The limited research evidence on adolescent romantic relationships is due in part to a lack of nationally representative data. No single national data set includes all of the information needed to assess teens’ early experiences with romantic relationships, their attitudes and expectations concerning romantic relationships and marriage, and their relationship outcomes in young adulthood. Therefore, for this report, we draw on data from four different sources: (1) the National Longitudinal Survey of Youth 1997 (NLSY97), (2) Monitoring the Future (MTF), (3) the 2002 National Survey of Family Growth (NSFG), and (4) the Youth Risk Behavior Survey (YRBS). We describe our use of these four data sets below.
National Longitudinal Survey of Youth 1997. With a large sample of young people, the NLSY97 is well suited for an analysis of high-school-aged youth. This survey, which began in 1997 and is conducted every year, follows respondents from their early teenage years into early adulthood. It was designed to be nationally representative of youth born from 1980 to 1984 (ages 12 to 16 in 1996). Further, the sample includes a large number of African American and Hispanic youth, which allows us to examine differences by race and ethnicity.
For this analysis, we focus on two waves of the NLSY97, one conducted when the respondents were in their teens and another that tracks them into their early 20s. The 1999 survey wave includes a sample of more than 6,600 teens between the ages of 15 and 18(1) and was conducted during the last year in which a large portion of the NLSY97 sample was still high school-aged. In Chapter II, we use the 1999 data to describe the youths’ families and their early relationship experiences. In Chapter IV, we use the 2005 wave of the NLSY97 to examine the relationship outcomes of this same cohort of young people when they were young adults between the ages of 20 and 25. In particular, we describe their patterns of dating, cohabitation, and marriage. All of the analyses are weighted to account for the survey sampling design and to reflect the national population in the appropriate year.
Monitoring the Future. Unfortunately, no data set currently tracks marriage attitudes and expectations for a nationally representative sample of teenagers. However, the MTF study tracks these items for a nationally representative sample of high school seniors. Since the mid-1970s, the MTF study has conducted an annual survey of 12th graders. Because the survey is conducted annually, the data can be used to track changes in the marriage attitudes and expectations of high school students over the past 30 years. Of course, trends in marriage attitudes among high school students may be different from trends among all teens. This potential difference should be kept in mind when interpreting results that are based on MTF data.(2)
The MTF study is best known for collecting and reporting information on youth substance use and risk behaviors (Bachman et al. 2002). However, the survey also includes a broad range of questions concerning attitudes toward marriage. For example, students are asked whether and when they expect to get married, whether they expect to stay married to the same person for life, whether they support the idea of couples living together before marriage, and whether they view marriage and family as important parts of their lives. We describe these measures in greater detail in Chapter III. The survey also collects data on demographic characteristics, information that can be used to examine how the attitudes and expectations of high school students differ by gender and racial and ethnic background.
Although the MTF study collects data for a large sample of more than 10,000 students per year, only a randomly selected subgroup of students are asked the survey questions concerning their attitudes toward marriage. Therefore, the analyses of MTF data presented in this report are based on smaller samples of about 2,500 students per year. When reporting results separately for different racial/ethnic groups, we pool data for two consecutive years to ensure that the sample sizes are large enough to generate precise estimates for each group. All of our analyses are weighted to account for the survey’s multistage sampling design.
National Survey of Family Growth. Additional information on teens’ attitudes toward marriage comes from the 2002 National Survey of Family Growth, a nationally representative survey of the non-institutionalized U.S. population ages 15 to 44 conducted by the National Center for Health Statistics (NCHS). For this report, we focus on data for the subsample of respondents ages 15 to 18, which includes 900 teenage boys and 909 teenage girls. Data from the NSFG can be used to examine how teens’ attitudes toward marriage differ by family background characteristics (such as income status and whether the teen lives with both biological parents), information that is not available in either the NLSY97 or the MTF. The NSFG also uses different measures of attitudes than the MTF, including measures of teens’ attitudes toward divorce and nonmarital childbearing. We describe these measures in greater detail in Chapter III. The NSFG oversamples African Americans and Hispanics, which allows us to conduct some of our analyses separately by racial and ethnic subgroups.(3) All of our analyses of the NSFG are weighted to account for the survey’s multistage sampling design. We also account for the sampling design when calculating statistical significance tests.
The Youth Risk Behavior Survey. The YRBS is a large, long-standing survey of high school students that was designed to monitor adolescent health risk behaviors, such as smoking, alcohol use, and sexual activity. The survey began in 1991 and is conducted every other year on a nationally representative sample of 10,000 to 16,000 9th through 12th grade students enrolled in public and private schools. For our purposes, the YRBS provides information on trends in sexual activity from 1991 to 2005, which we present in Chapter II.(4) As with the MTF study, the YRBS has the limitation of including only teens who are enrolled in high school. It, therefore, does not provide information on trends in sexual activity for all teens, including those who have dropped out of school.(5)