Search results

The Americans' Changing Lives (ACL) survey series is an ongoing, nationally representative, longitudinal study focusing especially on differences between Black and White Americans in middle and late life. These data constitute the first, second, third, fourth, fifth, and sixth waves in a panel survey covering a wide range of sociological, psychological, mental, and physical health items. Wave I of the study began in 1986 with a nation face-to-face survey of 3,617 adults ages 25 and up, with Black Americans and people aged 60 and over over-sampled at twice the rate of the others. Wave II constitutes face-to-face re-interviews in 1989 of those still alive. Survivors have been re-interviewed by telephone, and when necessary face-to-face, in 1994 (Wave III), 2001/02 (Wave IV), 2011 (Wave V), and 2019/21 (Wave VI).

Please note that for Wave VI, the majority of data collection occurred in 2019, with only a small subset (n=39) of participants surveyed in 2021.

ACL was designed and sought to investigate the following: (1) The ways in which a wide range of activities and social relationships that people engage in are broadly "productive," (2) how individuals adapt to acute life events and chronic stresses that threaten the maintenance of health, effective functioning, and productive activity, and (3) sociocultural variations in the nature, meaning, determinants, and consequences of productive activity and relationships. Among the topics covered are interpersonal relationships (spouse/partner, children, parents, friends), sources and levels of satisfaction, social interactions and leisure activities, traumatic life events (physical assault, serious illness, divorce, death of a loved one, financial or legal problems), perceptions of retirement, health behaviors (smoking, alcohol consumption, overweight, rest), and utilization of health care services (doctor visits, hospitalization, nursing home institutionalization, bed days). Also included are measures of physical health, psychological well-being, and indices referring to cognitive functioning.

Demographic information provided for individuals includes household composition, number of children and grandchildren, employment status, occupation and work history, income, family financial situation, religious beliefs and practices, ethnicity, race, education, sex, and region of residence.

The 1993 Detroit Area Study explored a variety of issues related to health, the effects of aging, living conditions, and participation in civic life in the Michigan tri-county area of Wayne, Oakland, and Macomb counties. A battery of questions probed respondents' perceptions of their health and mental state and those of their spouse, their ability to perform certain physical and mental activities, and the effect of their emotional state on their appetite and sleeping patterns. Other explored their feelings about neighborhood safety, means of transportation, relationships, accommodation, the portrayal of older people on television programs, and the treatment of older people by employers. The survey also sought respondents' opinions about government, their personal financial situation and problems, money management, savings and investments, and their life as a whole. Additional items questioned respondents about the frequency of their visits to the doctor, overnight hospitalization, chronic health conditions, smoking and drinking habits, and medical coverage, as well as electoral participation, political party preference, ideological leanings, class self-identification, assistance received from community organizations, family, and friends, personal regrets, and time spent watching television and engaging in pleasurable activities. Other questions gauged respondents' memory, vision, and motor skills. Respondents also provided demographic information on sex, age, marital status, race, ethnicity, religion, and education.

The Family Exchanges Study Wave 1 (FESI) was conducted in 2008 by the Institute for Survey Research at Temple University. The original 634 "target" or core sample was recruited from African American and White respondents aged 40-60 living in Philadelphia and the surrounding counties--Bucks, Chester, Delaware, and Montgomery. To be eligible for the study, respondents had to have at least one living parent and one living offspring over 18 years of age. Temple University sought to recruit the parents, spouse, and up to three offspring over 18 years of age into the study. All target, parent, and spouse surveys were conducted by telephone. Offspring were given the option of completing the survey by telephone or web. A total of 337 parents, 511 offspring (with another 80 by web and 1 listed as other for a total of 592), and 197 spouses were successfully recruited into the first wave of the study.

This collection includes four data files, one for each type of participant: target, spouse, parent, and offspring. For each of these participants, there are data related to relationships with other family members, perceptions of family members, and views on key social issues. Demographic information includes gender, marital status, education level, religion, age, race, ethnicity, and employment status.

The Family Exchanges Study (FESI) began in 2008 conducted by the Institute for Survey Research at Temple University. The original "target" or core sample was recruited from African American and White respondents aged 40-60 living in Philadelphia and the surrounding counties--Bucks, Chester, Delaware, and Montgomery. To be eligible for the study, respondents had to have at least one living parent and one living offspring over 18 years of age. Temple University sought to recruit the parents, spouse, and up to three offspring over 18 years of age into the study. All target, parent, and spouse surveys were conducted by telephone. Offspring were given the option of completing the survey by telephone or web.

