MeasuringMortalityFertilityAndNaturalIncrease5thEdition1994 (Pdfa)

Measuring Mortality,
Fertility, and
Natural Increase
A Self-Teaching Guide
to Elementary Measures
James A. Palmore
and
Robert W. Gardner
M
z
T h e U . S . Congress established the East-West C e n t e r i n 1960

to foster m u t u a l understanding and c o o p e r a t i o n a m o n g the
g o v e r n m e n t s and peoples of A s i a and the P a c i f i c region, i n c l u d i n g the U n i t e d States. P r i n c i p a l f u n d i n g f o r the C e n t e r
c o m e s f r o m the U . S . g o v e r n m e n t , w i t h a d d i t i o n a l support
p r o v i d e d by private agencies, i n d i v i d u a l s , and corporations,
and m o r e than 20 A s i a n and P a c i f i c g o v e r n m e n t s .
T h e C e n t e r p r o m o t e s responsible d e v e l o p m e n t , l o n g - t e r m
stability, and h u m a n d i g n i t y for a l l people i n the region and
helps prepare the U n i t e d States f o r c o n s t r u c t i v e i n v o l v e m e n t
i n A s i a and the P a c i f i c .
T h e Program o n P o p u l a t i o n c o n d u c t s research and offers professional e d u c a t i o n f o c u s i n g o n p o p u l a t i o n issues, w i t h e m phasis o n the analysis of d e m o g r a p h i c and h u m a n - r e s o u r c e
trends, their s o c i a l and e c o n o m i c causes a n d consequences,
and t h e i r p o l i c y i m p l i c a t i o n s i n A s i a , the P a c i f i c , a n d the
U n i t e d States. T o a c c o m p l i s h its goal and f u r t h e r the m i s s i o n of the East-West C e n t e r , the P r o g r a m cooperates w i t h
g o v e r n m e n t agencies, u n i v e r s i t i e s , and other o r g a n i z a t i o n s
t h r o u g h o u t the A s i a and P a c i f i c region and the U n i t e d States
and w o r k s c l o s e l y w i t h other programs of the C e n t e r .
East-West C e n t e r
1777 East-West R o a d
H o n o l u l u , H I 96848, U S A
Telephone: (808) 944-7145
Fax: (808) 944-7376
E - m a i l : EWCBOOKS@EWC.HAWAII.EDU
Measuring Mortality, Fertility, and

Natural Increase
Measuring Mortality, Fertility, and

Natural Increase
A Self-Teaching Guide to Elementary Measures
FIFTH EDITION
]ames A . Palmore and Robert W. Gardner
East-West Center
Honolulu
About
the
Authors
fames A. Palmore was a senior fellow in the East-West Center's Program

on Population and professor of sociology and population studies at the
University of Hawaii. He directed the university's Population Studies
Program from 1976 to 1991 and for 30 years taught graduate-level
courses in demography, survey design, and statistical analysis. He was
the author of numerous articles on Asian demography and coeditor of
several books, including Choosing
Asia and the United
States.
a Contraceptive:
Method
Choice in
Robert W. Gardner, who lives in Maine, teaches demographic methods at

Harvard University's School of Public Health and indexes demographic
books (including this one). For 20 years prior to 1992 he was a research
associate in the East-West Center's Program on Population and taught
courses in the Population Studies Program and the School of Public
Health at the University of Hawaii. During 1990-92 he served as
assistant director of the Program on Population. His publications include
Asian and Pacific Americans
in the United
States (with Herbert
Barringer and Michael Levin) and Migration Decision Making (edited

with Gordon F. De Jong).
1994 by East-West Center.
All rights reserved.
Second printing 1996.
Printed in the United States of America.
LIBRARY OF CONGRESS
CATALOGING-IN-PUBLICATION
DATA
Palmore, James A.
Measuring mortality, fertility, and natural increase : a selfteaching guide to elementary measures / James A. Palmore and Robert
W. Gardner. [Rev. ed.]
p.
cm.
Includes bibliographical references and index.
ISBN 0-86638-165-1 (pbk. : alk. paper) : $15.00
1. Demography. 2. Vital statistics. 3. MortalityStatistical
methods. 4. Fertility, HumanStatistical methods. I. Title.
HB849.4.P34 1996
304.6'01'5195dc20
94-36934
CIP
Published in 1994 by the East-West Center
1777 East-West Road, Honolulu, Hawaii 96848, U.S.A.
L i s t of figures and tables
Preface
vii
xi
1. Rates, ratios, percentages, and p r o b a b i l i t i e s
2. M o r t a l i t y
3. F e r t i l i t y , n a t u r a l increase, and r e p r o d u c t i o n rates
63
Appendixes
1. N o t a t i o n s and f o r m u l a s
131
2. R e l a t i o n s h i p between q and M values

x
3. A n s w e r s to selected exercises
135
139
4. C o u n t r i e s w i t h p o p u l a t i o n s of f e w e r t h a n 1 m i l l i o n
(1990 estimates or latest census)
References
Index
159
153
149
LIST O F F I G U R E S A N D T A B L E S
FIGURES
2.1.
A g e - s p e c i f i c death rates: G u a t e m a l a , 1985, and

