National Longitudinal Study of Adolescent Health
The database of the National Longitudinal Study of Adolescent Health (www.cpc.unc.edu/addhealth) was dearly won. The intention to create this database grew out of a realization on the part of researchers and policy makers that many public health decisions having to do with sexually transmitted diseases were being made on the basis of data collected half a century ago through the Kinsey sex surveys. It seemed absurd to be relying on data so old that it preexists AIDS to make decisions affecting the health and lives of people currently at risk for sexually transmitted diseases.
At the same time, these researchers and policy makers realized that there had never been a comprehensive survey of the health practices of adolescents.
These two needs drove a nearly decade-long fight in Congress to obtain funds to undertake the survey. Conservative lawmakers believed passionately that asking questions about the sexual lives of adolescents was way beyond anything government should sanction, let alone fund. So, year after year, money was placed in appropriation legislation to fund the study. And, year after year, the money was removed.
Many modifications in the research design were made, and, eventually, the go ahead was given to conduct the study. That some of the outcomes of the study seem to support conservative positions regarding sexuality had a noticeably positive effect on conservative support for the research. The survey, however, demonstrates in the best tradition of science that the outcomes of research are not set in advance.
To date, two waves of data collection have taken place. Wave I data were collected between September 1994 and December 1995. Wave II data were collected between April and August 1996. The survey is a school-based study of the health-related behaviors of adolescents in grades 7-12. It assesses the health status of participants and also explores the causes and contexts of their health-related behaviors.
Three contexts are examined: the social context, the personal decisions and practices themselves, and the personal and personality variables associated with particular behavioral decisions. Those three levels of data make this the richest dataset available for understanding the health behaviors of adolescents.
The sample size is over 90,000 adolescents from 80 communities chosen so as to make the sample representative of the U.S. by region, school type, urbanicity, ethnicity and school size. All respondents completed an in-school questionnaire. From these respondents, a core sample of 12,105 adolescents was taken. For these individuals, an additional in-home survey was conducted. Additional special oversamples of certain ethnic and racial groups were taken as well. There is also a large sample of adolescents with physical handicaps.
The questions it is possible to ask of these data are nearly limitless, and your opportunity to be one of those asking the questions is very real. The data are available in two forms-a public-use dataset and a restricted-access, contractual dataset. The sensitivity of these data makes protection of the respondents of cardinal importance. So the public use dataset contains only a subset of respondents. To gain access to the restricted-access data, you must be a certified researcher, and you must commit yourself to maintaining limited access. Deductive disclosure, that is, the ascertaining of the identity of a respondent by bringing together disparate pieces of data, can occur with this dataset. That is why anyone who uses the restricted dataset is obligated to protect respondents from that eventuality by submitting to a number of precautionary actions.
Don't be intimidated by the importance of the data. The Add Health Internet site is very well organized for the researcher, and the principal investigators are eager for researchers to wring the last scintilla of knowledge from the data. There is much more to be learned from these data than the causes of the health behaviors of adolescents. What do you want to know? The Add Health database may have your answer.
