# Total Fertility

## Definition

Total fertility is the mean number of children a woman would have by age 50 if she survived to age 50 and were subject, throughout her life, to the age-specific fertility rates observed in a given year. The total fertility is expressed as the number of children per woman.

## Method of Computation

Total fertility can be computed as the sum of age-specific fertility rates weighted by the number of years in each age group, divided by 1,000. Total fertility (TF) can be computed as follows:

Where fa is the age-specific fertility rate for women whose age corresponds to the five-year age group a. As a convention, the following seven five-year age groups are utilized: 15 to 19; 20 to 24; 25 to 29; 30 to 34; 35 to 39; 40 to 44; and 45 to 49.

## Data Sources

Data on total fertility can be obtained from three sources: civil registration systems, sample surveys and censuses. Civil registration systems are considered the best source of information on total fertility. However, some countries, especially those in the less developed regions, either lack a civil registration system or have a registration system whose coverage is too incomplete to be utilized for statistical purposes. Civil registration systems are considered to be complete if they cover 90 per cent or more of all live births taking place within a country or area.

In countries where civil registration systems are lacking, surveys and censuses can be utilized to estimate total fertility. Such information is generally obtained from questions regarding the number of live births that occurred in the household in the 12 months preceding the census or survey enumeration, questions on the date of birth of the last child born alive in the household or, in the case of surveys, retrospective birth histories. Compared to data from complete civil registration, such direct retrospective questions tend to yield less reliable estimates since they rely on the ability of individuals to recall with accuracy an event that took place several months or years before. Furthermore, while civil registration systems tend to generate annual estimates, the availability of survey and census data depends on the existence of adequate survey or census programmes. Censuses are generally conducted every 10 years. Surveys are undertaken at different intervals in different countries. In developing countries they typically take place every three to five years.

## Criteria for Estimate Selection

In World Fertility Data 2008 direct estimates produced by the national statistics office are given priority. When the direct estimates are unavailable, total fertility is computed based on available data (see also age specific fertility rates).

1. For civil registration systems, total fertility is obtained from country data reported by national statistical offices to the United Nations Statistics Division, the United Nations regional commissions or other regional statistical units. The data are also taken from reports and other analytical publications produced by national statistical offices.
2. For surveys, whenever the estimates are available in the survey report they are directly taken from the report. In other cases, if microdata are available, estimates are produced by the United Nations Population Division based on national data. The main surveys utilized are the Demographic and Health Surveys (DHS), the Reproductive Health Surveys (RHS), the World Fertility Survey (WFS), the Multiple Indicator Cluster Surveys (MICS), the Contraceptive Prevalence Surveys (CPS), and other nationally sponsored surveys.
3. For censuses, preference is given to data reported by national statistical offices to the United Nations Statistics Division. Data are also taken from census reports and other analytical publications produced by national statistics offices. Adjusted estimates are only used when reported by the national statistical office.

## Criteria for Selection and Reporting

Generally, only one source is provided per year for a country. When more than one source is available for the same period, preference is given to estimates based on civil registration. However, where such estimates are unavailable or incomplete, survey or census estimates are used. Registration data regarded as less than 90 per cent complete are used for countries where alternative sources are either not available or present problems of comparability, and where registration data can provide an assessment of trends. In countries with multiple survey programmes, sample surveys conducted on an annual or biennial basis are used where they exist. When such surveys are not available, other surveys using retrospective birth histories, censuses and other surveys are selected, in that order.

For each data point, information is provided on the source type (registration, census, survey or sample registration system) and the source (usually denoted by an acronym). A description of each acronym is given in the data file in the worksheet labeled "Sources". For surveys that are not part of an international survey programme, the full name is provided in the field "Survey name". Start and end years for the reference period covered by the data are reported when they differ from the calendar year.

Limitations depend on the data source utilized to produce the estimates:

1. For civil registration, estimates are subject to limitations that depend on the completeness of birth registration. Comparability of data is also affected by the treatment of infants born alive but who died before registration or within the first 24 hours of life, the quality of the reported information relating to age of the mother, and the inclusion of births from previous periods.
2. For survey and census data, the main limitations concern age misreporting, birth omissions, misreporting of the date of birth of the child and, in the case of surveys, sampling variability.
3. Notes are used to indicate any differences between the data presented and the definition of total fertility given above.

## Discrepancies

Generally, there is no discrepancy between estimates presented in World Fertility Data 2008 and the total fertility reported by national statistical offices. Estimates of total fertility reported in World Fertility Data 2008, however, may differ from those calculated by the United Nations Population Division and published in the World Population Prospects. The latter are best estimates based on all available demographic data. Total fertility rates presented in World Fertility Data 2008 are direct estimates from country data.

## Treatment of Missing Values

There is no attempt to provide estimates when country data are not available.

## Data coverage and periodicity

World Fertility Data 2008 contains data on total fertility for 224 countries or areas of the world. Data are provided, where available, for five reference dates: the closest years to 1970, 1985 and 1995, and the two most recent available years after 1999. Data are available for five reference dates for 169 countries or areas.