continued to decline during the fourth quarter under consideration (2005)(2006). During this period, the unemployment rate dropped to 5.1% in 2005 and dropped again in 2006 with an unemployment rate of 4.6%, almost down to what it was in 1999. Of the four quarters, one might identify the first and fourth quarters as better economies compared with the second and third quarters where opportunities of employment were not as great.

Analysis
All statistical analyses were conducted using SAS, version 9.2 (SAS Institute, Cary, North Carolina), and SAS-callable SUDAAN, version 10.01 (Research Triangle Institute, Research Triangle Park, North Carolina), on the Windows XP 64-bit platform. Stratum and primary sampling unit variables were used to adjust for the complex sampling design of NHANES, and MEC sampling weights were combined to obtain statistics weighted to the population and adjusted for nonresponse according to NHANES' analytic guidelines.
Differences between Black and White teens among selected characteristics were compared using χ 2 tests and odds ratios from simple logistic regression. Multiple logistic regression was used to estimate the odds of ever being pregnant for Black teens compared with White teens after adjusting for age, SES, parental status, and head of household. A second model added survey year to the other covariates to determine trends over time. A p value less than or equal to .05 was deemed significant for all statistical tests. Covariates in the full model were checked for multicollinearity, and the variance inflation factor (VIF) was below 2 for all variables (a VIF > 10 reveals a problem with collinearity).

Results
Pregnancy incidence for selected characteristics is given in Table 1. The pregnancy rate was found to be significantly higher for Blacks than Whites (21.73% vs. 9.06%, χ 2 = 43.59), for those with low SES compared with not-low SES (19.80% vs. 4.63%, χ 2 = 45.23), for teens living in single-parent households compared with teens living in intact homes (16.84% vs. 7.94%, χ 2 = 15.65), for teens living in households with a female head of household compared with those with a male head of household (14.41% vs. 7.97%, χ 2 = 7.39), and for adult teens (aged 18-19 years) compared with minor teens (aged 15-17 years; 19.57% vs. 6.65%, χ 2 = 36.32). The overall pregnancy rate has been dropping since the 1999-2000 period. However, there was a slight increase in 2005-2006. Crude odds ratios for selected characteristics are given in Table 2. All selected independent variables were found to have a significant relationship with the probability of ever being pregnant. Consistent with Wilson's (1987) analysis, the effect of low SES was found to have a much higher effect when compared with race. Single-parent household was found to have a slightly stronger effect when compared with female head of household. Adult teens (18-to 19-yearolds) were found to be more likely to say yes to ever being pregnant compared with minor teens (15-to 17-year-olds). It stands to reason that adult teens, being older, would be more likely to ever have been pregnant compared with minor teens. It is important to acknowledge that most teen pregnancy is due to adults getting pregnant. When the problem of teenage pregnancy is addressed, adult teens are often included in the rates provided, therefore painting a biased picture of our youth. Table 3 shows that race and SES remain significant after adjusting for other variables in the model providing support for both Hypothesis 1 that Black teenagers have a greater likelihood of reporting having ever been pregnant than White    (3.14, 8.24) in the crude model. Therefore, SES continues to have a greater effect on teenage pregnancy rates than race, indicating that low-SES teens are almost 4 times as likely to ever have been pregnant compared with not-low-SES teens, which is consistent with Hypothesis 3 that SES has a stronger effect on reporting having ever been pregnant than race. Single-parent household and female head of household were not significant in the adjusted model. This is consistent with Wilson's (1987) discussion that poor Black households have increasingly been headed by single-parent females. These results are inconsistent with both Hypothesis 4, supporting a unique effect of a single-parent household, and Hypothesis 5, positing a unique effect of female-headed household, separate and apart from race and SES. The odds ratio for age at screening also shows a reduction to 2.86 (1.83, 4.46) from 3.42 (2.28, 5.13) in the adjusted model. Table 4 examines the Race × SES × Time interaction. It reveals that the state of the economy determines the way in which race and SES affect teenage pregnancy and provides support for Hypothesis 6, that the effect of race on having ever been pregnant as a teenager is expected to differ depending on unemployment rates over time, and Hypothesis 7, that the effect of SES on having ever been pregnant as a teenager is expected to differ depending on unemployment rates over time. In -2000In and 2005In -2006, when the economy was good, Blacks were about equally as likely to ever have been pregnant compared with Whites (odds ratio = 1.18 [0.46, 3.04]; odds ratio = 1.01 [0.51, 1.99], respectively, n.s.). These two time periods also show that low-SES teenagers were about 3 times more likely to ever have been pregnant than not-low-SES teenagers (odds ratio = 3.16 [1.31, 7.66]; odds ratio = 2.91 [1.18, 7.17], respectively). Therefore, in a good economy, race does not decide teenage pregnancy-only poverty does.
