On this page

Specific Mortality
In subject area: Computer Science

Specific Mortality refers to death rates that are categorized

by factors such as age, sex, social class, race/ethnicity,
income, and area of residence for demographic, actuarial,
and public health analysis.

AI generated definition based on:

International Encyclopedia of the Social & Behavioral
Sciences, 2001

On this page

Chapters and Articles

You might find these chapters and articles relevant to this
topic.

Demographic Measurement:
Nuptiality, Mortality, Migration, and
Growth
M. Ní Bhrolcháin, in
International Encyclopedia of the Social & Behavioral
Sciences
, 2001

2 Mortality
The crude death rate (CDR)—deaths per 1,000 population—is
widely used as a basic measure of the level of mortality (see Table
2 for selected mortality measures). It is particularly strongly
influenced by the age composition of a population because of the
wide variation in death rates by age. Age-specific mortality rates—
deaths at a particular age per 1,000 population of that age—are at
the next level of detail, usually specific by sex. Where good quality
information is available, rates that are specific for single years of
age are produced, in preference to five-year age groups. Rates
specific by single years of age are of particular importance at the
younger and older parts of the age range, since in these age ranges
differences in rates between adjacent single years of age may be
sizable. Age may be grouped in a variety of ways, according to the
refinement required in analysis and the nature of available data.
However, even where five-year groups are used, the age groups
zero to one and one to four are often distinguished, because of the
relatively high death rate of infants, even in low mortality
countries.

Table 2. Selected mortality measures

Measure Definition Notes

Crude death rate 𝐷×1000 D=the number of deaths

𝑃

(CDR) Period occurring during a

particular year/period.
P=person years liveda during
the year or period.

𝐷
Age-specific death Mx= 𝑃 𝑥 Dx=deaths at age x during a
𝑥

rate Period; cohort ×1000 year or periodb. Px=person

years liveda at age x or in age
group x during a year or
periodb.

𝐷0
Infant mortality rate 𝐵
D0=deaths of infants under
(IMR) Period; cohort 12 months during a year or
periodb. B=births during a
year or periodb

𝐷0
Infant death rate 𝑃0
P0=person years liveda at
Period; cohort ages under one year during
a year or periodb.

𝐷𝑚
Maternal mortality 𝐵
Dm=maternal deaths during
ratio (sometimes a year or periodb.
referred to as the
maternal mortality
rate) Period; cohort

𝐷𝑚
Maternal mortality Pf15−49=woman years liveda
𝑃𝑓15 − 49
rate Period; cohort at ages 15–49 during a year
or periodb (age limits may
vary).

a
In specifying demographic rates, person-years lived,
woman-years lived and man-years lived (whether overall, or
at a particular age or in a given state) is the true
denominator. For practical purposes, person years lived in a
 period is estimated in various ways by different statistical
agencies. In the UK and USA, it is usually estimated by the
mid-year population (overall, or of a particular age etc.). See
also Demographic Techniques: Rates of the First and Second
Kind.
b
Where the rate relates to a year or period, events and person
years lived are those occurring during the year or period in
question. A cohort rate is based on events experienced and
person years lived by the cohort concerned.

The measurement of mortality is simplified to some extent by the

fact that death occurs ultimately to all and occurs only once.
These twin aspects mean that crude and age-specific death rates
are necessarily of type 1 (see Demographic Measurement: General
Issues and Measures of Fertility): the risk of death is universal and,
death occurring only once in a lifetime, nobody currently at risk of
death can have experienced the event previously. Death rates are
so closely and systematically associated with age, and to a lesser
extent sex, that specificity by age and sex is adequate for most
demographic purposes. Other dimensions of personal time are
required much less often in specifying death rates. However,
mortality rates may also be specific by social class, race/ethnicity,
income, area of residence (urban/rural or subnational region),
marital status, and other such socioeconomic characteristics.
Analysis by such factors is of interest in examining social and
economic variations in health and for actuarial and public health
purposes. With such subspecifications, death rates may also be of
type 2.

The infant mortality rate (IMR) is reported and published

extensively, partly as a demographic indicator but also as an
indicator of socioeconomic development. It is defined as the
number of deaths of infants (children under the age of one year)
per 1,000 births in a given year. It is, thus, not a rate as normally
specified since the numerator is not an estimate of the person-
years at risk of the event and will be inaccurate for this purpose to
the extent that there are year-on-year fluctuations in births. It is,
however, a convenient measure since it can be obtained from
simple counts of vital events and does not require population
estimates by age. It has a correctly specified counterpart in the
infant death rate, the number of deaths of infants (under the age of
one year) in a year per 1,000 person-years lived under age one.
Several mortality rates are distinguished also during the first year
of life, both because of the steep decline in death rates over the
first 12 months of life and because of the changing role of
endogenous (genetic, intrauterine, perinatal) and exogenous
(environmental, external) causes of death. Endogenous causes
predominate in the earliest postpartum period, with exogenous
factors growing in importance thereafter. The perinatal mortality
rate is defined as the number of late fetal deaths plus the number
of deaths within one week of birth per 1,000 total late fetal deaths
plus live births. The neonatal mortality rate is defined as the
number of deaths within one month (28 days) of birth per 1,000
live births, and the post-neonatal mortality rate as deaths between
28 days of birth and one year of age per 1,000 live births. The
probability of dying between birth and the fifth birthday is widely
referred to in the current demographic literature as the child
mortality rate, though the latter is more correctly defined as the
number of deaths of children under the age of five years per 1,000
person-years in the age group. The probability definition (also
sometimes termed child mortality risk) has become widespread
because under-five mortality is of particular interest in high-
mortality, less-developed societies, for which the indirect
methods (see Demographic Techniques: Indirect Estimation) used to
evaluate child mortality estimate a cohort probability rather than
a rate. High-mortality societies do not usually have good vital
registration statistics and it can be difficult to formulate accurate
assumptions about the average number of years lived by those
who die, and thus to obtain an estimate of person-years lived.

Maternal deaths are those that occur during pregnancy or within

42 days of the end of the pregnancy, due to a cause related to the
pregnancy or a condition aggravated by pregnancy; the term thus
includes abortion-related deaths. Maternal mortality is most often
represented by the maternal mortality ratio: the number of
maternal deaths per 100,000 live births in a period. This indicator
is referred to by many authors and in many sources as the
maternal mortality rate, although it is not in fact a true rate.
Strictly, the maternal mortality rate is the number of maternal
deaths per 100,000 women of reproductive age in a period, and
some authors employ the term in this sense. It should be clear
from the context which definition is in use.

Mortality specific by cause of death, of which maternal mortality

is an instance, is the principal way in which mortality indicators
are disaggregated beyond age and sex. The International
Classification of Disease, currently produced and revised from
time to time by the World Health Organization, is the standard
scheme by which cause of death is classified. Measures of
mortality by cause of death are of two kinds. Cause-specific
mortality rates refer to the number of deaths from a particular
cause during a period per 100,000 population, and may be
specific also by age and/or sex. A multiplier larger than 1,000 is
normally used for the purpose, since deaths from any one cause
will usually be relatively small in number. Cause-specific
mortality ratios, on the other hand, relate to the percentage of all
deaths that result from a particular cause, either overall or by age
and sex, and are of value where information on population
denominators is inaccurate or unavailable. Proportional mortality
analysis is used particularly in the study of occupational
mortality.

