LONGITUDINAL DEMOGRAPHIC SURVEYS

Longitudinal surveys are surveys that involve repeated data collection from individuals over time. They are of two types: specific-purpose surveys and panel surveys. Specific-purpose surveys collect information on a topic-specific basis at successive times from comparable populations but not the same individuals. These kinds of surveys may generically be termed longitudinal surveys, and have historically represented the dominant form of longitudinal data collection. They are most appropriate for comparing, for example, changes in demographic, socioeconomic, social-psychological, or health behaviors or attitudes in populations or subgroups of populations over time. Panel surveys involve the repeated interviewing, over time, of the same individuals. Surveys of this type permit examination of transitions in individual behaviors, attributes, or attitudes over time, and of linkages over time at the individual level. Analytically, these two forms of data collection can often complement each other, and they each have distinctive strengths and weaknesses. The surveys discussed in this article–selected longitudinal surveys with substantial demographic content undertaken in the United States–are of both types.

In addition to the brief appended bibliography, on-line addresses are provided that give access to more detailed accounts of the surveys described.

Health-and Fertility-Related Surveys

National Health Interview Survey (NHIS). The NHIS, the main longitudinal health data collection activity in the United States, is a continuing cross-sectional (non-panel) national survey of the civilian, non institutional population. The survey, entailing personal interviews, is conducted by the Bureau of the Census for the National Institutes of Health. Information is collected annually from about 40,000 households and over 100,000 individuals on basic demographics, illnesses, injuries, impairments, chronic conditions, activity limitations, utilization of health services, and other health topics. The survey has been conducted annually since 1957, although it has undergone several major changes. Since 1995, black and Hispanic households have been over-sampled. That is, they are over-represented in the survey compared to their actual representation in the population, in order to ensure sufficient numbers for statistically reliable group comparisons. A core questionnaire is completed each year including information of the type noted above; periodically, modules on selected topics such as disease prevention or cancer are added. The core questionnaire includes a basic family questionnaire, an adult questionnaire, a child questionnaire (information is collected from an adult), and a child immunization questionnaire.

The National Survey of Family Growth (NSFG). The NSFG collects detailed information about family relationships and fertility-related experiences from a representative national sample of adults of childbearing age. Five rounds were held from 1973 to 1995, covering only female respondents. Beginning with the sixth round in (2002), the sample (some 19,000 respondents) includes males. The sampling frame is the same as for the NHIS. Public user files for past rounds of the survey are available.

The National Health and Nutrition Examination Survey (NHANES). In contrast with the NHIS survey, NHANES focuses more directly on the collection of health measurement statistics. This survey has gone through several cycles. From 1960 through 1970, three National Health Examination Surveys (the name that it was known by prior to 1999) were conducted, covering (sequentially) chronic adult diseases, early and later childhood growth and development. Beginning in 1970, a new emphasis on health and nutrition was introduced, and increasing attention was given to over-sampling of minority groups in the population. Since 1999, NHANES has been an annual survey in which about 5,000 respondents are interviewed in each round. It entails detailed household interviews, physical examinations, and health and dietary interviews. Its data can be linked to related surveys such as the NHIS.

National Longitudinal Study of Adolescent Health (Add Health). Add Health is a panel survey of the health-related behaviors of adolescents who were in grades 7 through 12 as of the first survey wave (1994–1995). It is a fully representative national survey that focuses on adolescent sexuality, its social, psychological, and physiological correlates, and to some degree its determinants. Its premise is that families, friends, schools, and communities can encourage healthy or unhealthy behaviors among adolescents. In wave one, data were collected directly from about 90,000 youths in randomly selected schools in 80 U.S. communities. Three follow-up waves focused on a sub-sample of 20,000, who were interviewed at home–the last of them in 2001 to 2002, when the respondents were 18 to 25 years of age. Additional follow-up surveys are planned. Supplementary data collection from parents, school personnel, and siblings was also conducted. The sample includes a substantial number of minority youth. Public-use Add Health data are distributed by the Sociometrics Corporation on its web site.

