This measure divides the age distribution into two equal parts: one-half of the cases falling below the median value and one-half above the value. Median age is computed on the basis of a single-year-of-age distribution using a linear interpolation method.

There is some tendency for respondents to provide their age as of the date they completed the census questionnaire or interview, not their age as of the census reference date. The two-part question and editing procedures have attempted to minimize the effect of this reporting problem on tabulations. Additionally, the current census age question displays the census reference date prominently, and interviewer training emphasizes the importance of collecting age as of the reference date.

Respondents sometimes round a persons age up if they were close to having a birthday. For most single years of age, the misstatements are largely offsetting. The problem is most pronounced at age 0. Also, there may have been more rounding up to age 1 to avoid reporting age as 0 years. (Age in completed months was not collected for infants under age 1.) Editing procedures correct this problem.

There is some respondent resistance to reporting the ages of babies in completed years (i.e., 0 years old when the baby is under 1 year old). Instead, babies ages are sometimes reported in months. The two-part question along with enhanced editing and data capture procedures correct much of this problem before the age data are finalized in tabulations. Additionally, the current census age question includes an instruction for babies ages to be answered as 0 years old when they are under 1 year old.

Age heaping is a common age misreporting error. Age heaping is the tendency for people to overreport ages (or years of birth) that end in certain digits (commonly digits 0 or 5) and underreport ages or years of birth ending in other digits. The two-part question helps minimize the effect of age heaping on the final tabulations.

Age data for centenarians have a history of data quality challenges. The counts in the 1970 and 1980 Censuses for people 100 years and over were substantially overstated. Editing and data collection methods have been enhanced in order to meet the data quality challenges for this population.

It also has been documented that the population aged 69 in the 1970 Census and the population aged 79 in the 1980 Census were overstated. The population aged 89 in 1990 and the population aged 99 in 2000 did not have an overstated count. (For more information on the design of the age question, see the Comparability section below.)

Age data have been collected in every census. However, there have been some differences in the way they have been collected and processed over time. In the 2010 Census (as in Census 2000), each individual provided both an age and an exact date of birth. The 1990 Census collected age and year of birth. Prior censuses had collected month and quarter of birth in addition to age and year of birth. The 1990 Census change was made so that coded information could be obtained for both age and year of birth.

In each census since 1940, the age of a person was assigned when it was not reported. In censuses before 1940, with the exception of 1880, people of unknown age were shown as a separate category. Since 1960, assignment of unknown age has been performed by a general procedure described as imputation. The specific procedures for imputing age have been different in each census. (For more information on imputation, see 2010 Census: Operational Overview and Accuracy of the Data.)