Life expectancy is calculated as the number of years a person is expected to survive based on the statistical average.

The starting point used to calculate life expectancy is age-specific death rates among members of the relevant population. When plenty of data are available for a population, these age-specific death rates can be calculated simply by looking at the mortality rates for a given age. However, this method does not take into account any of the random statistical fluctuations between years of age and, today, several methods are used to adjust the data for these fluctuations such as the following:

• A mathematical formula such as an extended version of the Gompertz function is applied.
• When small amounts of data are available, a mortality table drawn up form a larger population can be adjusted to fit the data, by multiplying by a constant factor, for example,
• For large data sets, “smoothing” is applied to the mortality rates already experienced at given ages, by using cubic spines, for example.

The data needed to predict life expectancy in humans is relatively easy to identify compared with acquiring the data for industrial products or wild animals, for example. For wild animals, life expectancy can often only be calculated by capturing, marking and recapturing the animals. For long-lived products such as the components used in aircrafts for example, the life expectancy is calculated based on models such as accelerated aging.

To calculate the age-specific death rates, different data groups that are believed to be associated with different mortality rates (such as smokers versus non-smokers, for example) are considered separately. The data are then used to draw up a life table or actuarial table. These tables can be used to predict how likely it is that a person of a given age will die before their next birthday. From here, several points can be calculated, including:

• The person’s probability of surviving to any given age
• The life expectancy remaining for people of various ages