For the Wave 2 data collection, the Survey Research Center at Pennsylvania State University attempted to contact all FESI respondents again, as well as collect updated information for spouses/romantic partners, parents, and up to 4 age-eligible offspring. The survey instruments were drawn largely from the first wave of data collection. This collection includes eight data files. These data files include five main study datasets: target, spouse, spouse without target, parent, and offspring. This collection also includes three diary datasets: target diary, parent diary, offspring diary. For each participant, there are data related to relationships with other family members, perceptions of family members, and views on key social issues. Demographic information includes gender, marital status, education level, religion, age, race, ethnicity, and employment status.

The Midlife in the United States (MIDUS) is a collaborative, interdisciplinary investigation of patterns, predictors, and consequences of midlife development in the areas of physical health, psychological well-being, and social responsibility. A description of the study and findings from it are available at http://www.midus.wisc.edu.

The first wave of the MIDUS study (MIDUS 1 or M1) collected survey data from a total of 7,108 participants. The baseline sample was comprised of individuals from four subsamples: (1) a national RDD (random digit dialing) sample (n=3,487); (2) oversamples from five metropolitan areas in the U.S. (n=757); (3) siblings of individuals from the RDD sample (n=950); and (4) a national RDD sample of twin pairs (n=1,914). All eligible participants were non-institutionalized, English-speaking adults in the coterminous United States, aged 25 to 74.

Data from the samples were collected primarily in 1995/96. The survey (Project 1) dataset contains responses from a 30-minute Phone interview and two 50-page Self-Administered Questionnaire (SAQ) instruments. Of the 7,108 respondents who completed the Phone interview, 6,325 also completed the SAQ.

This updated version of the study is comprised of three primary datasets:

Dataset 1, Main, Siblings, and Twin Data, contains responses from the main survey of 7,108 respondents. Respondents were asked to provide extensive information on their physical and mental health throughout their adult lives, and to assess the ways in which their lifestyles, including relationships and work-related demands, contributed to the conditions experienced. Those queried were asked to describe their histories of physical ailments, including heart-related conditions and cancer, as well as the treatment and/or lifestyle changes they went through as a result. A series of questions addressed alcohol, tobacco, and illegal drug use, and focused on history of use, regularity of use, attempts to quit, and how the use of those substances affected respondents' physical and mental well-being. Additional questions addressed respondents' sense of control over their health, their awareness of changes in their medical conditions, commitment to regular exercise and a healthy diet, experience with menopause, the decision-making process used to deal with health concerns, experiences with nontraditional remedies or therapies, and history of attending support groups. Respondents were asked to compare their overall well-being with that of their peers and to describe social, physical, and emotional characteristics typical of adults in their 20's, 40's, and 60's. Information on the work histories of respondents and their significant others was also elicited, with items covering the nature of their occupations, work-related physical and emotional demands, and how their personal health had correlated to their jobs. An additional series of questions focusing on childhood queried respondents regarding the presence/absence of their parents, religion, rules/punishments, love/affection, physical/verbal abuse, and the quality of their relationships with their parents and siblings. Respondents were also asked to consider their personal feelings of accomplishment, desire to learn, sense of control over their lives, interests, and hopes for the future.

The Datasets previously numbered 2 and 3 have been removed to avoid redundancies, and all datasets have been renumbered. Please refer to the readme file.

Dataset 2, Twin Screener Data, provides the first national sample of twin pairs ascertained randomly via the telephone.

Dataset 3, Coded Text Responses, describes how open-ended textual responses in the MIDUS 1 Computer-Assisted Telephone Interview (CATI) and Self-Administered Questionnaire (SAQ) were transformed into categorical numeric codes. These codes are included in a stand-alone dataset containing only those cases (N=3,950) that contained text data in their responses.

Online Analysis Only: Datasets 1, 2, and 3 were merged together by the SU_ID variable to form "Merged Data with Weights (Online Analysis Only)" (Dataset 4) for online analysis capabilities.

MIDUS also maintains a Colectica portal, which allows users to interact with variables across waves and create customized subsets. Registration is required.

