DISC News contains articles about local, national, and international data issues.
It is published twice a semester by the library staff.
Editor: Joanne Juhnke, Special Librarian
Staff Contributors: Jack Solock, DISC Director
(Visit our PDF edition as well!)
How many people are uninsured in the USA? While the numbers trumpeted at townhalls and in news accounts may sound as though we have precise head counts, data users know better: it’s all about estimates from sample surveys, and it all depends on what survey you use.
There are several federal surveys that provide timely (within a year or two) data on the uninsured:
- Current Population Survey (CPS), Annual Social and Economic Supplement (ASEC)
- National Health Interview Survey (NHIS)
- Medical Expenditure Panel Survey (MEPS)
- American Community Survey (ACS)
On September 10, the Census Bureau released a report including the latest national estimates
on the uninsured from the 2008 CPS/ASEC (http://www.census.gov/prod/2009pubs/p60-236.pdf). The CPS is a monthly survey for collecting data on employment status; the Annual Social and Economic Supplement asks in March of each year about health insurance coverage, specifically regarding whether an individual has been uninsured for a full year. CPS data are widely used, with a sample size large enough to produce state-level estimates (across three years of data in the case of some smaller states.) The Kaiser Family Foundation uses their own analysis of CPS/ASEC data on the uninsured in their State Health Facts online (http://www.statehealthfacts.kff.org/). However, the CPS question asks people to think back an entire year; if some respondents actually give their current insurance status, the uninsured estimate may result in an overcount.
CPS ASEC data is available online at IPUMS-CPS at http://cps.ipums.org/cps/, and at DISC workstations through the CPS Utilities CD-ROM. (See story on CPS subscriptions, this issue).
Another source of data on the uninsured is the National Health Interview Survey (NHIS), conducted by the Census Bureau for the National Center for Health Statistics (http://www.cdc.gov/nchs/nhis.htm). The NHIS is a cross-sectional household interview survey with continuous sampling and interviewing throughout each year, focusing on a broad range of health topics among the US population, including access to health care. NHIS does quarterly “early release” reports on health insurance coverage 6 months after the quarter in question, most recently in July 2009 (http://www.cdc.gov/nchs/nhis/new_nhis.htm). The NHIS asks not only whether a person was uninsured for an entire year, but also if they were ever uninsured during a year, and whether they were uninsured at the time of the interview. The NHIS has too small a sample size, however, to provide state estimates other than in a few of the largest states. NHIS data can be found at the National Center for Health Statistics link above; at ICPSR (http://www.icpsr.umich.edu/icpsrweb/ICPSR/series/00040); and in harmonized form as the Integrated Health Interview Survey at http://www.ihis.us/ihis/.
The Medical Expenditure Panel Survey (MEPS), under the aegis of the Agency for Healthcare Research and Quality, uses a two-year panel for its household surveys, drawing a subsample of respondents to the NHIS. Since respondents are interviewed 5 times instead of just once, MEPS data can be used to analyze health insurance dynamics, i.e., how insurance coverage changes over time. The MEPS household surveys are supplemented by surveys of medical and health insurance providers. While the data are richer over time due to the panel component, the sample size is also smaller, leading to less reliable estimates for population subgroups and smaller geographical areas. The MEPS website is at http://www.meps.ahrq.gov/mepsweb/.
The latest addition to the field is the Census Bureau’s American Community Survey (ACS). The 2008 ACS includes a new point-in-time question on health insurance that had not been asked in the 2007 or previous questionnaires. The first ACS single-year estimates on the uninsured were released on September 22, 2009, for geographic areas with total populations of 65,000 or more. In 2010, the health insurance question will have been asked long enough to be included in ACS three-year estimates, for geographic areas with populations of 20,000 or more. While the ACS has the benefit of a large sample size, the insurance question only covers whether the person is insured at the time of the survey, with no information about the length of time uninsured or whether the person was uninsured at any time during the year. Both tabulated data and PUMS (Public Use Microdata Sample) data for the ACS are available from the Census Bureau’s American FactFinder site at http://factfinder.census.gov/; ACS PUMS data is also available at IPUMS USA at http://usa.ipums.org/usa/.
