|Sponsor:||The Urban Institute|
|Description:||The National Survey of Americas Families (NSAF) is part of The Urban Institutes Assessing the New Federalism project. Its purpose is to track the effects of recent federal policy changes decentralizing many social programs, and to provide a comprehensive look at the well-being of children and non-elderly adults. The survey provides quantitative measures of child, adult and family well-being in America, with an emphasis on persons in low-income families. The survey gathers data on economic, health and social characteristics of children and families in order to estimate well-being. Specific topics include: participation in government programs; employment; earnings and income; economic hardship; educational attainment; training; family structure; housing arrangements; health insurance coverage; access to and use of health services; health status; psychological well-being; participation in religious and volunteer activities; knowledge of social services; and attitudes about work, welfare, health care and childbearing. In 2002, interviews were conducted with more than 40,000 families, yielding information on more than 100,000 people. Earlier rounds of the survey were conducted in 1997 and 1999.|
|Relevant Policy Issues:||Measurement of Health Status, Factors Contributing to Measured Health Disparities, Income Status, Economic Assistance Program Participation Rates, Economic Opportunity, Housing Quality, and Measure of Child Well Being and Child Care arrangements.|
|Unit of Analysis:||Individual or family level.|
|Identification of AI/AN/NA:||What is (your/NAMEs) race?
(Interviewers are instructed to probe by reading categories if necessary and if the respondent says Native American, to verify by asking: I am recording this as 'American Indian' is that right?)
The race question was identical in the 1999 study. In 1997, the race question did not include the follow-up probe for Native American.
|AI/AN/NA Population in Data Set:||There are four person-level data files available for public use. These files include the Adult Pair, Random Adult, Childless Adult, and Focal Child. Additionally, there are two family-level files that do not contain race information, but can be merged with the individual-level data files.
The 2002 Focal Child File contains one record for each child (ages 0 17) included in the NSAF study. Out of 34,332 completed interviews in this file, 491 respondents are identified as American Indian/Native American/Aleutian or Eskimo.
There are also three data files for the adult respondents included in the NSAF. In the survey, some questions are asked about both the respondent and his or her spouse or partner (if one exists), others are asked about either the respondent or his or her spouse or partner (randomly chosen), and still others are asked about only the respondent.
The 2002 Adult Pair File contains records for both the respondent and his/her spouse or partner. Out of 70,577 completed interviews in this file, 1,018 individuals are identified as American Indian/Native American/Aleutian or Eskimo.
The 2002 Random Adult File contains data elements from the extended interview that are specific to a randomly selected adult (this is a subset of the Adult Pair File). The random adult data set contains variables based on questions asked during the extended interview that were specific to a randomly selected adult (either the respondent or the spouse/partner). This situation occurs only in sections E (Past Year Health Insurance Coverage) and F (Health Care use and Access) of the NSAF questionnaire. Out of 49,507 completed interviews in this file, 752 respondents are identified as American Indian/Native American/Aleutian or Eskimo.
The 2002 Childless Adult File contains data elements representing households without children, where up to two childless adults between the ages of 18 and 64 were selected for interviewing. Out of 15,279 completed interviews, 248 respondents are identified as being American Indian/Native American/Aleutian or Eskimo. This file is not available on the NSAF Online Statistical Analysis webpage, but can be downloaded seperately at: http://anfdata.urban.org/drsurvey/login.cfm
|Geographic Scope:||Geographic indicator variables on the public use data files include a state indicator (including D.C.), the 5-digit Federal Information Processing Standards (FIPS) county code for counties with more than 250,000 persons in the 13 NSAF focal states, and the 4 Census regions of Northeast, Midwest (formerly North Central), West, and South.