In 2005, 592 African Americans from Milwaukee were added to the MIDUS sample to examine health issues in minority populations (for more details, see Midlife in the United States (MIDUS 2): Milwaukee African American Sample [ICPSR #22840]). Respondents were interviewed in their homes using a Computer Assisted Personal Interview (CAPI) survey protocol and asked to complete and return a Self-Administered Questionnaire (SAQ). Afterwards these individuals were eligible for participation in the same research protocol as the national MIDUS 2 sample, including cognitive, daily stress, biomarker, and neuroscience projects.

With support from the National Institute on Aging, a second wave of survey data collection on the Milwaukee sample was begun in 2016. The survey consisted of a 2.5 hour CAPI interview followed by a 45-page mailed SAQ. CAPI survey data was collected for 389 individuals, realizing a 78 percent response rate, adjusted for mortality and other eligibility criteria. Data collection for this follow-up wave largely repeated baseline assessments, with additional questions in selected areas (e.g., economic recession experiences, childhood experience with race, etc.). Following successful completion of the CAPI and SAQ protocols, individuals were eligible for participation in cognitive, daily stress, biomarker, and neuroscience projects.

In 2011-2014, the MIDUS Refresher study recruited a national probability sample of 3,577 adults, aged 25 to 74, designed to replenish the original MIDUS 1 baseline cohort and paralleling the five decadal age groups of the MIDUS 1 baseline survey [ICPSR 2760]. The MIDUS Refresher survey employed the same comprehensive assessments as those assembled on the existing MIDUS sample, but with additional questions about the effect of the economic recession of 2008-09.

The MIDUS Refresher collection is split into two datasets: Aggregate Data and Coded Text Data. The Coded Text Dataset provides coded responses to open-ended question items in the Aggregate Dataset. The survey data collection (Project 1) [MIDUS, ICPSR 2760] consisted of a 30-minute phone interview followed by two 50-page mailed self-administered questionnaires. Survey data were collected on demographic, psycho-social, and physical and mental health information. This new crosssectional MIDUS sample allows the examination of period effects on health (mental and physical) related to the economic recession by comparing the pre-recession MIDUS 1 sample with the post-recession MIDUS Refresher sample. A further objective of the MIDUS Refresher sample was to strengthen cross-project analyses in MIDUS by increasing the sample sizes available for testing hypotheses dealing with the interplay of key factors (e.g., socioeconomic status, gender, psychosocial factors, biological factors) in mid- and laterlife health. To that end, the MIDUS Refresher sample followed the same multi-disciplinary protocol established in the main MIDUS sample, in that after completing the survey protocol respondents were asked to complete a cognitive assessment by phone (Project 3) [MIDUS 3, ICPSR 36346] and later became eligible to participate in daily diary assessments (Project 2) [MIDUS 2, ICPSR 4652] biomarker assessments (Project 4) [MIDUS 2: Biomarker Project, ICPSR 29282] and neuroscience assessments (Project 5) [MIDUS 2: Neuroscience Project, ICPSR 28683].

The MIDUS Refresher was funded by the National Institute on Aging as two separate but related efforts: The MIDUS Refresher younger decades (MRY), was fielded in November, 2011, and recruited over 2,100 new participants aged 25 to 54; Funding was later added for the MIDUS Refresher older decades (MRO), which was fielded in June, 2013 and recruited over 1,400 new participants aged 55 to 74.

Demographic variables include age, sex, gender, race, religion, and marital status.

In 2012-2013, the MIDUS Milwaukee Refresher study recruited a sample of 508 Milwaukee African American adults, aged 25 to 64, designed to replenish the Midlife Development in the United States (MIDUS 2): Milwaukee African American Sample, 2005-2006 (ICPSR 22840). This sample was also designed to increase the number of racial minorities included in the broader MIDUS study. The MIDUS Milwaukee Refresher survey employed the same assessments (demographic, psycho-social, and physical and mental health) as those assembled on the existing MIDUS sample, but with additional questions about the effect of the economic recession of 2008-09. A sample of African Americans from Milwaukee County, Wisconsin, was stratified by age, gender, and income.

