|Data Source:||U.S. Census Bureau|
Universe: Housing units
|Excerpt from:||Social Explorer, U.S. Census Bureau; 2000 Census of Population and Housing, Summary File 3: Technical Documentation, 2002.|
|Summary File 3 Technical Documentation -> Appendix C. Data Collection and Processing Procedures -> Glossary -> Imputation|
When information is missing or inconsistent, the Census Bureau uses a method called imputation to assign values. Imputation relies on the statistical principle of "homogeneity," or the tendency of households within a small geographic area to be similar in most characteristics. For example, the value of "rented" is likely to be imputed for a housing unit not reporting on owner/renter status in a neighborhood with multiunits or apartments where other respondents reported "rented" on the census questionnaire. In past censuses, when the occupancy status or the number of residents was not known for a housing unit, this information was imputed.
Internet Questionnaire Assistance (IQA)
An operation which allows respondents to use the Census Bureau's Internet site to (1) ask questions and receive answers about the census form, job opportunities, or the purpose of the census and (2) provide responses to the short form.
Interpolation frequently is used in calculating medians or quartiles based on interval data and in approximating standard errors from tables. Linear interpolation is used to estimate values of a function between two known values. Pareto interpolation is an alternative to linear interpolation. In Pareto interpolation, the median is derived by interpolating between the logarithms of the upper and lower income limits of the median category. It is used by the Census Bureau in calculating median income within intervals wider than $2,500.
|Summary File 3 Technical Documentation -> Appendix B. Definitons of Subject Characteristics -> Housing Characteristics -> Year Structure Built|
The data on year structure built were obtained from answers to long-form questionnaire Item 35, which was asked on a sample basis at both occupied and vacant housing units. Year structure built refers to when the building was first constructed, not when it was remodeled, added to, or converted. For housing units under construction that met the housing unit definition-that is, all exterior windows, doors, and final usable floors were in place-the category "1999 or 2000" was used for tabulations. For mobile homes, houseboats, RVs, etc., the manufacturers model year was assumed to be the year built. The data relate to the number of units built during the specified periods that were still in existence at the time of enumeration.
Median year structure built
Median year structure built divides the distribution into two equal parts: one-half of the cases falling below the median year structure built and one-half above the median. Median year structure built is computed on the basis of a standard distribution (see the "Standard Distributions" section under "Derived Measures"). Median year structure built is rounded to the nearest whole number. Median age of housing can be obtained by subtracting median year structure built from 2000. For example, if the median year structure built is 1967, the median age of housing in that area is 33 years (2000 minus 1967). (For more information on medians, see "Derived Measures".)
Limitation of the data
Data on year structure built are more susceptible to errors of response and nonreporting than data on many other items because respondents must rely on their memory or on estimates by people who have lived in the neighborhood a long time.
Data on year structure built were collected for the first time in the 1940 census. Since then, the response categories have been modified to accommodate the 10-year period between each census. In the 1980 census, the number of units built before 1940 appeared to be underreported. In an effort to alleviate this problem, a "Don't know" category was added in 1990. Responses of "Don't know" were treated like blanks and the item was allocated from similar units by tenure and structure type. However, this led to an extremely high allocation rate for the item (28 percent). A 1996 test proved inconclusive in determining whether a "Don't know" category led to a more accurate count of older units, but the test showed the allocation rate for this item was greatly reduced by the elimination of the "Don't know" category. As a result, "Don't know" was deleted for Census 2000.