2.1 Age Standardization

Because of the substantial variation in death rates by age (and in
age structure across time and place), cross-national, areal, or time
comparisons of the level of mortality require that death rates are
standardized for age. Direct standardization, in which a standard
age distribution is employed, yields a directly standardized death
rate. Indirect standardization, in which a standard set of age-
specific mortality rates is chosen, yields two indicators: the
standardized mortality ratio (SMR) and the indirectly standardized
death rate. The SMR gives an index figure expressing the level of
mortality in the index population relative to that of the standard
population, set usually to 100: it is simply the ratio of observed to
expected deaths in the index population, with expected deaths
based on the standard population's age-specific rates. Multiplying
the SMR by the CDR in the standard population, gives the
indirectly standardized death rate in the index population. Direct
standardization requires that the age-specific mortality rates are
known and precisely estimated in the populations to be
standardized. Where the age-specific rates are either unknown or
subject to sizable errors, the indirect method must be used.
Statistical theory singles out a version of indirect standardization
as optimal when a multiplicative model holds for the force of
mortality.

2.2 Life Expectancy and Life Tables

Life expectancy at birth, e0, can be considered both an indicator of
the ‘timing’ of mortality and also a summary of the overall level of
mortality in a population. It measures timing in that it represents
the expected or average number of years lived by a person or
group subject, throughout their lifetime, to a given set of age-
specific mortality rates. Because death occurs universally and
removes an individual permanently from the population, any
increase in death rates at any age is necessarily reflected in a
shorter life expectancy, while a decline in death rates at any age is
necessarily accompanied by an increase in life expectancy. The
expectation of life is influenced both by the level of mortality in a
population and by the pattern of mortality by age. As a result, two
populations with the same overall life expectancy may differ in
their patterns of age-specific death rates. Life expectancy is
calculated from a life table (see Life Table), and so is standardized
for age. The life table in question may be based on the age-specific
mortality rates obtained in a calendar period or those of a birth
cohort. Where based on a period life table, the expectation of life
and other life table functions are synthetic cohort measures and
share the shortcomings of all synthetic multiplicative measures
(see Demographic Measurement: General Issues and Measures of
Fertility).

The life table is a central analytical technique in demography,

having a large variety of uses. It is, however, fairly demanding of
data since it requires information on age-specific death rates and
thus on the distribution of deaths by age as well as population
estimates by age. Because detailed data of this kind are often
unavailable for historical populations and for less developed
nations, sets of model life tables have been developed
corresponding to varying levels and patterns of mortality (Coale
et al. 1983, United Nations 1982). Such compilations allow
estimates of the level and/or age pattern of mortality to be made
from information that is insufficient in itself to determine these
(see Demographic Techniques: Indirect Estimation).

View chapter

Explore book

Read full chapter

URL: https://www.sciencedirect.com/science/article/pii/B0080430767047562

Mortality Crossover
I.T. Elo, in
International Encyclopedia of the Social & Behavioral
Sciences
, 2001

1 Empirical Evidence
One of the most widely noted mortality crossovers is that for
African Americans and whites in the United States. Throughout
the twentieth century, mortality estimates based on vital statistics
and census data have consistently shown black death rates to
exceed white rates until some age above the mid-seventies at
which point white rates exceed black age-specific death rates. The
greatest relative disadvantage for African Americans is observed
in middle age, and the age at which the two rates cross has
increased over time. Investigations of cause-specific mortality
have revealed a similar crossover in cause-specific death rates,
although the relative differentials by age and the age at which a
crossover occurs appear to vary by cause (Elo and Preston 1997,
Manton et al. 1979, Manton and Stallard 1997).
Evidence from linked data sources, where information for deaths
and population at risk comes from a single source, also suggests a
racial crossover in mortality. Kestenbaum (1992) has conducted a
careful analysis of mortality based on Medicare and Social
Security data for African and white Americans at ages 85 and
above. Black mortality exceeded or was the same as white
mortality up to age 87 for males and up to age 88 for females, at
which ages the rates crossed over. Similarly, mortality estimates
based on the National Longitudinal Mortality Study and small-
scale, longitudinal data sources show a crossover in black and
white death rates at the oldest ages (Elo and Preston 1997).

A mortality crossover is also observed for Hispanics and non-

Hispanic whites in the United States, but at a younger age than for
blacks and whites. Hispanic Americans have higher mortality than
non-Hispanic whites at younger ages and lower mortality after
about age 40–45 with the proportionate gap between Hispanic
and white rates increasing with age. At least some of the Hispanic
advantage appears to be due to the exceptionally low mortality of
the foreign-born. Hispanic populations in Latin America are also
known for having unusually low death rates at older ages relative
to their level of mortality at younger ages (Elo and Preston 1997,
Dechter and Preston 1991).

Empirically observed mortality crossovers are not limited to the

United States. Coale and Kisker (1986) compared death rates in
selected life tables for developed and developing countries with a
similar life expectancy at age five. In each pair of life tables
considered, death rates at younger ages in the developing country
exceeded those of the developed country, while at the oldest ages
the relative differentials were reversed. Based on indicators of
data quality, the authors concluded that the observed crossovers
were due to poor data quality at the oldest ages. Nam et al. (1978)
investigated mortality crossovers by comparing life table death
rates for pairs of national populations where one population was
considered low and the other high income. Of the 1,035 pairs of
mortality curves examined, in only 16 percent of the original pairs
were the authors' findings consistent with their original
hypothesis with lower income population having higher mortality
at younger ages but lower mortality at older ages.

More recently, Manton and Vaupel (1995) have suggested that a

crossover in death rates at the very oldest ages also occurs
between white Americans and populations in England, France,
Sweden, and Japan. In Japan and the European countries, life
expectancy at birth is notably higher than in the United States.
Death rates, however, begin to converge around age 65, and above
age 80 the US rates appear to fall below those of the European
countries and Japan.
 View chapter

Explore book

Read full chapter

URL: https://www.sciencedirect.com/science/article/pii/B0080430767021276

Adult Mortality in the Less

Developed World
I. Timæus, in
International Encyclopedia of the Social & Behavioral
Sciences
, 2001

1 Introduction
Measures of mortality are required to calculate life expectancy.
Demographers use data on current and past mortality by age to
forecast mortality trends and produce population projections (see
Population Forecasts). Epidemiologists use statistics on cause-
specific mortality to investigate the etiology of diseases that kill
adults. Social scientists investigate socioeconomic inequalities in
adult mortality and behavior that influences adults' risk of dying.
Thus, data on adult mortality are important inputs into health
policy making and program evaluation, into actuarial work in the
public and private sectors, and into economic and social planning
more generally.