Area probability sampling methods were used to identify potential respondents. Field interviewers screened households to determine if they contained any African American adults. There was additional screening to achieve an appropriate age/gender distribution in a manner similar to what was done for the original MIDUS sample (Midlife in the United States (MIDUS 1), 1995-1996 [ICPSR 2760]). Milwaukee respondents were interviewed in their homes using a 2.5-hour Computer Assisted Personal Interview (CAPI) protocol and afterwards asked to complete a self-administered questionnaire (SAQ). All measures paralleled those used in the larger MIDUS samples. In addition to successful completion of the survey, participants were asked to complete a cognitive assessment by phone. Some respondents were eligible to participate in additional MIDUS projects: daily diary assessments, biomarker assessments, and neuroscience assessments.

In 2011-2014, the MIDUS Refresher study recruited a national probability sample of 3,577 adults, aged 25 to 74, designed to replenish the original MIDUS 1 baseline cohort and paralleling the age groups of the MIDUS 1 baseline survey in 1995. The MIDUS Refresher survey employed the same comprehensive assessments as existing MIDUS sample, with additional questions about the effect of the Great Recession in 2008-09. Survey data were collected on demographic, psychosocial, and health and well-being information.

In 2022-2024, the second wave of survey data (MIDUS Refresher 2) was collected on longitudinal participants, including the questions (1) repeated from the Refresher 1, (2) new impacts of the COVID-19 pandemic replacing the questions related to the Great Recession in the Refresher 1, and (3) the additional questions in selected areas (e.g., AD8, IADL, family history of dementia). This new longitudinal MIDUS data allow examination of period effects on health and well-being related to the COVID-19 pandemic by comparing the pre-pandemic MIDUS Refresher 1 data with the post-pandemic MIDUS Refresher 2 data. Further, the longitudinal MIDUS Refresher datasets (Wave 1 and Wave 2) allow investigation of the two major macro-level historic events, the Great Recession and the COVID-19 pandemic, on health and well-being across various population groups in the U.S.

In 2012-2013, the MIDUS Milwaukee Refresher study recruited a sample of 508 Milwaukee African American adults, aged 25 to 64, designed to replenish the original MIDUS Milwaukee 1 survey that was designed to increase the number of racial minorities in 2005. The MIDUS Milwaukee Refresher survey employed the same comprehensive assessments (demographic, psycho-social, and physical and mental health) as those on the existing MIDUS sample, with additional questions about the effect of the Great Recession in 2008-09.

In 2023-2024, the second wave of survey data was collected from 327 longitudinal Refresher Milwaukee participants, including the questions (1) repeated from the Refresher 1, (2) new impacts of the COVID-19 pandemic replacing the questions related to the Great Recession in the Refresher 1, and (3) the additional questions in selected areas (e.g., AD8, IADL, family history of dementia). This new longitudinal MIDUS Refresher Milwaukee data allow examination of period effects on health and well-being related to the COVID-19 pandemic by comparing the pre-pandemic MIDUS Refresher Milwaukee 1 data with the post-pandemic MIDUS Refresher Milwaukee 2 data. Further, the longitudinal MIDUS Refresher datasets (wave 1 and wave 2) allow investigation of the two major macro-level historic events, the Great Recession and the COVID-19 pandemic, on health and well-being.

Downloads of Add Health require submission of the following information, which is shared with the original producer of Add Health: supervisor name, supervisor email, and reason for download. A Data Guide for this study is available as a web page and for download.

The National Longitudinal Study of Adolescent to Adult Health (Add Health), 1994-2018 [Public Use] is a longitudinal study of a nationally representative sample of U.S. adolescents in grades 7 through 12 during the 1994-1995 school year. The Add Health cohort was followed into young adulthood with four in-home interviews, the most recent conducted in 2008 when the sample was aged 24-32. Add Health combines longitudinal survey data on respondents' social, economic, psychological, and physical well-being with contextual data on the family, neighborhood, community, school, friendships, peer groups, and romantic relationships.

