Friday, December 18, 2009

American Community Survey (2003): National Population Statistics

As described in a previous post, I found that the Cornell University Disability Statistics website offered materials that seemed likely to help me estimate the number of people with disabilities (PWDs) in a particular county in the U.S.  The next step was to understand something about the general population terminology and concepts that formed the basis for census calculations.

I had learned that the American Community Survey (ACS) provided the most frequently updated and most recent county-by-county data.  The Guide to Disability Statistics from the American Community Survey (Weathers, 2005, p. 28) indicates that estimates of the numbers of PWDs depend directly upon the number of questions asked. According to Weathers (p. 4), the ACS approach has several limitations:

First, the ACS is limited to six questions that are used to identify the disability
population [as a whole] and it does not allow one to identify the prevalence of specific health conditions (e.g., cancer, paralysis, HIV/AIDS, etc.). Second, the ACS definition does not explicitly include important societal and environmental factors that may contribute to a disability such as discrimination and lack of reasonable accommodations. Finally, the ACS does not capture the population living in “group quarters.” Group quarters include individuals living in institutions, college dormitories, and other types of group quarters. This is a very important limitation in that it may leave out an important segment of the population with disabilities. The Census Bureau plans to address this last limitation of the ACS by including a sample of persons living in group quarters beginning in 2006.
By contrast, at the low end of the scale, according to the Guide to Disability Statistics from the Current Population Survey - Annual Social and Economic Supplement (March CPS) (Burkhauser & Houtenville, 2006, p. 19), the Current Population Survey asks just one question about disabilities -- regarding, specifically, the presence or absence of a work limitation -- and thus "misses a large part of the broader population with disabilities based on an ICF [i.e., International Classification of Functioning] disability conceptualization."  On the other extreme, as Burkhauser & Houtenville (2006, p. 19) also note, the Survey of Income and Program Participation (SIPP) uses 93 disability-related questions.  Generally, they say, "data sets that ask more questions to identify a population with disabilities and that contain a broader disability conceptualization will capture a larger disability population."

Weathers (2005) quantifies the difference made by these varying operationalizations of the concept of disability.  Weathers uses data from 2003.  To analyze these data, it may be helpful to begin by clarifying the meanings of some terms that the Census Bureau uses.

First, the resident population of the U.S. includes all U.S. residents living in the 50 states and the District of Columbia.  That is, it excludes residents of Puerto Rico and American possessions abroad, as well as U.S. citizens living abroad, whether in the military or as expatriates.  This post does not examine the question of whether PWDs may be disproportionately represented in the military -- whether, that is, the percentages of PWDs in their home counties would be noticeably higher (particularly in some age groups and locations) if those members of the Armed Forces were available for inclusion within the count of PWDs under the ACS.

The resident U.S. population as of July 1, 2003 was estimated to be 290,210,914.  According to the 2003 ACS Subject Definitions (p. 13), the resident population is divided into people who live in housing units (i.e., households) and those who live in group quarters.  "Housing units" can include any occupied "unit," including tents and railroad cars; but "If all the people staying in the unit at the time of the interview are staying there for two months or less, the unit is considered to be temporarily occupied, and classified as 'vacant'" (p. 14).  The data include transients (i.e., those who stay only briefly in any one place), however, by including them if there is no place where they usually stay.  Note that "Vacant units are excluded from the housing inventory if they are open to the elements; that is, the roof, walls, windows, and/or doors no longer protect the interior from the elements. Also, excluded are vacant units with a sign that they are condemned or they are to be demolished" (p. 14).  It preliminarily appears, then, that the 2003 ACS data undercount people who are not living in any kind of housing unit but, rather, are living outdoors (e.g., under a bridge) -- among whom, again, there may be a disproportionate number of PWDs.

The Census Bureau indicates that residents of group quarters were sampled in 1999 and 2001, but the 2003 data exclude people living in group quarters.  For purposes of introducing some relevant terms, then, it will be convenient to refer to the 2000 census.  According to the Statistical Abstract of the United States for 2003, table 77 (examining 2000 census data), about 7.8 million people (2.8% of the total resident population of 281,424,602 in 2000) lived in group quarters.  These people were divided almost equally into institutionalized and noninstitutionalized group quarters.  Among the institutionalized individuals in group quarters, by far the largest groups were males aged 18-64 in correctional institutions (44%) and females aged 65+ in nursing homes (29%).  Other institutions include mental hospitals and wards and juvenile facilities.  Among noninstitutionalized individuals in group quarters, 55% were living in college dormitories.  Other noninstitutional group quarters include military quarters, religious group quarters, shelters, and group homes. 

If 2.8% of the total resident population lived in group quarters in 2003, as was the case in 2000 (above), that would mean that just over 8 million (i.e., 2.8% of 290,210,914) lived in group quarters in 2003.  Subtracting those 8 million from the total resident population would produce something like the 282,909,885 people whom the Census Bureau estimated as the total household population in 2003.

The total household population included people of all ages.  In the analysis of disabilities done by Weathers (2005), however, the focus is on people who were between 18 and 64 years of age in 2003.  In the 2003 ACS, to use the slightly adjusted number provided by Weathers, the Census Bureau estimates that there were 176,395,446 people in this age group.  This number is quite close to the 176,305,337 estimated by the SIPP for that year, and roughly midway between the estimates of 172,761,000 from the NHIS and 179,132,544 provided by the CPS.

I will be using the ACS figure in the next step, which addresses the numbers and percentages of people with disabilities within that larger total.  At this point, I join Weathers (2005) in turning from the estimate of the total numbers of people with disabilities in the population, and focus (at least for the time being) upon the prevalence of disabilities among people of traditional working (i.e., post-high school, pre-retirement) age.