A life table provides a full description of mortality at all ages (see

Life Table). Various measures based on the life table have been
proposed as summary indices of adult mortality, including life
expectancy at 15 years or some other age chosen to represent the
onset of adulthood. However, to calculate life expectancy
accurately one needs reliable data on old age mortality. No such
data exist for most less developed countries. For this reason, and
to distinguish the death of working age adults from mortality in
old age, adult mortality is often measured using indices such as
life table survivorship or partial life expectancy that summarize
mortality between two ages representing the onset of adulthood
and old age respectively. In particular, the World Bank and World
Health Organization have adopted the life table probability of
dying between exact ages 15 and 60 (denoted 45q15) as their
preferred measure of adult mortality (World Bank 1993, WHO
1999).

View chapter
 Explore book

Read full chapter

URL: https://www.sciencedirect.com/science/article/pii/B0080430767021264

Demographic Measurement:
Nuptiality, Mortality, Migration, and
Growth
M. Ní Bhrolcháin, in
International Encyclopedia of the Social & Behavioral
Sciences
, 2001

2.1 Age Standardization

Because of the substantial variation in death rates by age (and in
age structure across time and place), cross-national, areal, or time
comparisons of the level of mortality require that death rates are
standardized for age. Direct standardization, in which a standard
age distribution is employed, yields a directly standardized death
rate. Indirect standardization, in which a standard set of age-
specific mortality rates is chosen, yields two indicators: the
standardized mortality ratio (SMR) and the indirectly standardized
death rate. The SMR gives an index figure expressing the level of
mortality in the index population relative to that of the standard
population, set usually to 100: it is simply the ratio of observed to
expected deaths in the index population, with expected deaths
based on the standard population's age-specific rates. Multiplying
the SMR by the CDR in the standard population, gives the
indirectly standardized death rate in the index population. Direct
standardization requires that the age-specific mortality rates are
known and precisely estimated in the populations to be
standardized. Where the age-specific rates are either unknown or
subject to sizable errors, the indirect method must be used.
Statistical theory singles out a version of indirect standardization
as optimal when a multiplicative model holds for the force of
mortality.

View chapter

Explore book

Read full chapter

URL: https://www.sciencedirect.com/science/article/pii/B0080430767047562
Demography
John R. Weeks, in Encyclopedia of Social Measurement, 2005

Measuring Mortality
Demographers generally are interested in more than just the total
count of vital events such as birth and death, and a variety of rates
have been developed to evaluate the risk of death. The ability to
measure accurately varies according to the amount of information
available, and as a consequence the measures of mortality differ
considerably in their level of sophistication. The least
sophisticated and most often quoted measure of mortality is the
crude death rate.

Crude Death Rate

The CDR is the total number of deaths in a year divided by the
average total population. In general form:

CDR = 𝑑𝑝 × 1000 (6)

where d represents the total number of deaths occurring in a

population during any given year, and p is the total average (mid-
year) population in that year. It is called crude because it does not
take into account differences by age and sex in the likelihood of
death. Nonetheless, it is frequently used because it requires only
two pieces of information, total deaths and total population,
which often can be estimated with reasonable accuracy, especially
in developing countries where the cost of censuses and vital
registration systems may limit their availability.

Differences in the CDR between two countries could be due

entirely to differences in the distribution of the population by age,
even though the chance of dying is actually the same. Thus, if one
population has a high proportion of old people, its crude death
rate will be higher than that of a population with a high
proportion of young adults, even though at each age the
probabilities of death are identical. For example, in 2000 Mexico
had a crude death rate of 4 per 1000, scarcely one-third of the 11
per 1000 in Lithuania in that year. Nonetheless, the two countries
actually had an identical life expectancy at birth of 72 years. The
difference in crude death rates is accounted for by the fact that
only 5% of Mexico's population was aged 65 and older, whereas
the elderly accounted for 13% of Lithuania's population. Mexico's
crude death rate was also lower than the level in the United States
(9 per 1000 in 2000). Yet in Mexico a baby at birth could expect to
live 5 fewer years than a baby in the United States. The younger
age structure in Mexico puts a smaller fraction of the population
into the older high-risk ages, even though the actual probability of
death at each age is higher in Mexico than in the United States.
Age- and Sex-Specific Death Rates
To measure mortality at each age and for each sex we must have a
vital registration system (or a large survey) in which deaths by age
and sex are reported, along with census or other data that provide
estimates of the number of people in each age and sex category.
The age- and sex-specific death rate (nMx or ASDR) is measured as
follows:

𝑑 (7)
𝑀𝑥 = 𝑛 𝑥
𝑝 × 100, 000
𝑛 𝑛 𝑥

where ndx is the number of deaths in a year of people of a

particular age group in the interval x to x + n (typically a 5-year age
group, where x is the lower limit of the age interval and n is the
width of the interval in years of age) divided by the average
number of people of that age, npx in the population (again, usually
defined as the mid-year population). It is typically multiplied by
100,000 to get rid of the decimal point.

In the United States in 2000, the ASDR for white males ages 65–74
(denoted as 10M65) was 2955 per 100,000, whereas for white
females it was 1900 per 100,000, approximately two-thirds that
for males. In 1950 the ASDR for white males ages 65–74 was 4865
per 100,000, and for white females 3243. Thus, over the last half
of the twentieth century, the death rates for males ages 65–74
dropped by 39%, whereas for females the decline was 41%,
keeping women well ahead of men in terms of survival at these
ages. If the death data are sufficiently detailed, it is possible to be
even more specific in the calculation of rates. Thus, it is possible
to calculate rates that are, for example, specific to causes of death,
specific to marital status, or specific to racial/ethnic groups.

Age-Adjusted Death Rates

It is possible to compare CDRs for different years or different
regions, but it is analytically more informative if the data are
adjusted for differences in the age structure of the populations
prior to making those comparisons. The usual method is to
calculate age-specific death rates for two different populations
and then apply these rates to a standard population. For this
reason, this method is also known as standardization. The formula
for the age- adjusted death rate (AADR) is as follows:

AADR = ∑ 𝑛WS𝑥 × 𝑛𝑀𝑥 (8)

𝑖

where nwsx is the standard weight representing this age group's

proportion in the total population and nMx is the age-specific
death rate as calculated in formula (6). We can apply this
methodology to compare the CDR in Egypt in 1996 (6.19 deaths
per 1000 in population) with the United States in that same year
(8.70 deaths per 1000 in population). We use the United States
population as the standard weight and apply the age-specific
death rates for Egypt to the United States age-sex structure in
1996 to see what the CDR would be in Egypt if its age-sex
structure were identical to that in the United States. The result is
that the AADR for Egypt in 1996 was 15.08 deaths per 1000—
nearly twice that of the United States.
Life Tables
Another very useful way of standardizing for age is to calculate
the expectation of life at birth, also known as life expectancy. This
measure is derived from a life table, which was first used in 1662
by John Graunt to uncover the patterns of mortality in London.
Life expectancy can be summarized as the average age at death
for a hypothetical group of people born in a particular year and
being subjected to the risks of death experienced by people of all
ages in that year. An expectation of life at birth for U.S. females in
1999 of 79.4 years does not mean that the average age at death in
that year for females was 79.4. What it does mean is that if all the
females born in the United States in the year 1999 have the same
risks of dying throughout their lives as those indicated by the age-
specific death rates in 1999, then their average age at death will
be 79.4. Of course, some of them would have died in infancy
whereas others might live to be 120, but the age-specific death
rates for females in 1999 implied an average of 79.4. Note that life
expectancy is based on a hypothetical population, so the actual
longevity of a population is measured by the average age at death.
Because it is undesirable to have to wait decades to find out how
long people are actually going to live, the hypothetical situation
set up by life expectancy provides a quick and useful comparison
between populations. One of the limitations of basing the life
table on rates for a given year, however, is that in most instances
the death rates of older people in that year will almost certainly
be higher than will be experienced by today's babies when they
reach that age. This is especially true for a country that is in the
midst of a rapid decline in mortality, but even in the United States
in the twenty-first century, current life tables are assumed to
underestimate the actual life expectancy of an infant by 5 or more
years.