Japan, 1989
3.1.
12
A g e - s p e c i f i c f e r t i l i t y rates: C o s t a R i c a , 1984,
G u a t e m a l a , 1985, Japan, 1989, and U n i t e d States,
1988
TABLES
75
1.1.
R a t i o s f r e q u e n t l y used i n d e m o g r a p h i c w o r k
1.2.
C a l c u l a t i o n of the n u m b e r of person-years l i v e d
d u r i n g one year i n a h y p o t h e t i c a l s m a l l t o w n h a v i n g
a p o p u l a t i o n of 700 persons o n January 1 and v e r y
erratic d e m o g r a p h i c b e h a v i o r
2.1.
H i g h e s t and l o w e s t crude death rates, by region:

recent years
2.2.
11
H i g h e s t and l o w e s t age-specific death rates, b y sex:

recent years
2.3.
13
A g e - s p e c i f i c and crude death rates f o r three h y p o thetical populations
2.4.
16
A g e - s p e c i f i c death rates and p o p u l a t i o n s f o r M a i n e

and S o u t h C a r o l i n a : 1930
2.5.
17
A g e s t a n d a r d i z a t i o n of crude death rates: Japan

(1989) and C h i l e (1989-90)
2.6.
23
F o r m u l a s for direct age-standardization of the c r u d e

death rate f o r t w o h y p o t h e t i c a l p o p u l a t i o n s
2.7.
Standardized death rates for selected places and

years
2.8.
25
26
Infant m o r t a l i t y rates for selected c o u n t r i e s , by age

and region: recent years
32
viii
2.9.
List of Figures and Tables
H i g h e s t and l o w e s t i n f a n t m o r t a l i t y rates, by region:

latest available data
34
2.10. C o m p l e t e l i f e table f o r females: U n i t e d States,

1979-81
38
2.11. A b r i d g e d l i f e table for females: U n i t e d States,

1988
48
2.12. A g e c o m p o s i t i o n of the stationary p o p u l a t i o n a n d

the actual p o p u l a t i o n for U . S . w h i t e females:
1988
52
2.13. C r u d e death rates and life-table death rates f o r U . S .

w h i t e females: selected years, 1900-88
53
2.14. S u r v i v o r s to exact age x(ej of 100,000 m a l e i n f a n t s

[t ): S r i L a n k a , 1920-81, and U n i t e d States, 1900-02
0
to 1988
55
2.15. E x a m p l e s of h i g h and l o w values of l i f e e x p e c t a n c y

at b i r t h (e ) for m a l e s and females: recent years
0
57
2.16. L i f e expectancies at selected exact ages f o r m a l e s and

females, by color: U n i t e d States, 1988
57
2.17. L i f e expectancies for w h i t e males and females

at exact ages 0, 40, and 70: U n i t e d States,
1850-1988
3.1.
58
H i g h e s t and l o w e s t crude b i r t h rates, by region:

recent years
3.2.
D i s t r i b u t i o n of countries b y l e v e l of crude b i r t h rate:

1980s
3.3.
68
Average crude rates of n a t u r a l increase, by region:

1985-90
3.4.
67
69
A p p r o x i m a t e n u m b e r of years a p o p u l a t i o n takes to
double, triple, and quadruple i n size, g i v e n s p e c i f i e d
rates of g r o w t h (based o n the c o m p o u n d interest
f o r m u l a of P = P [l+r\ ]
n
70
3.5.
E s t i m a t e s of m i d y e a r p o p u l a t i o n s , by region: selected
years, 1650-1990
3.6.
71
L o w e s t and highest age-specific f e r t i l i t y rates per

1,000 w o m e n : 1980s
3.7.
76
A g e - s p e c i f i c f e r t i l i t y rates per 1,000 w o m e n :

H u t t e r i t e s and a l l U . S . w o m e n , a r o u n d 1940
3.8.
77
D i s t r i b u t i o n of m a j o r c o u n t r i e s and territories by
level of age-specific f e r t i l i t y rates: recent years
3.9.
ix
77
B i r t h rates by l i v e b i r t h order and percentage change

i n rates: U n i t e d States, selected years,
1942-88
83
3.10. C r u d e b i r t h rates and d i r e c t l y standardized b i r t h

rates: selected places and dates
86
3.11. C r u d e b i r t h rates and d i r e c t l y standardized b i r t h

rates: U n i t e d States, selected years, 1940-88
87
3.12. G e n e r a l f e r t i l i t y rates f o r selected countries, agestandardized by the direct m e t h o d : late 1980s

(standard = Sweden, 1988)
88
3.13. O b s e r v e d general f e r t i l i t y rates and age-standardized

general f e r t i l i t y rates w i t h Sweden (1988), India
(1990), and the R e p u b l i c of Korea (1990) as standard
p o p u l a t i o n s : Egypt, P h i l i p p i n e s , Sweden, U n i t e d
States, and Japan
89
3.14. C a l c u l a t i o n of total f e r t i l i t y rates f o r the U n i t e d

