AGE MEASUREMENT

In most countries, throughout the year following their seventh birthday a child who was asked "How old are you?" would say he or she is seven years old. In demographic terminology, this response represents the child's age last birthday or age in completed years. Other possible definitions of age include "age

TABLE 1

at the nearest birthday" and "age next birthday"; these definitions are no longer used in censuses or surveys. The term exact age is applied to the time elapsed since birth: an infant born on April 20, 2000, attains age 2.03 years on May 1, 2002. An East Asian practice of reckoning age involves assigning age one at birth and then increasing it by one at each subsequent New Year. The East Asian age, thus calculated using the lunar calendar (the lunar year is shorter than the solar year by a few days), may exceed the corresponding Western age by as much as three years for the elderly beyond the age of 70 years. Given the East Asian age, the animal year of birth, and whether the birthday falls between the New Year's day and the census or survey date, the Western age can be calculated. In situations where a direct question on age is unlikely to produce useful answers, a person's age may be calculated as the difference between the year of the census or survey and the reported year of birth.

The United Nations' recommendation for the 1970 round of population censuses was to use the following definition of age: "the estimated or calculated interval of time between the date of birth and the date of the census, expressed in completed years" (United Nations 1967, p. 41). The recommendation indicates that information on age can be collected by asking a direct question on age, one on the date of birth, or both.

Errors in Age Data

The frequency distribution of age may show irregularities (see Table 1), which may be real, reflecting past patterns of mortality (e.g., age-selective war casualties), fertility (e.g., a baby boom, birth heaping in auspicious years and deficits in inauspicious years), or migration, or may reflect errors in the data resulting from omission, multiple inclusion, or inaccuracy in reported age. Notice the tendency in Table 1 for the counts to peak at ages ending in 0 and 5, and to a lesser extent in ages ending in 2 and 8. Such heaping patterns reflect digit preference, the tendency to report ages ending in certain digits.

The pattern of digit preference varies among societies, and it also depends on the procedure used in collecting age data. The use of a question on date of birth, in combination with a direct question on age, tends to reduce such irregularities in the age data. When age is estimated as the difference between the census or survey year and the reported year of birth, age heaping occurs because of the preference shown for years ending in certain digits (such as 1900, 1910, 1920, and so on). Similar problems arise when age data are obtained by attempting to pinpoint the year of birth with reference to a list of historical events.

Age heaping has an inverse association with literacy level. The tendency to prefer ages ending in 0 and 5 is, however, widespread. Certain age preferences are no doubt culturally determined (e.g., the tendency found in some countries to avoid the number 13).

The term age shifting refers to deliberately giving an inaccurate age or date of birth. The elderly are prone to exaggerate their age, particularly if old age brings with it higher status; young men may under-state or overstate their age, if by doing so they stand to benefit in some way (e.g., avoid or, as the case may be, qualify for military duty).

Correction Methods Applicable to Age Data

The importance of obtaining accurate information on age stems partly from the fact that many demo-graphic features–such as reproductive behavior and geographic mobility–show distinct age patterns. Deficiencies in age data may lead to misleading patterns of such characteristics. The use of age groups for cross-classifications partially overcomes the problem.

In choosing age groups, it is advisable to have the "preferred" ages (displaying marked heaping) located toward the middle of each interval rather than at the end points. By centering the age groups at distribution peaks, the adjacent, depleted ages are brought within the interval.

If interest centers on the age distribution itself, rather than in cross-tabulations involving age as one of the variables, then techniques such as graduation can be used to produce age distributions smoothed over the observed irregularities.

Requirements for Improved Age Reporting

There are essentially three requirements for correct age reporting.

1. Availability of information: A respondent unsure of his or her age is unlikely to give the correct response to a query about age. In populations with complete birth registration, the age data tend to be more accurate, other things being the same. Also, the data are of better quality in population segments of higher literacy levels.
2. Use of appropriate methods of data collection: As mentioned, following a direct question on age by one on the date of birth may improve the data quality. Also, when the enumerators are free to adjust the responses (e.g., to hide digit preferences), the final age data may reflect to a large extent variation in the procedures used for adjustment.
3. Use of appropriate data processing and reporting procedures: The quality of the collected data may be excellent, but the reported age distribution still may be inaccurate, if, for example, only the year of birth, not the exact date of birth, is used at the tabulation stage.

See also: Data Assessment; Estimation Methods, Demo-graphic.

bibliography

Carrier, Norman, and John Hobcraft. 1973. Demo-graphic Estimation for Developing Societies. London: London School of Economics.

Prassat, Roland. 1985. The Dictionary of Demography, ed. Christopher Wilson. New York: Blackwell.

Scott, Christopher, and George Sabagh. 1970. "The Historical Calendar as a Method of Estimating Age." Population Studies 24: 93–109.

Shryock, Henry S., Jacob S. Siegel, and Associates. 1971. The Methods and Materials of Demography, Volume 1. Washington, D.C.: U. S. Government Printing Office.

United Nations. 1967. "Manual IV-Methods of Estimating Basic Demographic Measures for Incomplete Data." Population Studies 39.

——. 1967. "Principles and Recommendations for the 1970 Population Censuses." Statistical Papers, Series M, 44.

Krishnan Namboodiri