After the economy declined, the effects of race and SES increased. In 2001-2002, Black teenagers were found to be about 4 times more likely than White teenagers to ever have been pregnant (odds ratio = 3.93 [1.85, 8.35]), and low-SES teenagers were found to be over 5 times more likely than not-low-SES teenagers to ever have been pregnant (odds ratio = 5.54 [2.35, 13.04]). In 2003-2004, when unemployment rates were at their highest, being a Black teenager made the likelihood of ever having been pregnant almost 7 times that of a White teenager (odds ratio = 6.72 [2.76, 16.32]). This is the time when the effect of race far exceeded the effect of low SES. During this time period, low-SES teenagers are twice as likely to ever have been pregnant, yet this relationship is only close to statistically significant (odds ratio = 2.21 [0.97, 5.06], p = .0598). Consistent with the main effects model (Table 3), neither the effect of singleparent household nor female head of household were found to be statistically significant. Also consistent with the main effects model, adult teenagers are found to be about 3 times more likely than minor teenagers to have ever reported being pregnant (odds ratio = 2.97 [1.95, 4.52]). The Race × SES × Time interaction as reflected in the contrast statements shed additional light on how race and SES interact over time. 5 In 1999-2000, low-SES teenage Blacks were found to be more likely than not-low-SES teenage Blacks to ever have been pregnant (odds ratio = 3.24 [1.07, 9.78]). This same contrast was found to be significant for Whites (odds ratio = 3.09 [0.99, 9.65]). No additional three-way interactions were found in 2005-2006. In a bad economy, some interesting three-way interactions do show up, however. In 2001-2002, low-SES White teenagers were found to be more likely than not-low-SES White teenagers to ever have been pregnant (odds ratio = 13.07 [3.40, 50.24]). This pattern was not found for Blacks. Not-low-SES Black teenagers were found to be more likely than not-low-SES White teenagers to ever have been pregnant (odds ratio = 9.26 [2.40, 35.70]). In 2003-2004, low-SES Black teenagers were found to be more likely than low-SES White teenagers to ever have been pregnant (odds ratio = 3.11 [1.00, 9.72]). Not-low-SES Black teenagers were also found to be more likely than not-low-SES White teenagers to ever have been pregnant (odds ratio = 14.49 [2.91, 72.11]). Figures 1 and 2 provide a different way of looking at the patterns of the effect of race and SES over time on teenage pregnancy rates (percentage of respondents in specified group saying "yes" to having ever been pregnant) and provide details that examining overall effects do not. These figures show additional support for Hypothesis 6 and 7 and support for Hypothesis 8 that the rates of having ever been pregnant as a minor are expected to differ when compared with the rates for all teenagers depending on unemployment rates over time. Figure 1 shows rates of having ever been pregnant for all teenagers (15-19) and Figure 2 for minor teenagers (15-17) only. Figure 1 shows that starting with a good economy in 1999-2000, low-SES Blacks were more likely to say "yes" to having ever been pregnant compared with low-SES Whites whereas not-low-SES Blacks were about equally as likely as not-low-SES Whites to say "yes" to having ever been pregnant. During 2001During -2002During and 2003During -2004 and times of worsening unemployment rates, the pregnancy rates of Blacks exceeded the pregnancy rates of Whites in their respective SES categories. In addition, the pregnancy rates of not-low-SES Blacks exceeded those rates for all Whites and were only slightly lower than those for low-SES Blacks in 2003-2004. Noteworthy is the widening gap between Blacks and Whites within their respective SES categories during poor economies and the very narrow gap between low-SES and not-low-SES Blacks in the quarter where unemployment rates peak. This peak of unemployment rates is also a time when the pregnancy rates for all Blacks exceed those for all Whites. In 2005-2006, when the economy picked up, the pregnancy rates for low-SES Blacks exceeded those of low-SES Whites. However, the rates for not-low-SES Blacks fell below those of not-low-SES Whites.
