The Census Bureau collects data on disability primarily through the American Community Survey (ACS) and the Survey of Income and Program Participation (SIPP). The definitions of disability are not always alike so caution should be taken when making comparisons across surveys. Generally, the SIPP estimates of disability prevalence are broader and encompass a greater number of activities on which disability status is assessed. The ACS has a more narrow definition but is capable of producing estimates for states, counties, and metropolitan areas. Because the ACS has replaced the decennial long-form as the source for small area statistics, there is no disability data in the 2010 Census.

In addition to these recent data sources, the Census Bureau has also produced disability estimates from the 2000 Census, and the Current Population Survey Annual Social and Economic Supplement (CPS ASEC). Other Federal agencies also collect and report disability statistics. Depending on your needs, one survey may be more suitable than another.

During the past 40 years, the accepted definitions of disability have been changing along side shifting attitudes and perceptions about people with disabilities. Prior to the 1960s and 70s, the medical model dominated how disability was defined. For example, a person with leg paralysis would have been considered disabled based solely on his or her physical condition.

Today, the social model of disability supports a definition as the result of a complex interaction between a person. and his or her environment. The same person with leg paralysis may be considered to have a disability due to their physical impairment as well as the barriers in the environment that prevent full social participation.

As public perception of disability has changed over time, so have the goals of programs supporting people with disabilities. In the past, the emphasis was to provide support to people with disabilities primarily through cash benefits and other replacements to earned income. Today, the emphasis has shifted to supporting independence and promoting involvement in all aspects of society.

The Census Bureau and other federal agencies that collect data about individuals with disabilities face two primary challenges:

• The process of measuring a complex, multi-dimensional concept in a survey format is difficult.
• The constantly evolving concepts and perceptions of disability require survey professionals to continuously develop measurement approaches that adapt to new definitions.

The development of the international classification of functioning, disability, and health (ICF) by the World Health Organization reflects this new perspective. The ICF is meant to codify a classification of functioning, disability, and health "to provide a unified standard language and framework for the description of health and health-related state.” The ICF views disability as an umbrella term and does not provide a single way to determine disability status. Thus, it is important to state clearly what aspect of the disablement process is being captured by a specific set of questions.

• In the past, the emphasis was to provide support to people with disabilities primarily through benefits.
• Today the emphasis has shifted to supporting independence and promoting involvement in all aspects of society.

Because the definition of a disability varies, the collection of disability data depends on the purposes for which the data are being used and survey methodologies employed in collecting the information. General purpose surveys like the ACS use mail-based paper questionnaires, telephone and personal interviews to collect a “core” definition of disability. This set of questions is now being used in a number of other federal surveys such as the Current Population Survey (CPS), the National Health Interview Survey (NHIS), and the National Crime Victimization Survey (NCVS). This definition has also been selected as the standard for all U.S. Department of Health and Human Services sponsored data collection, per Section 4302 of the Affordable Care Act.

Because self-response about disability is sensitive to contextual factors, the same questions can produce different prevalence estimates in surveys. For instance, the questions in the CPS (a labor force survey) may produce different estimates than the same questions in the ACS because the context of the survey may prime respondents to consider their ability to do activities with regard to the over-arching theme of the survey.