Life-table calculations, as shown in Table I for U.S. females for

1999, begin with a set of age- and sex-specific death rates, and the
first step is to find the probability of dying during any given age
interval. Table I is called an abridged life table because it groups
ages into 5-year categories rather than using single years of age.
The probability of dying (nqx) between ages x and x + n is obtained
by converting age/sex-specific death rates (nMx) into probabilities.
A probability of death relates the number of deaths during any
given number of years (that is, between any given exact ages) to
the number of people who started out being alive and at risk of
dying. For most age groups, except the very youngest (less than 5)
and oldest (85 and older), for which special adjustments are
made, death rates (nMx) for a given sex for ages x to x + n may be
converted to probabilities of dying according to the following
formula:

(𝑛)( 𝑀𝑥 )
𝑞𝑥 = 1 + (𝑎)(𝑛)(
𝑛 (9)
𝑀𝑥 )
𝑛 𝑛

This formula is only an estimate of the actual probability of death

because the researcher rarely has the data that would permit an
exact calculation, but the difference between the estimation and
the “true” number will seldom be significant. The principal
difference between reality and estimation is the fraction a, where
a is usually 0.5. This fraction implies that deaths are distributed
evenly over an age interval, and thus the average death occurs
halfway through that interval. This is a good estimate for every
age between 5 and 84, regardless of race or sex. At the younger
ages, however, death tends to occur earlier in the age interval. The
more appropriate fraction for ages 0–1 is 0.85 and for ages 1–4
(shown in Table I as 1–5) is 0.60. Another special case is the oldest
age group. Deaths tend to occur later rather than earlier in the
interval at the oldest ages. However, Table I uses 85+ as the oldest
age group and so avoids the specific age detail at the very oldest
ages. Note that because the interval 85+ is open-ended, going to
the highest age at which people might die, the probability of
death in this interval is 1.0000—death is certain.

Table I. Calculation of Life Expectancy for U.S. Females, 1999

hyp
(5)
Probabilities
of death
(4) Age- proportion (
specific of persons Nu
(2) Number (3) Number death alive at ali
of females of deaths rates in beginning begi
(1) Age in the in the the who die
interval, population, population, interval, during inte
x to x + n n Px n Dx nMx interval, nqx

0–1 1,867,649 12,291 0.00658 0.00654 10

1–5 7,383,117 2274 0.00031 0.00123 9

 hyp
(5)
Probabilities
of death
(4) Age- proportion (
specific of persons Nu
(2) Number (3) Number death alive at ali
of females of deaths rates in beginning begi
(1) Age in the in the the who die
interval, population, population, interval, during inte
x to x + n n Px n Dx nMx interval, nqx

5–10 9,741,935 1510 0.00016 0.00077 9

10–15 9,538,922 1593 0.00017 0.00083 9

15–20 9,587,530 3998 0.00042 0.00208 9

20–25 8,841,667 4244 0.00048 0.00240 9

25–30 9,150,709 5161 0.00056 0.00282 9

30–35 9,959,530 7629 0.00077 0.00382 9

35–40 11,332,470 13,123 0.00116 0.00577 9

40–45 11,231,542 19,015 0.00169 0.00843 9

45–50 9,855,838 24,817 0.00252 0.01251 9

50–55 8,447,622 32,498 0.00385 0.01905 9

55–60 6,692,991 41,443 0.00619 0.03049 9

60–65 5,546,089 54,812 0.00988 0.04822 9

65–70 5,110,451 79,166 0.01549 0.07457 8

70–75 4,909,038 118,514 0.02414 0.11384 7

75–80 4,272,506 163,496 0.03827 0.17463 7

80–85 3,003,063 194,124 0.06464 0.27824 5

85+ 2,934,837 436,152 0.14861 1.00000 4

Source: Death rate data from the U.S. National Center for
Health Statistics; other calculations by the author.

In Table I the age-specific death rates for females in 1999 in the

United States are given in column (4). In column (5) they have
been converted to probabilities of death from exact age x (for
example, 10) to exact age x + n (for example, 10 + 5 = 15). Once the
probabilities of death have been calculated, the number of deaths
that would occur to the hypothetical life-table population is
calculated. The life table assumes an initial population of 100,000
live births, which are then subjected to the specific mortality
schedule. These 100,000 babies represent what is called the radix
(l0). During the first year, the number of babies dying is equal to
the radix (100,000) times the probability of death. Subtracting the
babies who died (1d0) gives the number of people still alive at the
beginning of the next age interval (l1). These calculations are
shown in columns (7) and (6) of Table I. In general:

𝑑 = ( 𝑞𝑥 )(𝑙𝑥 ) (10)
𝑛 𝑥 𝑛

and

𝑙𝑥 + 𝑛 = 𝑙𝑥 − 𝑑𝑥 (11)
𝑛

The final two columns that lead to the calculation of expectation

of life are related to the concept of number of years lived. During
the 5-year period, for example, between the 5th and the 10th
birthdays, each person lives 5 years. If there are 98,000 people
sharing their 10th birthdays, then they all have lived a total of
5 × 98,000 = 490,000 years between their 5th and 10th birthdays.
Of course, if a person dies after the 5th but before the 10th
birthday, then only those years that were lived prior to dying are
added in. The lower the death rates, the more people there are
who will survive through an entire age interval and thus the
greater the number of years lived will be. The number of years
lived (nLx) can be estimated as follows:

𝐿 = 𝑛(𝑙𝑥 − 𝑎𝑛 𝑑𝑥 ) (12)
𝑛 𝑥

The fraction a is 0.50 for all age groups except 0 to 1 (for which
0.85 is often used) and 1 to 5 (for which 0.60 is often used).
Furthermore, this formula will not work for the oldest, open-age
interval (85+ in Table I) because there are no survivors at the end
of that age interval and the table provides no information about
how many years each person will live before finally dying. The
number of years lived in this group is estimated by dividing the
number of survivors to that oldest age (l85) by the death rate at
the oldest age (M85):

𝐿85 + = 𝑀8585
𝑙 (13)