The overall rates for the two SES groups of Blacks showed countercyclical fertility: Low pregnancy rates are observed during 1999-2000 and 2005-2006 when the unemployment rates were low relative to higher pregnancy rates in 2001-2002 and 2003-2004, when the unemployment rates  were high. For not-low-SES Whites, procyclical fertility is apparent. High pregnancy rates are observed in the better economies of 1999-2000 and 2005-2006 relative to lower pregnancy rates in 2001-2002 and 2003-2004 when economies were worse. Low-SES Whites showed a continual decline in pregnancy rates until the last quarter when they showed an increase.
When pregnancy rates for minors are separated from the pregnancy rates for all teenagers (aged 15-19 at the age of screening), a somewhat different picture appears for low-SES minors. A continual decline in teen pregnancy rates is observed for low-SES Blacks, revealing a reduction in the pregnancy rates for Black minors between the first and second quarters not detected when the rates of having ever been pregnant for 15-to 17-year-olds is not separated from those of 18-to 19-year-olds. This decline, along with increases in the last quarter for low-SES Whites, brought the pregnancy rates for low-SES Blacks below those of low-SES Whites in the last quarter. This finding was not detected when considering all teens. No consistent pattern for low-SES White minors is detected. Not-low-SES Blacks show signs of countercyclical fertility and not-low-SES Whites show signs of procyclical fertility consistent with the results in Figure 1. Of concern is that during the quarter when unemployment rates were at their peak, the pregnancy rates for not-low-SES Black minors exceeded those rates for all other groups; however, their rates decreased so extensively in the last quarter as the economy picked up that they were lower than all other groups.

Discussion
The model of main effects shows how race and SES influence teenage pregnancy with SES being the stronger effect. This is consistent with the theoretical framework that Wilson (1987) uses to explain forms of social dislocation like teenage pregnancy. These results also show the continued importance of race emphasized by Ladner (1971). Not included were neighborhood effects, which are likely to be highly correlated with the individual effects of race and SES because aggregate level characteristics are made up of individual characteristics. Effects of single-parent household and female head of household are not found to be statistically significant in any of the multivariate analyses indicating that these factors have more to do with race and SES when predicting teenage pregnancy. Multivariate analyses show adult teenagers to be about 3 times as likely as minor teenagers to report having ever been pregnant. This reinforces the need to look at the rates for minors and adults separately when considering interventions aimed at reducing teen pregnancy as a social problem because most teen pregnancies are from adult teenagers (Hamilton et al., 2010;Martin et al., 2009).
The three-way interaction model reveals the effects of race and SES to be more complex than what the main effects models revealed. Being Black is found to have a statistically significant effect on teenage pregnancy only in poor economies. Teen pregnancy during better economies is due to differences in class. In good economies when there are greater job opportunities, Black teens may choose education, work, or career over motherhood. These findings complement the literature that reveals educational and occupational aspirations to be related to teenage pregnancy (Hellerstedt et al., 2001;Young et al., 2001). Educational and occupational aspirations may be higher in a good economy when opportunities are more apparent. This is also consistent with countercyclical fertility and the findings of Colen et al. (2006). In a poor economy, when the effect of race is significant, Blacks increase their rates of teen pregnancy compared with Whites, and the effect of being Black supersedes that of class.
When examining the pregnancy rates for different groups according to race and SES, patterns of countercyclical fertility were for Black teenagers of all ages and only for affluent Black minors when 15-to 17-year-olds are separated from the 15-to 19-year-olds. This finding supplements the findings of Colen et al. (2006) because it specifies which SES groups among Blacks follow this pattern. Times of economic decline would be those with fewer opportunities, and it is expected that racial minorities would experience the effects more greatly than those from the majority group. This indicates that prejudice and discrimination may be more prevalent when resources are scarce. Any gains in financial security that more affluent Blacks achieve may be short lived because they are more vulnerable than Whites in a poor economy. The fact that affluent Black minors show increases in teenage pregnancy during times of economic decline might reflect the struggle that affluent Blacks have in keeping their position in the economic social structure. They may work longer hours providing less supervision for their children relative to affluent Whites. Poor Blacks have more consistent economic circumstances, remain poor regardless of the economy, and are therefore less responsive to the economy.