Employment-and Income-Focused Surveys

The Current Population Survey (CPS). The CPS is a nationwide survey of about 50,000 to 60,000 civilian non-institutional households conducted monthly for the U.S. Bureau of Labor Statistics (BLS) by the Bureau of the Census. Dating from the 1940s, the CPS provides a continuous profile of the changing American population and is the primary source of information on the employment characteristics of the population. In addition to the core monthly employment and unemployment data, this survey also collects extensive demographic, social, and economic data. Recent CPS outputs of general interest include an annual demographic supplement that includes statistics on work experience over the year, income, migration, and household composition, and periodical supplements on race and ethnicity (in 1995 and 2000), marital history (1995), fertility (2000), and school enrollment (annually in October). Special topics vary from year to year, depending on current issues of interest to government agencies. While the CPS is usually thought of as a continuing cross-sectional survey, it also has some short-term panel qualities. Households selected for inclusion remain in the survey for four months, are skipped for eight months, and then return for an additional four months; hence, a relatively large sample is available for short-term panel analyses.

The Panel Study of Income Dynamics (PSID). The PSID, begun in 1968, is essentially a longitudinal survey of a representative national sample of United States population, and the households in which they reside. The study is conducted by the Survey Research Center at the University of Michigan. It emphasizes the dynamic aspects of economic and demographic behavior, but its content includes explanatory and outcome measures drawn from several disciplines. PSID has sought to maintain continuity over time in the collection of basic data items, especially on the source and amount of income, family structure and demographic behavior, labor market activity, housing, and geographic mobility. Occasional survey supplements have covered such topics as wealth accumulation, neighborhood characteristics, health care, and child development.

The sampling methodology for the PSID is complex. For example, adults who form their own families become respondent units in subsequent rounds. Largely as a result of this, the sample size grew from 4,800 families in 1968 to over 7,000 by 2001. Since 1997, the survey has been conducted biennially.

The child development study, begun in 1997, involves a sub-sample of about 3,500 children, providing detailed information on care-giving, within-family time use of children, and selected measures of children's cognitive, emotional, and behavioral development–all of which can be readily linked with the main PSID data file. PSID data are available to public users on the Internet.

The National Longitudinal Surveys (NLS). NLS is a generic term that encompasses a set of longitudinal panel surveys that have been conducted since 1966. Data were still being collected regarding four cohorts in 2002: (1) women who were 14 to 24 when first interviewed in 1968 or 30 to 44 in 1967–about 5,300 women; (2) men and women 14 to 21 in 1978–about 8,000 persons; (3) younger and older children of the NLSY 79 female respondents–about 8,000 subjects; and (4) men and women ages 12 to 16 when first interviewed in 1997–about 8,000 subjects. Other than the child data collected in the NLSY97 survey, all these surveys focused on labor market dynamics. Their data also cover a wide range of complementary behavioral and attitudinal data on education, training, and family and household structure, enabling researchers to explore linkages between dimensions of employment and other factors. The 1979 survey's child data included in-depth information on psychometrics and other dimensions of child development. Information and data for these surveys are available from the Center for Human Resource Research at Ohio State University on its website.

Survey of Income and Program Participation (SIPP). SIPP is a continuing series of national panel surveys undertaken by the Bureau of the Census, designed to collect data on income, labor force, participation in government transfer programs, and general demographic conditions. Sample sizes range from 14,000 to 36,000 households. The survey is built around a core of labor force, program participation, and income questions. At its outset in 1984, SIPP was intended to measure the effectiveness of transfer programs; over time, it has evolved to become more of an omnibus survey that meets a wide range of research objectives. Topical modules have included personal histories, childcare and support, and school enrollment. Public use data for selected waves of the survey are available, and additional information may be found on the Internet.

An add-on to SIPP is the Survey of Program Dynamics (SPD), conducted from 1997 through 2002 with a sample of 18,000 households. The SPD survey includes an extensive set of questions regarding the children of the interviewed adult–questions on schooling, health, and child-focused activities–and a self-report from adolescent children about a wide range of their activities and behaviors.