The results of these calculations are shown in column (8) of Table

I. The years lived are then added up, cumulating from the oldest
to the youngest ages. These calculations are shown in column (9)
and represent Tx, the total number of years lived in a given age
interval and all older age intervals. At the oldest age (85+), Tx is
just equal to nLx. But at each successively younger age (e.g., 80 to
85), Tx is equal to Tx at all older ages (e.g., 85+, which is T85) plus
the number of person-years lived between ages x and x + n (e.g.,
between ages 80 and 85, which is 5L80). Thus, at any given age:

𝑇𝑥 = 𝑇𝑥 + 𝑛 + 𝐿𝑥 (14)
𝑛

The final calculation is the expectation of life (ex), or average

remaining lifetime. It is the total years remaining to be lived at
exact age x and is found by dividing Tx by the number of people
alive at that exact age (lx):

𝑒𝑥 = 𝑇𝑙 𝑥 (15)
𝑥

Thus for U.S. females in 1999, the expectation of life at birth (e0)
was 7,939,099/100,000 = 79.4, whereas at age 55 a female could
expect to live an additional 27.3 years.
Other Widely Used Death Rates
There are several other mortality rates that are routinely and
widely used in demographic research. The infant mortality rate
(IMR) is actually the first line of the life table, but it can be
calculated on its own. It is the number of deaths during the first
year of life (d0) per 1000 live births (b):

IMR =
𝑑0
× 1000 (16)
𝑏

The IMR is extremely sensitive to overall levels of

health and well-being and is highly correlated with life
expectancy, so on its own it provides a good index of a
population's mortality situation. As recently as the
nineteenth century, IMRs of 500 (meaning that one-
half of all babies born died before reaching their first
birthday) were not unheard of. As of the beginning of
the twenty-first century, the lowest rates were in Japan
and Sweden (3 per 1000) and the highest rates were in
several West African countries (such as Sierra Leone
with 153 per 1000).
Although IMR measures infant deaths from birth through the first
year of life, the most dangerous time for infants is just before and
just after birth. There are special measures of infant death that
take these risks into account. For example, late fetal mortality
refers to fetal deaths that occur after at least 28 weeks of
gestation. Neonatal mortality refers to deaths of infants within 28
days after birth. Postneonatal mortality, then, covers deaths from
28 days to 1 year after birth. In addition, there is an index called
perinatal mortality, which includes late fetal deaths plus deaths
within the first 7 days after birth. All of these values are typically
divided by the number of live births. Technically, they should be
divided by the number of conceptions in order to measure the risk
of various pregnancy outcomes, but the data on conceptions are
generally not available or are not very reliable even if they are
available.

Maternal mortality refers to those deaths that are associated with

complications of pregnancy and childbearing. The maternal
mortality ratio measures the number of maternal deaths per
100,000 live births. At the end of the twentieth century, the world
average was estimated by the World Health Organization to be
400 per 100,000. Another way of measuring maternal mortality
that takes into account the number of pregnancies that a woman
will have is to estimate a woman's lifetime risk of a maternal
death. As a woman begins her reproductive career by engaging in
intercourse, the question that is asked is: What is the probability
that she will die from complications of pregnancy and childbirth?
This risk represents a combination of how many times she will get
pregnant and the health risks that she faces with each pregnancy,
which are influenced largely by where she lives. For the average
woman in the world, that probability is 0.013 or 1 chance in 75,
but for women in sub-Saharan Africa, the risk is 1 in 11.

View chapter

Explore book

Read full chapter

URL: https://www.sciencedirect.com/science/article/pii/B0123693985002735

Demographic Measurement:
Nuptiality, Mortality, Migration, and
Growth
M. Ní Bhrolcháin, in
International Encyclopedia of the Social & Behavioral
Sciences
, 2001

4 Population Growth
The difference between the crude birth rate and the crude death
rate is known as the crude rate of natural increase (CRNI) and
represents the rate at which a population is growing or declining
in a year purely as a result of vital events. Since migration also
influences population growth, the overall population growth rate
is CBR–CDR±the crude net migration rate. Where population
growth rate is inferred from the change in total population during
a period of t years, the growth rate is estimated as r=ln (Pt/P0)/t.
The intrinsic rate of natural increase is a theoretical quantity,
referring to the growth rate of a stable population (see Population
Dynamics: Theory of Stable Populations) having a fixed set of age-
specific mortality and fertility rates. It can be calculated for a real
population, but should be interpreted with care since it
represents the growth rate that would be obtained in the
hypothetical circumstance that current age-specific rates were to
remain fixed and that the population were closed to migration.
The difference between the actual population growth rate and the
intrinsic rate of natural increase is between the observed growth
rate and the hypothetical growth rate that would apply if current
vital rates were to obtain in a stable population. Further indicators
relevant to the theoretical analysis of population growth are
outlined in the article Demographic Measurement: General Issues
and Measures of Fertility: the gross and net reproduction rates
(GRR and NRR; see also Population Dynamics: Theory of Stable
Populations). While the GRR and NRR may be calculated for actual
(empirical) populations, an important proviso must be borne in
mind: since the NRR accurately reflects growth prospects only in
populations that are stable (see Population Dynamics: Theory of
Stable Populations), when calculated for an actual population it
takes no account of prospective growth/decline resulting either
from migration nor does it take account of the growth potential
inherent in population age structure (see Population Dynamics:
Momentum of Population Growth). The same is true of the total
fertility rate (TFR). A TFR of 2.1 in low-mortality countries is
widely referred to in the demographic literature as ‘replacement
level fertility,’ that is, the level of fertility necessary for population
replacement in the long run. But this is potentially misleading.
Low-mortality populations with a TFR of below 2.1 can and do
continue to grow because of in-migration, declines in mortality,
and/or population momentum. Equally, low-mortality
populations with a TFR of 2.1 or more can and do decline because
of out-migration, changes in mortality, and/or population
momentum.

View chapter

Explore book

Read full chapter

URL: https://www.sciencedirect.com/science/article/pii/B0080430767047562

Paleodemography
James W. Wood, in Encyclopedia of Social Measurement,
2005

Demographic Nonstationarity
Another shortcoming of traditional paleodemographic life-table
analysis is that it assumes that the population under investigation
was stationary—that it was closed to migration and had an
intrinsic rate of increase equal to zero, age-specific schedules of
fertility and mortality that were unchanging over time, and an
equilibrium age distribution induced by those age-specific birth
and death rates. Only in this special case is the empirical age
distribution of skeletons expected to have a simple,
straightforward relationship to the cohort age-at-death column in
the life table.

As demographers have long realized, the age structure of a

nonstationary population (and thus the number of its members at
risk of death at each age) is more sensitive to changes in fertility
than to similar changes in mortality. Thus, age-at-death
distributions from different populations are at least as likely to
reflect fertility differences as genuine differences in mortality.
This incontrovertible fact of mathematical demography has given
rise to the odd notion that paleodemographic age-at-death
estimates are more informative about fertility than mortality. In
fact, all we can ever hope to estimate about fertility from such
data is the crude birth rate, which is scarcely a measure of fertility
at all. But if we could correct for demographic nonstationarity, we
could extract quite a bit of information about age-specific
mortality from skeletons, and perhaps even estimate the
population's growth rate.