Procyclical fertility patterns were found for affluent Whites when both teenage age groups are considered and also when only minor teenagers are considered. This is not consistent with Colen et al.'s (2006) findings of no effect of state unemployment rates on White teenage pregnancy rates. Their study was different in that they used different control variables; SES was not taken into account, and data were collected at the state aggregate level rather than the individual level. Their study was also different in that they were able to examine the teen pregnancy rates for minor and adult teenagers separately due to the availability of this data at the state level.
A finding of great importance is that the pregnancy rates for poor minor Blacks continue to decline throughout all time periods examined and drastically for the last quarter for affluent minor Blacks. Furthermore, the rates for minor Whites, although sporadic, increase and exceed those of poor minor Blacks in the last quarter, regardless of SES.
No consistent pattern for low-SES White minors is detected, and the fluctuations occur in a manner inconsistent with the economy, indicating that poor White minors may be responsive to other variables. Because of the recent increase in their pregnancy rates, one might propose that programs aimed at reducing teen pregnancy may be more effective for Blacks or indicate a lack of a cohesive White community in dealing with the problem. Apparently, teen pregnancy is becoming more of a problem for affluent and poor White minors of late compared with their Black counterparts as reflected in their recent rates. This finding would not have been detected without separating minors from adults. More analyses need to be conducted on the 15 to 17 age group because teenage pregnancy among minors is where the real social problem lies and where policy needs to be informed.

Summary and Conclusion
Race and SES affect teenage pregnancy but change as a function of the state of the economy. During a healthy economy, race ceases to have a unique impact, and the variation in teen pregnancy becomes a result of SES. During troubled economic times, however, being Black as compared with being White, is to increase the likelihood of being pregnant as a teenager. When teenage pregnancy rates according to race, SES, and age are observed over time, additional insights are revealed. In particular, more recently, White minors appear to be more pregnancy prone than Black minors. Because this study attempts to provide additional insight into the problem of teenage pregnancy according to race and SES by controlling for age and state of the economy, it can be concluded that Black teenagers should no longer be seen as the model for the problem of teen pregnancy. One limitation of this study is that it includes a narrow set of control variables. Further research should consider the remainder of the decade studied as well as when unemployment rates approach double digits for the nation. Similar analyses should also consider other racial/ethnic groups. Also not evaluated in this study are ways in which the rates of military service of Black men in differing economies may influence teen pregnancy rates. Research efforts should focus on the reasons that Blacks have decreased teen pregnancy rates over Whites. Social programs and community efforts aimed at reducing teenage pregnancy should be explored for the ways they influence Blacks and Whites differently. Cultural context and the ways in which different communities intervene (or not) in assisting parents in socializing youth should also be explored. These ways should be contemplated within race, SES, and immigration status groups. Studies should also examine minor teenagers differently than adult teenagers when deliberating on the problem of teenage pregnancy.
This study was not intended to be a policy piece and has not considered the wide body of literature that has policy implications; however, the authors make certain recommendations. A poverty stricken environment impedes the normal development of adolescents in becoming contributing members of society (Edelman & Ladner, 1991;Ladner, 1971;Musick, 1991;Sum & Fogg, 1991;Wilson, 1987Wilson, , 1996Wilson, , 1987Wilson, , 2003Wilson, , 2010. All recommendations center on problems due to poverty and changing economies, and contribute to self-sufficiency. Edelman and Ladner (1991) consider selfsufficiency to be the most important aim for policy considerations. First and foremost, basic safety needs must be addressed before personal development can take place. When one is striving for mere survival, orientation to the future is not likely to be a focus. The challenges Black girls face because of poverty, as documented in Ladner's (1971) study, are compounded by the violence and destruction of lives that accompany drugs and gangs in more recent years (Edelman & Ladner, 1991). Greater emphasis should be placed on community policing and neighborhood watch programs. In addition to the government's role in this matter, Edelman and Ladner (1991) stress the importance of community in providing solutions and for the implementation of those solutions. Community involvement is even more important given the current economy and limited resources. Involving the community may also contribute to self-sufficiency. The poor themselves are an untapped resource. Everyone's productivity is needed (Edelman & Ladner, 1991).