Education Surveys

NLS—72/HS&B/NELS:88. Three longitudinal studies of youth, essentially covering the years since 1972, are panel surveys that follow a series of high school cohorts over time. These surveys are: (1) The High School Class of 1972 (NLS—72), a national sample of about 19,000 high school students followed from when they were seniors through 1986;(2) The High School and Beyond survey (HS&B), a sample from the class of 1980 initially comprising about 30,000 sophomores (followed to 1992) and 12,000 seniors (followed to 1986); and (3) The National Education Longitudinal Survey, 1988 (NELS:88) that followed about 18,500 8th graders from 1988 to 2000. These surveys collected detailed information about school progression and transitions to the work force, and a variety of demographic and family information.

Early Childhood Longitudinal Study (ECLS). The aim of these two panel surveys is to provide data on child development from infancy to the beginning of adolescence. The kindergarten component follows a nationally representative sample of about 22,000 children who attended kindergarten from 1998 to 1999 through the fifth grade; the birth cohort component follows 13,500 children born in 2001 from nine months of age through first grade. Both surveys examine the effects of family, school, community, and individual characteristics on a child's development. The samples include significant minority representation.

Aging Surveys/Family Processes Surveys

The Longitudinal Study of Aging. This panel survey was initiated in 1984 as a supplement on aging (SOA) to the NHIS. It included about 7,500 persons aged 70 and over, and reinterviews were conducted with most respondents in 1986, 1988, and 1990. The primary objective was to obtain data to (1) describe the continuum of movement from functional independence through dependence, including institutionalization and death; and (2) provide morbidity and mortality statistics by various demographic characteristics. These individual interview data have been linked with various other data files from the NHIS, and other forms of available records. The data are available on CD ROM.

Health and Retirement Study (HRS)/Study of Assets and Health Dynamics Among the Oldest Old (AHEAD). The HRS, a panel survey, was initiated by the National Institute on Aging in 1990. The first data collection wave in 1992 included over 12,600 persons in 7,600 households who were members of the 1931 to 1941 birth cohort. The sample includes an over-representation of minority respondents. The core sample has been reinterviewed at two-year intervals. The interviews cover health, retirement behavior and plans, family structure, income and employment, and related topics. The data can be linked with several administrative data sets including social security earnings data and the national death index.

Beginning in 1993, a parallel survey of 7,447 respondents (and spouses) aged 70 and over, the AHEAD study, was initiated. Additional birth cohorts have been added to this data set since its inception, and in 1998 the AHEAD and HRS samples were merged. The data are collected by the Survey Research Center at the University of Michigan.

National Survey of Families and Households (NSFH). This panel survey represents a first attempt to comprehensively interview a large nationally representative sample of the population about behaviors, attributes, and attitudes regarding a full range of family-linked activities for a wide variety of family types. Personal interview waves were conducted in 1987 to 1988, 1992 to 1993, and 2001 to 2002. The first wave includes 13,007 respondents in 9,637 households, with an over-representation of minority household units, single-parent families, families with step-children, cohabiting couples, and recently married persons, as well as selected focal children in the household. The second wave followed up on a large proportion of the original respondents, current and past spouses, partners and other core family members. The third wave consists of telephone interviews with a subset of the second wave respondents. The large sample size for family units under-going transitions, in conjunction with the depth of behavioral and attitudinal detail, permit comprehensive examination of family processes and transitions.

See also: Census; Databases, Demographic; Demographic Surveys, History and Methodology of.

bibliography

Bureau of Labor Statistics, U.S. Department of Labor. 2002. The National Longitudinal Surveys Handbook. Columbus: Center for Human Resource Research, Ohio State University.

Carley, M. L., K. L. Muller, E. A. McKean, and E. L. Lang. 1997. National Survey of Family Growth: Cycle V, 1995: A User's Guide to the Machine-Readable Files and Documentation. (Data Set N8–O5). Los Altos, CA: Sociometrics Corporation, Data Archive on Adolescent Pregnancy and Pregnancy Prevention.