Let f0(a) be the expected age-at-death distribution for a single

birth cohort in the target population. If the target population was
stationary, the same distribution holds for all deaths occurring in
the population. But even if we cannot take it for granted that the
population was stationary, it may be reasonable to assume that it
was stable. That is, we may be able to make all the assumptions
listed above for the stationary population, except allowing for the
possibility of a nonzero growth rate. (The assumption of stability
is much less restrictive than that of stationarity: even when
fertility and mortality rates are changing and migration is
occurring, most human populations still closely approximate a
stable age distribution at any given time.) In a stable but
nonstationary population, the age-at-death distribution is only
partly a function of age-specific mortality; it is also influenced by
the number of living individuals at risk of death at each age,
which is influenced in turn by population growth. More precisely,
the probability density function for ages at death in a stable
population with growth rate r is

−ra −ra
μ(𝑎)S(𝑎)𝑒 𝑓0 (𝑎)𝑒
𝑓𝑟 (𝑎) = ∞ −rx = ∞ −rx ,
∫ μ(𝑥)𝑆(𝑥)𝑒 dx ∫ 𝑓0 (𝑥)𝑒 dx
0 0

where μ(a) is the force of mortality and S(a) is the survival

function at age a. This expression also applies to all the skeletons
accumulated by a stable population over some more or less
extended span of time—for example, the period over which
skeletons were deposited in a particular cemetery. In principle,
then, we can treat fr(a) as the Pr(a|θ) function in the Rostock
protocol and estimate r as an additional parameter of the model, if
we can assume that the population was stable. If it was not stable,
at least approximately, we have probably reached the limits of
what we can ever hope to learn about age-specific mortality from
skeletal samples.

View chapter

Explore book

Read full chapter

URL: https://www.sciencedirect.com/science/article/pii/B0123693985003819

Mortality: The Great Historical

Decline
T. Bengtsson, in
International Encyclopedia of the Social & Behavioral
Sciences
, 2001

2 Explanations of the Mortality Decline

An understanding of the great mortality decline can be based on
two approaches, namely period and cohort explanations. One
should also make a distinction between factors that rely almost
entirely upon the actions of man and those beyond deliberate
human control.

2.1 Period Explanations

A widely accepted multifactorial explanation (United Nations
1953) is based on period factors that depend upon our activities
such as public health reforms, advances in medical knowledge,
improved personal hygiene, and rising income and standards of
living. This type of explanation is consistent with demographic
transition theory and with the theory of the epidemiological
transition.

The demographic transition theory was once one of the most

widely accepted ‘truths’ in social sciences and one of the main
pillars of the UN paradigm of how to promote developing
countries. It stated that mortality was high and fluctuated
strongly before the demographic transition. The transition was
said to begin with a decline in the level of mortality and with a
dampening of its fluctuations. This was regarded as a
consequence of a modernization of the economies (the
agricultural and industrial revolutions). A rapid increase in
population size would follow because fertility adjusted much later
due to a slow response in institutions and norms that influence
fertility (Davis 1945, Notestein 1953).

The idea of a stable level of mortality prior to modernization is a

central feature in the theory, but before the mid-eighteenth
century the mortality level was far from stable, as was shown for
England by Wrigley and Schofield in their classic book (1981).
Similar patterns had been demonstrated by many other scholars
as well, but only for individual parishes or institutions, or for
smaller regions like southern Sweden (Bengtsson and Oeppen
1993) or northern Italy (Galloway 1994). Since the mortality level
prior to the great mortality decline was not stable, it is not
appropriate to use it to date the beginning of the demographic
transition. Another finding that questions the demographic
transition theory is the long-term movement in fertility in the
past. Since such changes took place in the past, one cannot justify
the assumption that fertility norms changed slowly and kept
fertility high for many decades after the start of the demographic
transitions (Bengtsson and Ohlsson 1994).

On a broader front, McKeown (1976) questioned the multifactorial

explanation of the great mortality decline. He argued that a single
factor, namely better nutrition, could explain almost the entire
great mortality decline. His criticism was based on a study of
cause-specific mortality in England and Wales from 1838 to 1947,
where he observed that two-thirds of the mortality decline was
due to a reduction in infectious diseases. In later work he also
analyzed mortality rates and economic development for other
countries and further back in time, though not in such detail as for
England and Wales.

McKeown argued that medical advances had little influence on

mortality trends before the breakthrough of sulfonamides and
antibiotics in the 1930s and 1940s. In earlier periods the only
disease that one could cure was diphtheria. Very few deaths,
however, were due to diphtheria, and it was already retreating
when the antitoxin came in use around 1900. For periods before
1838, McKeown argued that inoculation and vaccination for
smallpox had little or no impact on the general mortality decline.
Vaccination started in England at the end of the eighteenth
century, but it was not widespread until after 1840, at which time
it became available at public expense. Others have drawn the
same conclusion for Sweden, where inoculation never became
widespread. Only about 50,000 individuals were inoculated
between 1750 and 1800 (Sköld 1997). If inoculation provided total
safety against death from smallpox, some 7,500 would have been
saved, assuming a fatality rate of 15 percent. This saving should be
compared with the 300,000 who died from smallpox in Sweden
during this period. Thus, even though inoculation saved some
lives and possibly slowed the spread of the disease, it was not a
major factor behind the initial mortality decline.

A similar story can be told about vaccination. In all the Nordic

countries, the smallpox mortality declined before vaccination
started in the first years of the nineteenth century and well before
vaccination became common in the 1820s. Even though the level
of mortality in smallpox was low already when vaccination
started, the disease continued to be fatal for many decades. There
were outbreaks in Finland and Sweden as late as in the 1880s,
most likely because not all children were vaccinated (for an
overview of the Nordic countries, see Bengtsson 1992). In
England, vaccination did not become widespread until the latter
part of the nineteenth century. Thus neither inoculation nor
vaccination was of major importance for the initial mortality
decline but it is likely that vaccination ensured that the disease
never came back once it had essentially been eradicated.

The improvement of personal hygiene may have had some effect

on mortality in England and Wales after about 1880, when a
decline in intestinal infections coincided with substantial
improvements in water supply and sewage control, which must
have reduced the incidence of waterborne infections. Since
changes in mortality from intestinal infections were only a
smaller part of the general mortality decline prior to, say, 1870,
the general reduction must have been due mostly to factors other
than improved personal hygiene, according to McKeown. For the
same reason the type of public health measures discussed above
must have had little impact on the great mortality decline prior to
1870 and only a partial effect afterwards. Other public health
measures, such as the breastfeeding campaigns that took place in
Sweden in the 1830s and later, were not discussed by McKeown.
Fridlizius (1984) has argued for Sweden that these campaigns
might very well have had an impact on infant mortality, but that
they started much later than the decline in mortality. Incidentally,
childhood mortality went up somewhat during the period when
breastfeeding campaigns started.