Add Health Wave I data collection took place between September 1994 and December 1995, and included both an in-school questionnaire and in-home interview. The in-school questionnaire was administered to more than 90,000 students in grades 7 through 12, and gathered information on social and demographic characteristics of adolescent respondents, education and occupation of parents, household structure, expectations for the future, self-esteem, health status, risk behaviors, friendships, and school-year extracurricular activities. All students listed on a sample school's roster were eligible for selection into the core in-home interview sample. In-home interviews included topics such as health status, health-facility utilization, nutrition, peer networks, decision-making processes, family composition and dynamics, educational aspirations and expectations, employment experience, romantic and sexual partnerships, substance use, and criminal activities. A parent, preferably the resident mother, of each adolescent respondent interviewed in Wave I was also asked to complete an interviewer-assisted questionnaire covering topics such as inheritable health conditions, marriages and marriage-like relationships, neighborhood characteristics, involvement in volunteer, civic, and school activities, health-affecting behaviors, education and employment, household income and economic assistance, parent-adolescent communication and interaction, parent's familiarity with the adolescent's friends and friends' parents.

Add Health data collection recommenced for Wave II from April to August 1996, and included almost 15,000 follow-up in-home interviews with adolescents from Wave I. Interview questions were generally similar to Wave I, but also included questions about sun exposure and more detailed nutrition questions. Respondents were asked to report their height and weight during the course of the interview, and were also weighed and measured by the interviewer.

From August 2001 to April 2002, Wave III data were collected through in-home interviews with 15,170 Wave I respondents (now 18 to 26 years old), as well as interviews with their partners. Respondents were administered survey questions designed to obtain information about family, relationships, sexual experiences, childbearing, and educational histories, labor force involvement, civic participation, religion and spirituality, mental health, health insurance, illness, delinquency and violence, gambling, substance abuse, and involvement with the criminal justice system. High School Transcript Release Forms were also collected at Wave III, and these data comprise the Education Data component of the Add Health study.

Wave IV in-home interviews were conducted in 2008 and 2009 when the original Wave I respondents were 24 to 32 years old. Longitudinal survey data were collected on the social, economic, psychological, and health circumstances of respondents, as well as longitudinal geographic data. Survey questions were expanded on educational transitions, economic status and financial resources and strains, sleep patterns and sleep quality, eating habits and nutrition, illnesses and medications, physical activities, emotional content and quality of current or most recent romantic/cohabiting/marriage relationships, and maltreatment during childhood by caregivers. Dates and circumstances of key life events occurring in young adulthood were also recorded, including a complete marriage and cohabitation history, full pregnancy and fertility histories from both men and women, an educational history of dates of degrees and school attendance, contact with the criminal justice system, military service, and various employment events, including the date of first and current jobs, with respective information on occupation, industry, wages, hours, and benefits. Finally, physical measurements and biospecimens were also collected at Wave IV, and included anthropometric measures of weight, height and waist circumference, cardiovascular measures such as systolic blood pressure, diastolic blood pressure, and pulse, metabolic measures from dried blood spots assayed for lipids, glucose, and glycosylated hemoglobin (HbA1c), measures of inflammation and immune function, including High sensitivity C-reactive protein (hsCRP) and Epstein-Barr virus (EBV).

Wave V data collection took place from 2016 to 2018, when the original Wave I respondents were 33 to 43 years old. For the first time, a mixed mode survey design was used. In addition, several experiments were embedded in early phases of the data collection to test response to various treatments. A similar range of data was collected on social, environmental, economic, behavioral, and health circumstances of respondents, with the addition of retrospective child health and socio-economic status questions. Physical measurements and biospecimens were again collected at Wave V, and included most of the same measures as at Wave IV.

The overall goal of Wave VI was to better understand life course trajectories, determinants, and consequences of critical dimensions of aging, health, and health disparities among U.S. early midlife adults. Data collection took place from 2022 to 2025, with participants between the ages of 39 and 51, with an average age of 44. Beyond longitudinal survey measures, newly added questions included those on cumulative stress, discrimination, despair, work-life balance, memory, physical limitations, and caregiving. Continuing from previous waves, home exams collected physical measurements and biospecimens with most of the same measures as Wave V.

The National Longitudinal Study of Adolescent to Adult Health (Add Health) Parent Study Public Use collection includes data gathered as part of the Add Health longitudinal survey of adolescents. The original Add Health survey is a longitudinal study of a nationally representative sample of adolescents in grades 7-12 in the United States during the 1994-1995 school year. In Wave 1 of the Add Health Study (1994-1995), a parent of each Add Health Sample Member (AHSM) was interviewed. The Add Health Parent Study gathered social, behavioral, and health survey data in 2015-2017 from the parents of Add Health Sample members who were originally interviewed at Wave 1 (1994-1995). Wave 1 Parents were asked about their adolescent children, their relationships with them, and their own health.