Curtain, T. R., S. J. Ingels, S. Wu, and R. Heuer. 2002. National Education Longitudinal Study of 1988, Base Year to Fourth Follow-Up Data File User's Manua (NCES 2002–323). U.S. Department of Education. Washington, D.C.: National Center for Education Statistics. Retrieved fromNational Education Longitudinal Study: 1988–2000 Data Files and Electronic Codebook System–Base Year through Fourth Follow Up ECB/CD-ROM, Public Use.

Hill, Martha S. 1992. The Panel Study of Income Dynamics: A User's Guide. Newbury Park, CA: Sage Publications.

Kelley, M. S., and J. L. Peterson. 1997. The National Longitudinal Study of Adolescent Health (Add Health), Waves I and II, 1994-1996: A User's Guide to the Machine-Readable Files and Documentation (Data Sets 48–50, 98, A1–A3). Los Altos, CA: Sociometrics Corporation, American Family Data Archive.

U.S. Department of Education, National Center for Education Statistics. 2000. National Education Longitudinal Study: 1988–2000 Data Files and Electronic Codebook System–Base Year through Fourth Follow-Up ECB/CD-ROM, Public Use. [NCES 2002–322 CD-ROM]. Washington, D.C.: U.S. Department of Education, National Center for Education Statistics.

Westat. 2001. Survey of Income and Program Participation Users' Guide Third Edition. Washington, D.C.: Mathematica Policy Research.

Zill, Nicholas, and Margaret Daly, eds. 1993. Researching the Family: A Guide to Survey and Statistical Data on U.S. Families. Washington, D.C.: Child Trends.

internet resources.

Child Development Supplement to the Panel Study of Income Dynamics. Sponsor: National Institute of Child Health and Human Development. <http://www.isr.umich.edu/src/childdevelopment/home.html/>.

Current Population Survey. Bureau of Labor Statistics and Bureau of the Census. <http://www.bls.census.gov/cps//>.

Early Childhood Longitudinal Study. National Center for Educational Statistics. <http://www.nces.ed.gov/ecls/>.

Health and Retirement Study: A Longitudinal Study of Health, Retirement, and Aging Sponsored by the National Institute of Aging. Institute for Social Research, University of Michigan. <http://www.umich.edu/~hrswww>.

High School and Beyond: Overview. National Center for Education Statistics. <http://nces.ed.gov/surveys/hsb//>.

National Center for Health Statistics: Monitoring the Nations Health. Centers for Disease Control and Prevention (CDC). <http://www.cdc.gov/nchs/nhis.htm/>.

National Center for Health Statistics: Monitoring the Nations Health. Centers for Disease Control and Prevention (CDC). <http://www.cdc.gov/nchs/about/major/nsfg/nsfgback.htm/>.

National Center for Health Statistics: Monitoring the Nations Health. Centers for Disease Control and Prevention (CDC). <http://www.cdc.gov/nchs/nhanes.htm/>.

National Education Longitudinal Study of 1988: NELS 88 Overview. National Center for Education Statistics. <http://nces.ed.gov/surveys/nels88//>.

National Longitudinal Survey of the High School Class of 1972: Overview. National Center for Education Statistics. <http://nces.ed.gov/surveys/nls72//>.

National Survey of Families and Households. NICHD and NIA. <http://www.ssc.wisc.edu/nsfh/>.

Ohio State University Center for Human Resource Research. Ohio State University. <http://www.chrr.ohio-state.edu//>.

Overview of the Panel Study of Income Dynamics. National Science Foundation. <http://www.isr.umich.edu/src/psid/overview.html/>.

Research Design, Facts at a Glance. Add Health. <http://www.cpc.unc.edu/addhealth/facts.html/>.

Research Practice Feedback. Sociometrics Corporation. <http://www.socio.com/>.

United States Census Bureau. <http://www.sipp.census.gov/sipp/>.

Frank Mott

Thomas Gryn