Additional sources and surveys for data on the uninsured include:
- The Small Area Insurance Health Insurance Estimates (SAHIE) - see Crossroads Corner
- The Survey of Income & Program Participation (SIPP): http://www.census.gov/sipp/
- National Survey of America’s Families: http://www.urban.org/center/anf/nsaf.cfm and through ICPSR
- Community Tracking Survey: http://www.hschange.com/index.cgi?data=11 and through ICPSR
The Data and Information Services Center’s Current Social Science Research Reports (CSSRR) has recently changed its format. Originally produced as three weekly e-mail newsletters covering the latest Internet releases in Economics, Sociology, and Health, the service has now been transformed into a weblog, at http://www.disc.wisc.edu/reports/cssrindex.html.
On that page, users can find an explanation for the change as well as a link to the weblog. In its new form, CSSRR is updated constantly throughout the day as we become aware of new items so that users can look in whenever they please, or subscribe to CSSRR as an RSS feed. For users who are more comfortable with the old indexed report format, the index categories are available on the right side of the page. The reports can also be searched or browsed chronologically. For those who only wish to look at the blog periodically, we are still keeping an e-mail subscription list for weekly notifications that the blog has been updated.
The weblog is more flexible for users, who can now control how the information is organized, and is also an easier way for us to present the information.
We are always interested in feedback on the form or content of the CSSRR weblog. Please send any feedback you might have to firstname.lastname@example.org.
UW-Madison’s membership in the Inter-university Consortium for Political and Social Research (ICPSR) provides the campus community with access to an immense archive of social science data for downloading and online analysis. Much has happened at ICPSR since the previous issue of DISC News. The ICPSR website, at http://www.icpsr.umich.edu/, has a new look and new features, which are detailed on the ICPSR blog at http://icpsr.blogspot.com/2009/07/visual-design-and-navigation.html.
Upcoming ICPSR Webinars
On October 8 and 9, 2009, ICPSR will be hosting webinars on teaching with data and quantitative literacy. On October 8 the topics will be Using Data in Teaching (11am-1pm); Delivering Research Opportunities to Undergraduates (1pm-2pm); and Tools for Bringing Data into the Classroom (2pm-3pm). The October 9 topic is Quantitative Literacy: Assessment and Enhancement (12pm-2pm). These sessions are free and open to the public. For more information, visit http://icpsr.blogspot.com/2009/09/icpsr-webinars-on-using-data-in.html.
Social Science Variables Database (SSVD)
The Social Science Variables Database (SSVD) enables users to search for variables across ICPSR datasets. DISC News last reported on the SSVD in February 2004, when the database was a pilot project and held only 63 studies, a tiny fraction of ICPSR’s total holdings. The revised SSVD, at http://www.icpsr.umich.edu/ICPSR/ssvd/index.jsp, holds approximately 1300 studies, about 20 percent of ICPSR’s holdings excluding US Census data. The total number of searchable variables this represents is around 1.2 million.
A sample search on the term “uninsured” returned 280 results. The results list includes the variable name and label, the text of the question, and the title of the study in which the variable is found. A click on the variable name or label leads to a page about the variable with an unweighted frequency table and summary statistics, plus a link to the study home page and links for further variable searching within the study or series.
Multimedia Tutorials at ICPSR Data User Help Center
ICPSR has an increasingly broad array of user help options. One of the newer sources under the Help item on the top menu is the Data User Help Center, where users can find audio and video tutorials on topics from finding and downloading ICPSR data, to using ICPSR data in statistical software, to depositing data with ICPSR. Several of the tutorials can also be found at the ICPSRweb channel on YouTube: http://www.youtube.com/user/ICPSRWeb.