Helleiner argued in 1957 that the population of Western Europe

must have increased from the mid-eleventh century to the late
thirteenth century and from the mid-fifteenth century to the end
of the sixteenth century. The population increase observed in the
eighteenth century should therefore not be unique except that
mortality decline started from a higher level and went on longer
than before. Helleiner and others scholars have argued that these
changes in mortality were spontaneous (natural), and in 1973 the
UN added natural factors as a fifth determinant of the great
mortality decline. Later, others like Fridlizius (1984), Perrenoud
(1984), and Schofield (1984) have asserted that a change in the
virulence of pathogens initiated the great mortality decline. The
basic idea is that the virulence of pathogens changes
spontaneously over time. Virulence in an organism is generally
understood as its ability to overcome host defenses. At this point
it is important to note that some pathogens develop more quickly
in a malnourished host while others do not depend on such
circumstances to produce very high mortality. (Influenza
epidemics, for example, are neither created nor sustained in
malnourished persons, and they never were.)

McKeown's main argument on this issue, though not proved, is

that the initial mortality development during the latter part of the
eighteenth century was an integrated part of the great mortality
decline, which continued for the next two centuries and therefore
cannot be due to a spontaneous reduction in the virulence of
pathogens. According to McKeown, what is left is nutrition as an
explanation not only for the decline in infectious diseases from
the mid-nineteenth century onwards but also for the initial phase.
This reasoning is part of his attempts at finding one explanation
for the entire mortality decline.

In criticism, Fogel (1994) noted that McKeown only discussed the

nutritional intake, the diet, and not the claims from the body to
maintain itself and build up cells. Thus, McKeown only took into
account gross nutrition, not net nutrition, and the latter must be
more closely related to health and mortality. We will come back
to this issue shortly.

While Helleiner regarded the virulence hypothesis as a period

explanation outside of man's control, Kunitz (1986) did not agree.
He argued that the disappearance of epidemics, such as plague,
typhus, and smallpox, is associated with the growth of stable
governments, the expansion and integration of nation states, and
the emergence of disciplined armies in the seventeenth and
eighteenth centuries. These advancements made it possible to
control epidemics by means of isolation and better hygiene.
Population growth and loss of regional and local isolation then
converted epidemics into relatively more benign childhood
diseases. The mortality in the latter diseases varied depending on
nutritional status and the availability of adequate nursing, as did
mortality in the nonspecific pneumonia–diarrhea complex, which
became relatively more important when epidemic diseases
declined. Poverty, crowding, poor sanitation, and hazardous work
conditions became more important as determinants of death,
particularly in the cities. As a result, the eighteenth century saw a
divergence in the life expectancy in the upper and lower classes,
according to Kunitz. His empirical evidence for this conclusion is
rather weak since it is only based on a comparison of aggregate
infant mortality rates in England, France, Sweden, Germany, and
Spain, and with rough estimates of the standard of living in these
countries at the end of the eighteenth century.

The first part of Kunitz' explanation, namely the control of

epidemics by means of isolation and better hygiene, is much in
line with Easterlin's arguments. He argued that technology (in a
broad sense) is the main determinant of the great mortality
decline (Easterlin 1999). Easterlin argued that that the mortality
decline from the beginning of the nineteenth century onwards is
not due to economic growth and market developments but to
public policy initiatives and new institutions based on new
knowledge of disease. This could explain the simultaneous
mortality patterns among countries despite differences in
economic development.
2.2 Cohort Explanations
Cohort explanations of the mortality decline refer to features that
initially affect only certain young age groups but that may have
long-lasting effects for the groups involved. We mainly have
improvements in childhood conditions in mind, or even
conditions during the fetal stage that have lasting effects on
health and on the lifespan. Net nutrition is seen as the main
determinant of the development of cells, which is most rapid
during the fetal stage and which diminishes gradually until the
body is fully developed at an age around 20, or a little later. Net
nutrition is what is left for the development of cells after
nutritional claims needed for other life-supporting claims and
claims to sustain work are accounted for. A low net nutrition
could thus either be due to a low nutritional intake or to
additional claims coming from diseases. Also, many diseases not
only make their own claims on nutrients but also make it more
difficult for the body to absorb nutrients in general. This is the
case for infectious diseases. If cells and organs do not develop,
children may become stunted, become shorter in general and less
healthy. Thus we can basically differentiate between two types of
cohort explanations for the mortality decline, namely (a)
increased nutritional intake during the fetal stage and/or early
years of life, and (b) decreased claims during the fetal stage or
early years of life due to a reduction in diseases in the mother or
the child.

The importance of early childhood conditions for later life has

probably been well-known since time immemorial. The idea is
that each generation seems to carry along the same relative
mortality from childhood to old age. Both medical and historical
research on this matter has expanded rapidly over the last couple
of decades (for an overview, see Elo and Preston 1992).
In this connection, the work by Barker (1994 and elsewhere) has
been of major importance. He has summarized the medical
evidence for the importance of the nutrition of the fetus and
newborn infant for the health of the adult. Fogel (1996) has
advocated these ideas in historical research perhaps more than
anyone else. In addition, Preston and van de Walle (1978) for
urban France, and Fridlizius (1989) for Sweden, emphasize the
importance of cohort factors for the mortality decline.

Steckel (1983) and Fogel (1996 and previous work) use final body
height as a measure of net nutrition and health during childhood.
Individuals who had well-nourished and healthy mothers and
were well-nourished themselves during the fetal stage have a
lower risk of death during infancy. If they are well-nourished and
healthy their cells and organs develop better and they reach
greater heights and live a longer life. Since net nutrition and not
gross nutrition determines health there is no direct link between
gross nutrition and height. Improvements in health and height
may either be a result of better nutrition (better diet), or fewer
claims due to diseases. Thus a decline in the prevalence of
smallpox, for example, has a positive effect on height and extends
the lifespan, everything else being equal. The problem is that it is
difficult to evaluate how much of the improvement in health is
due to diet and how much is due to lower claims. Calculating diets
for premodern populations is a difficult task (Fogel 1996), and it is
even more difficult to calculate claims. Still, historical records
show similarities between trends in height and gross
consumption (Fogel 1994, 1996), indicating that trend in claims
has been of minor importance. If that is the case, then McKeown's
focus on gross nutrition can be justified.

Undernourishment, whether due to a low or badly composed

intake, or due to an increased claim because of diseases, may lead
to a stunting of height or weight and illness, diseases and
mortality later in life instead of increasing mortality in the
immediate future. The immediate relationship—i.e., the period
link between economy and mortality—is therefore much weaker
than what Malthus believed, according to Fogel. The rather weak
short-term relationship often found between prices and deaths
for many European countries, shown by Lee (1981, 1993),
Galloway (1988) and others, supports this interpretation (Fogel
1994). Thus cohort factors matter more to the mortality decline
than period ones, according to Fogel.

A warning against an uncritical use of cohort-mortality notions is

implied in the findings by Fridlizius (1989), who has shown that
the mortality of the Swedish cohorts born in the beginning of the
nineteenth century deviated greatly from what would have been
anticipated given their mortality at younger ages. He called them
deformed cohorts and saw the excess mortality for men as due to
high levels of alcoholism during the 1830s and 1840s. Such life-
styles lifted their mortality above what would otherwise be their
‘programmed’ level. Other analyses of aggregate cohort mortality
data also bring any strong cohort effect of conditions in early life
into question.