The Add Health Parent Study interview is a comprehensive survey of Add Health parents' family relations, education, religious beliefs, physical and mental health, social support, and community involvement experiences. In addition, survey data contains cognitive assessments, a medications log linked to a medications database lookup table, and household financial information collection. The survey also includes permission for administrative data linkages and includes data from a Family Health History Leave-Behind questionnaire. Interviews were conducted with parents' spouse/partner when available.

Research domains targeted in the survey and research questions that may be addressed using the Add Health Parent Study data include:

• Health Behaviors and Risks Many health conditions and behaviors run in families; for example, cardiovascular disease, obesity and substance abuse. How are health risks and behaviors transmitted across generations or clustered within families? How can we use information on the parents' health and health behavior to better understand the determinants of their (adult) children's health trajectories?
• Cognitive Functioning and Non-Cognitive Personality Traits What role does the intergenerational transmission of personality and locus of control play in generating intergenerational persistence in education, family status, income and health? How do the personality traits of parents and children, and how they interact, influence the extent and quality of intergenerational relationships and the prevalence of assistance across generations?
• Decision-Making, Expectations, and Risk Preferences Do intergenerational correlations in risk preferences represent intergenerational transmission of preferences? If so, are the transmission mechanisms a factor in biological and environmental vulnerabilities? Does the extent of genetic liability vary in response to both family-specific and generation-specific environmental pressures?
• Family Support, Relationship Quality and Ties of Obligation How does family complexity affect intergenerational obligations and the strength of relationship ties? As parents near retirement: What roles do they play in their children's lives and their children in their lives? What assistance are they providing to their adult children and grandchildren? What do they receive in return? And how do these ties vary with divorce, remarriage and familial estrangement?
• Economic Status and Capacities What are the economic capacities of the parents' generation as they reach their retirement years? How have fared through the wealth and employment shocks of the Great Recession? Are parents able to provide for their own financial need? And, do they have the time and financial resources to help support their children and grandchildren and are they prepared to do so?

Understanding How Personal Networks Change: Wave 1 (UCNets) is the University of California Berkeley Social Networks Study, a longitudinal study with the objective of understanding how network composition changes over time as a result of life course transitions - e.g., graduation, marriage, retirement or widowhood - and how these changes are related to health status and outcomes. Using mostly addressed-based sampling from six San Francisco Bay Area counties, the study recruited participants in two age groups (cohorts) - 21-30 year-olds and 50-70 year-olds -- to maximize the possibility of experiencing a life transition. There are 3 waves of interviewing. The first interview, Wave 1, was May 2015 through January 2016. The re-interview for Wave 2 was conducted in Feb-June 2017, and Wave 3 was during Feb-May 2018. In Wave 1, the majority of respondents completed a face-to-face (FTF) survey, with the rest filling out a comparable web survey. In Wave 2 about half completed the FTF survey, and in Wave 3, about one-quarter. The survey contains items regarding households, personal networks, family milestones, employment, health status and behavior, personality and demographic characteristics.

The Wisconsin Longitudinal Study (WLS) is a long-term study of a random sample of 10,317 men and women who graduated from Wisconsin high schools in 1957. The WLS provides an opportunity to study the life course, intergenerational transfers and relationships, family functioning, physical and mental health and well-being, and morbidity and mortality from late adolescence through 2025. WLS data also cover social background, youthful aspirations, schooling, military service, labor market experiences, family characteristics and events, social participation, psychological characteristics and retirement.

Survey data were collected from the original respondents (the graduates) in 1957, 1964, 1975, 1992, 2004, 2011, 2020 and 2022; from their parents in 1957 and 1964; from a selected sibling in 1977, 1994, 2005, 2011, 2020 and 2022; from the spouse of the original respondent in 2004; and from the spouse of the selected sibling in 2006.

The collection of cognitive function data among graduates and siblings under the ILIAD study (MPIS: Sanjay Asthana, Michal Engelman, and Pamela Herd) began in 2020 to measure risk and resilience of Alzheimer's disease and related dementia (ADRD) as the cohort reaches their 80s. Data from 2020 and 2022 have been released, and 2025 and 2027 wave releases are forthcoming.

The WLS also has extensive administrative linkages, including with the 1940 and 1950 census, Social Security records, voting participation, and Medicare claims data.