The DISC subscription to the CPS-Utilities / Current Population Survey March Supplement from Unicon will be discontinued as of the end of September 2009.
For an outline of other sources for CPS data, DISC News in September 2005 featured “Many Avenues to CPS Data” at http://www.disc.wisc.edu/pubs/Newsletters/sep05news.html. In recent years, IPUMS CPS at http://cps.ipums.org/ has emerged as a robust and current source for data from the ASEC/March Supplement, with the March 2009 data having been released at IPUMS in mid-September. Please feel free to contact DISC staff with any questions concerning this decision.
The April 2009 issue of DISC News mentioned the Census Project as a worthy source for information on the upcoming 2010 decennial census. Since then, the organization has announced a new weblog endeavor, the Census Project Blog, for “readers supporting an accurate 2010 Census.” The “policy and operational issues” that the blog addresses have so far addressed concerns about the Census in Schools initiative and the Census Barriers, Attitudes and Motivators Survey (CBAMS) which informs Census outreach and communications efforts. The Census Project Blog can be found at http://censusprojectblog.org/.
by Joanne Juhnke
Crossroads Corner highlights web sites recently added to the searchable Internet Crossroads in Social Science Data on the DISC web site.
Small Area Health Insurance Estimates
The Small Area Health Insurance Estimates (SAHIE) program from the U.S. Census Bureau, at http://www.census.gov/did/www/sahie/, provides state and county-level estimates of health insurance coverage status in the US. The program’s most recent data release was in August 2009, encompassing estimates for the year 2006. At the state level the estimates are categorized by age, sex, race, Hispanic origin, and income; while at the county level the categories are by age, sex, and income. The estimate models combine data from various sources, including the Annual Social and Economic Supplement of the Current Population Survey, Census 2000, the Census Bureau’s Population Estimates Program, aggregated federal tax returns, participation records for the Supplemental Nutrition Assistance Program, County Business Patterns, and Medicaid and Children’s Health Insurance Program (CHIP) participation records.
Though the 2008 American Community Survey began providing single-year estimates at the county level in September 2009 for counties with populations of 65,000 or more, the SAHIE remains the only program that estimates health insurance status for all US counties.
Gallup-Healthways Well-Being Index
The polling firm Gallup and health-management firm Healthways have teamed up to produce the Gallup-Healthways Well-Being Index, at http://www.well-beingindex.com/. Conceived as a 25-year initiative, the project is interviewing over 1,000 U.S. adults each day, asking questions designed to reflect how Americans evaluate their health, their work, and their lives in general. The index aims to become “the official measure for health and well-being” for the United States.
Component indices include life evaluation, emotional health, physical health, healthy behaviors, work environment, and basic access. The site carries national graphs of the indices since January 2008, monthly reports, and daily snapshot findings. The site also links to a companion resource hosted by America’s Health Insurance Plans (AHIP) that creates annual reports based on Well-Being index data for each of the Congressional Districts in the US.
Looking for Data on Health?
The Economic and Social Data Service (ESDS) of the United Kingdom has put together a set of resource pages called “Looking for Data on Health?” at http://www.esds.ac.uk/themes/health/. The site’s definition of “data on health” is intentionally broad and cross-disciplinary, referencing data at ESDS on issues such as child development, access to care, behavior and lifestyle, diet and nutrition, personal attitudes, government expenditure, and health policy. The pages include details on how to search and browse ESDS collections for health data; video tutorials for analyzing ESDS data online; and case studies for how ESDS data on health might be used. The site also links to various additional sources for data on health.
While some of the sources have an international component, the data for the most part focuses on the United Kingdom, consistent with ESDS’ mission as a national data service. Users from outside the UK can register to access much of the data in the ESDS catalogue at http://www.esds.ac.uk/aandp/access/online_form.asp (UW-Madison is among the choices in the drop-down menu of institutions). However, certain data such as the ESDS International macrodata and UK Census data are licensed for UK users only.