The NSAF was designed to produce national estimates of the population under 65. Additionally, state estimates are possible for the 13 states that contained oversamples: Alabama, California, Colorado, Florida, Massachusetts, Michigan, Minnesota, Mississippi, New Jersey, New York, Texas, Washington, and Wisconsin. There is also a public use NSAF data set available for the California NSAF, conducted in 1997, 1999, and 2002 (containing 113 AI/AN individuals).
|Date or Frequency:||The 2002 study was conducted from February to November. Previous rounds of data collection include the 1997 and 1999 study. At this time there are no plans for future rounds of data collection.|
|Aggregation:||Researchers who wish to aggregate data across the three years of data collection for the NSAF study should first examine all survey items of interest to make certain the question text and the response options are identical. Although every effort was made to keep question wording unchanged between the rounds, there were some improvements made to some questions. Researchers should also examine the design of each wave very carefully. Changes were made in each wave that could affect survey weights and estimation; for example, in the third round of data collection, the sample size for nontelephone households in the study areas was reduced.
The number of AI/AN respondents for the 1997 and 1999 NSAF that could be included in an aggregation are below:
AI/AN Respondents in 1999:
AI/AN Respondents in 1997:
|Data Collection Methodology:||All interviews were conducted on the telephone by interviewers working in central interviewing facilities, using computer-assisted telephone interviewing (CATI) technology. In-person interviewers used cellular telephones to connect respondents in nontelephone households to the interviewing centers for the CATI interview.|
|Participation:||Optional, with incentives|
|Response Rate:||2002: overall child response rate = 55.1%
2002: overall adult response rate = 51.9%
1999: overall child response rate = 62.4%
1997: overall child response rate = 65.1%
|Sampling Methodology:||The sample is representative of the civilian, non-institutionalized population under age 65. As with the prior two rounds of data collection (conducted in 1997 and 1999), the 2002 survey included oversize samples drawn in 13 states (listed under Geographic Scope) to allow for the production of reliable estimates at the state level. The oversize state samples are supplemented with a balance of the United States sample to allow the creation of estimates at the national level as well.
The sampling frame consisted of a list-assisted, random-digit dialing (RDD) sample of telephone numbers supplemented by an area probability sample of nontelephone households. A short screening interview was used to identify and sample households based on age composition and household income. Once household eligibility was sampled, subsequent questions were asked to identify the children (age 0 to 17) or adults (age 18 to 64) in the household. Once this list was compiled, the CATI program sampled up to two children or up to two adults for subjects on the extended interview. If children were sampled, a series of questions was asked to determine the name and relationship of the person most knowledgeable about the selected child or children (the most knowledgeable adult).
|Analysis:||There are a series of methodology reports that accompany the NSAF public use data. Detailed information on calculating design effects (DEFF) and standard errors can be found in the following reports in the 2002 NSAF Variance Estimation, Report No. 4, located at http://www.urban.org/UploadedPDF/900716_2002_Methodology_4.pdf; and the NSAF Public Use File Users Guide, Report No. 11, located at http://www.urban.org/UploadedPDF/900760_2002_Methodology_11.pdf.
Child, all races: DEFF = 1.77, effective sample size = 10,828
|Strengths:||Data are collected on many key policy issues, including family well-being measures such as assistance receipt, socioeconomic status and education. There are multiple years of data available.
For most survey questions the item nonresponse rates were very low, often less than 1 percent. For survey items with significantly higher levels of item nonresponse (such as income), missing responses were imputed using a standard hot deck method, for all three rounds of data collection.
Low-income families were oversampled. Even with this oversampling method, the NSAF contains a relatively small average margin of error for state-level estimates of low-income children and adults.
|Limitations:||No major limitations were identified.|
|Access Requirements and Use Restrictions:||To access the public use data, researchers must register with the Urban Institutes website and agree to their terms of confidentiality. The data are available at no cost.|
|Contact Information:||The Urban Institute
Assessing the New Federalism Policy Center
2100 M Street, NW
Washington DC 20037
The public use data can be accessed: http://anfdata.urban.org/drsurvey/login.cfm.
Additionally, researchers can perform web-based analysis of the NSAF survey data using the online analysis tool at: http://www.urban.org/center/anf/analysisprelogin.cfm.
"report.pdf" (pdf, 1.56Mb)
"apa.pdf" (pdf, 107.76Kb)
"apb.pdf" (pdf, 88.97Kb)
"apc.pdf" (pdf, 233.07Kb)