While Fogel was mainly concerned with chronic malnutrition, we

may also note that temporary disturbances in nutrition, in
particular during some periods at the fetal stage, may also have
long-lasting effects on health, since later improvements cannot
always compensate for prior loss (Barker 1994). Therefore,
temporary malnutrition due to lack of nutrients or to an increased
disease load, whether of mother or of child, may have long-lasting
effects on health and lifespan. It has, for example, been shown for
Sweden that the disease load in the first year of life has a strong
impact on mortality from airborne infections and other diseases
in later life (Bengtsson and Lindström 2000). Consequently,
elimination of temporary malnutrition, which is still a major
problem in many countries of the world, may cause another long-
term cohort decline in mortality.
2.3 Mortality Fluctuations
Another feature of the great mortality decline, as observed by
demographers, is that the mortality fluctuations became much
smaller in the nineteenth century than they were previously. Its
consequences for the balance between the population and the
economy in the preindustrial society, and also its effects on the
mortality decline, have been controversial. There are two schools
of thought, and they have reached very different conclusions.
According to those who analyze mortality crises (Meuvret 1946,
Goubert 1965), the reduction in such crises was a major factor
behind the secular decline in mortality. This view gained support
during the 1960s and 1970s. Wrigley and Schofield (1981)
showed, however, that the elimination of crises accounted for
only a small fraction (about 10 percent) of the secular decline in
mortality in their analysis based on a large national sample of
English parishes. Many others also obtained similar results (see
Ortega Osona 2000 for an overview). This view is now questioned
by Ortega Osona, who has found that the conclusion of Wrigley
and Schofield holds true for England but not for Northern Italy
and New Castile, where the reduction of the variability in
mortality is important for the initial decline in mortality.

View chapter

Explore book

Read full chapter

URL: https://www.sciencedirect.com/science/article/pii/B0080430767021161

Population Dynamics: Classical

Applications of Stable Population
Theory
T.W. Pullum, in
International Encyclopedia of the Social & Behavioral
Sciences
, 2001

4 Applications that Require Schedules of

Rates
A number of analyses of population dynamics have resulted from
examining the sensitivity of population characteristics to
disturbances of a stable population. These applications require
knowledge of the age-specific schedules for mortality (say mx for
each age x), of fertility (say fx for each x), and for any other
transitions in the model. For example, Preston (1974) took partial
derivatives of the basic stable population formulas to estimate the
effect of various age-specific mortality changes upon the intrinsic
growth rate, birth rate, and age composition. He found that the
effects could be represented as simple functions of the age
pattern of the cumulative changes in the mortality rates. Probably
the best known of these sensitivity analyses has been on the topic
of population momentum, which in its simplest form measures
the eventual effect on population growth of an immediate shift to
replacement fertility (see Population Dynamics: Momentum of
Population Growth).

The classical stable population model has no in- or out-migration,

but several authors have used the model in innovative ways to
evaluate the impact of emigration (Keyfitz 1971) or immigration
(Lesthaeghe et al. 1988). Mitra (1990) considered how a
population could maintain stationarity through a combination of
below-replacement fertility and a steady stream of immigration.
With assumptions about the ages and fertility of the immigrants,
it is possible to determine the necessary size of the stream of
immigration. This model is relevant, at least as a first
approximation, to the current mix of below-replacement fertility
and immigration in Europe and North America.

The classical formulation of stable population theory only allows

individuals to be born, procreate, and die, but later developments
permit individuals to move between several statuses or states
during their lifetime, as in multistate models (see Multi-state
Transition Models in Demography). In one obvious application the
states represent regions of a territory (Feeney 1970). In another,
Keyfitz developed a linkage between aging and career mobility in
stable populations (1973; the title of this paper is misleading
because he included stable as well as stationary populations). He
supposed that advancement in status across a set of discrete
categories is essentially based on the steepness of the age
distribution. In a setting of high growth, age (as a proxy for
experience) tends to produce earlier advancement because there
are relatively few people at higher levels. In a setting of low
growth, seniority is less valued, because there are relatively many
people at higher levels, and promotion will take longer. The
approach is again a type of sensitivity analysis, in which the age at
reaching a sequence of discrete gateways is expressed as a
function of stable population parameters (in particular, the
intrinsic growth rate).

View chapter

Explore book

Read full chapter

URL: https://www.sciencedirect.com/science/article/pii/B0080430767021070

Demographic Measurement:
Nuptiality, Mortality, Migration, and
Growth
M. Ní Bhrolcháin, in
International Encyclopedia of the Social & Behavioral
Sciences
, 2001

2.2 Life Expectancy and Life Tables

Life expectancy at birth, e0, can be considered both an indicator of
the ‘timing’ of mortality and also a summary of the overall level of
mortality in a population. It measures timing in that it represents
the expected or average number of years lived by a person or
group subject, throughout their lifetime, to a given set of age-
specific mortality rates. Because death occurs universally and
removes an individual permanently from the population, any
increase in death rates at any age is necessarily reflected in a
shorter life expectancy, while a decline in death rates at any age is
necessarily accompanied by an increase in life expectancy. The
expectation of life is influenced both by the level of mortality in a
population and by the pattern of mortality by age. As a result, two
populations with the same overall life expectancy may differ in
their patterns of age-specific death rates. Life expectancy is
calculated from a life table (see Life Table), and so is standardized
for age. The life table in question may be based on the age-specific
 mortality rates obtained in a calendar period or those of a birth
cohort. Where based on a period life table, the expectation of life
and other life table functions are synthetic cohort measures and
share the shortcomings of all synthetic multiplicative measures
(see Demographic Measurement: General Issues and Measures of
Fertility).

The life table is a central analytical technique in demography,

having a large variety of uses. It is, however, fairly demanding of
data since it requires information on age-specific death rates and
thus on the distribution of deaths by age as well as population
estimates by age. Because detailed data of this kind are often
unavailable for historical populations and for less developed
nations, sets of model life tables have been developed
corresponding to varying levels and patterns of mortality (Coale
et al. 1983, United Nations 1982). Such compilations allow
estimates of the level and/or age pattern of mortality to be made
from information that is insufficient in itself to determine these
(see Demographic Techniques: Indirect Estimation).

View chapter

Explore book

Read full chapter

URL: https://www.sciencedirect.com/science/article/pii/B0080430767047562

Related terms:
Mortality Risk, Cancer Patient,

Female Population, Stationary Population,

Transition Rate, Mortality Rate,

All content on this site: Copyright © 2024 Elsevier B.V., its licensors, and contributors. All rights are
Population Growth Rate,
reserved, including those for text and data mining, AI training, and similar technologies. For all open
access content, the Creative Commons licensing terms apply.

Stable Population Theory, Total Fertility Rate.

Recommended publications

Rehabilitation Robots for Neurorehabilitation in High-,

Low-, and Middle-Income Countries

Book • 2023
Computers & Mathematics with Applications

Journal

Computational Statistics & Data Analysis

Journal

Biologically Inspired Cognitive Architectures

Journal