The second recommendation has to do with changing perceptions of opportunity. This study found fewer teenage pregnancies in better economies. Perceptions of opportunity during a poor economy is particularly important. The efforts of social workers, counselors (especially junior and high school counselors), and program planners should take into account which occupations are in demand when counseling youth and planning interventions relevant to education and work. For instance, in today's economy, opportunities for employment still exist in the medical field. This is important because the literature stresses the role of education and work for upward mobility. This recommendation is consistent with the work of William Julius Wilson that highlights the importance of perceived opportunities. Adolescents are not likely to become self-sufficient adults if they perceive their efforts to be fruitless.
Also relevant to changing perceptions of opportunity is the availability of role models for at-risk and troubled youth. Young women who have escaped poverty will make excellent role models for new generations of poor adolescents. Schools (all levels) are the logical site to place role models. Partnerships between public schools and universities are one way of providing role models. Upwardly mobile young women enrolled in the university have already been involved in after-school programs through internships. Pregnant minor programs like Cal-Safe in California might network with the university to provide role models who have already had a child as a teen and managed to succeed. Local churches are also a logical place for mentoring. Black churches have already made efforts to encourage their youth to succeed in school and pursue careers. Collins (2006) recommends collaborations between social work practitioners and Black churches in dealing with at-risk youth for problems such as teenage pregnancy.
The third recommendation has to do with the ways in which programs target at-risk youth (the poor and racial minorities). In light of this study's findings, one might speculate that poor Black minors may have responded to efforts to reduce teen pregnancy whereas more affluent Black minors have not. One might also consider that the extended kinship networks in the Black community and efforts within Black churches to monitor and regulate the behavior of youth may be stronger in less affluent neighborhoods. Teenage pregnancy is not a respecter of persons. Affluent and White adolescents are subject to many of the same challenges their poor and minority counterparts encounter in a society with relaxed sexual norms. However, one must consider that there are different ways of growing up. All children have the same needs for warmth and safety, but more affluent White children and poor Black children face different realities (Coles, 1991). Steinberg (1991) describes the bifurcation of the adolescent experience in contemporary America. He emphasizes the interaction between the developmental process and the social context. Context may impede psychosocial development, increase chances of psychological and social pathology, and "call into question the very logic of adolescence itself as it exists" for some children (Steinberg, 1991, 35). Sum and Fogg (1991) describe an extended period of economic adolescence for those with no postsecondary schooling and especially for those who are male and racial/ethnic minorities. Life skills classes are a likely avenue to provide assistance to all groups. The curriculum should be examined in light of the kinds of problems all adolescents grapple with in today's society and especially for ways in which they contribute to self-sufficiency. Fourth, intervention should be aimed at all ages. Ladner's (1971) study reveals that peer groups for poor Black girls become a major influence at an earlier age than the developmental literature purports. Peers rather than parents were found to be the preferred source of consultation and information before puberty. Edelman and Ladner (1991) also stress the importance of adolescent intervention because it is not too late to make a difference.
The fifth area of concern revolves around the relational problems of poor Black adolescents, which must be remedied through relationships (Musick, 1991). The parents of these adolescents, who have themselves been products of the same developmental disabilities and blocked opportunities, are also important targets for remediation. The lack of adequate male role models and examples of healthy male-female relationships also present problems in forming relationships. Interventions that target only the girls are not likely to have the needed impact as long as adolescent males are left out of intervention. Because the family is the primary mechanism for socialization, healthy relationships within the family are likely to provide the environment where children learn to be productive members of society. Economic security strengthens the capacity for young adults to marry and form stable relationships (Sum & Fogg, 1991). Economic security comes from taking advantage of available educational and career opportunities. These are opportunities that need to be made more transparent, especially during troubling economic times.
The last recommendation is to examine more thoroughly the complexities that accompany race and class. Analyses of interaction effects, such as those in this study, yield important findings. More qualitative studies like Ladner (1971) should take place in light of changes in Black and White communities. Such studies will be useful in identifying specific elements that contribute to the recent reduced minor teenage pregnancy rates for Blacks and the increased rates for Whites. When resources are limited, cost-effective alternatives must be proposed. This is the only way that growing up in poverty to become self-sufficient members of society, that which Ladner (1991) refers to as "mission impossible" (p. 1) becomes possible.

Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding
The author(s) received no financial support for the research and/or authorship of this article.