States: 1957, 1976, and 1988
93
3.15. C a l c u l a t i o n of the gross r e p r o d u c t i o n rates f o r C o s t a

R i c a : 1984
96
3.16. E s t i m a t e d crude b i r t h rates and gross r e p r o d u c t i o n

rates for w o r l d regions: 1985-90
98
3.17. D i s t r i b u t i o n of countries b y w o r l d region and l e v e l of

gross r e p r o d u c t i o n rate: 1985-90
99
3.18. C a l c u l a t i o n of the gross and net r e p r o d u c t i o n rates

and the length of a generation for the n o n w h i t e
p o p u l a t i o n : U n i t e d States, 1988
100
3.19. Measures of reproduction: selected countries, recent

years
102
3.20. G r o s s and net r e p r o d u c t i o n rates: Europe, G r e a t

D e p r e s s i o n years, p o s t - W o r l d War II, and recent
past
104
3.21. G r o s s and net r e p r o d u c t i o n rates, by color: U n i t e d

States, 1905-10 to 1988
105
3.22. Percentage ever m a r r i e d and n u m b e r of c h i l d r e n ever

b o r n for w o m e n of ages 4 0 - 4 4 and 3 0 - 3 4 : U n i t e d
States, selected years 1940-90
113
3.23. N u m b e r of c h i l d r e n ever b o r n per 1,000 w o m e n and

per 1,000 ever-married w o m e n , by age: U n i t e d
States, selected years 1940-90
114
3.24. Average n u m b e r of c h i l d r e n ever b o r n per w o m a n

a m o n g ever-married w o m e n of ages 4 5 - 4 9 , by
p r o v i n c e and u r b a n / r u r a l residence: Indonesia, 1980
and 1990
115
3.25. B i r t h p r o b a b i l i t i e s w i t h i n successive b i r t h i n t e r v a l s
2, 3, and 4 - 8 , by d u r a t i o n of i n t e r v a l and contraceptive use status: P h i l i p p i n e s and R e p u b l i c of K o r e a ,
1973-74
122
3.26. Year i n w h i c h any b i r t h i n t e r v a l had to begin, g i v e n a

w o m a n ' s age at the b e g i n n i n g of the i n t e r v a l and her
age at the t i m e of being i n t e r v i e w e d , f o r a s u r v e y
t a k i n g place i n 1990: a l l intervals
126
R o n a l d F r e e d m a n f i r s t suggested the idea of w r i t i n g a series

of self-teaching guides about e l e m e n t a r y d e m o g r a p h i c measures. F o l l o w i n g up o n h i s idea, f a m e s P a l m o r e drafted t w o
short m a n u a l s i n the f a l l of 1969: Measuring
Self-Teaching Guide to Elementary
Mortality:
Measures and Measur-
ing Fertility and Natural Increase: A Self-Teaching Guide to

Elementary
Measures.
In 1971 revised drafts of the m a n u a l s
were p u b l i s h e d as Papers of the East-West P o p u l a t i o n Institute, N o s . 15 and 16. Several m o r e r e v i s i o n s f o l l o w e d .

R o b e r t G a r d n e r j o i n e d t h i s enterprise f o r the f o u r t h
e d i t i o n , p u b l i s h e d i n 1983. W i t h that e d i t i o n w e c o m b i n e d
the t w o m a n u a l s i n t o a single short textbook. T h e Guide,
as
it has c o m e to be called, has gone t h r o u g h seven p r i n t i n g s

and been translated i n t o C h i n e s e , Indonesian, and A l b a n i a n .
W e hope y o u w i l l f i n d this f i f t h e d i t i o n as u s e f u l . W h i l e
the basic content of t h i s e d i t i o n r e m a i n s the s a m e as i n the
f o u r t h e d i t i o n , w e have r e w r i t t e n the text to m a k e the m a terial m o r e accessible, updated the tables to i n c l u d e m o r e
recent data, used m o r e recent research and advanced m e t h ods as examples i n the text, and changed the exercises to
reflect the present d e m o g r a p h i c s i t u a t i o n .
T h i s v o l u m e m a y be u s e f u l f o r several k i n d s of reader.
In i n t r o d u c t o r y courses o n p o p u l a t i o n issues, u s u a l l y n e i ther the i n s t r u c t o r n o r the students w a n t to spend m u c h class
t i m e d i s c u s s i n g s u c h b a s i c d e m o g r a p h i c measures as the
crude b i r t h rate. T h e G u i d e is designed to f a m i l i a r i z e graduates, undergraduates, and even advanced h i g h s c h o o l students
w i t h m o s t measures pf m o r t a l i t y , f e r t i l i t y , and n a t u r a l i n crease that are l i k e l y to be encountered i n s u c h a course. It
m a y also be used as an i n t r o d u c t o r y text i n courses o n d e m o graphic methods. Since most demographic methods
are
readily m o d i f i e d f o r other k i n d s of s o c i a l s c i e n c e measurem e n t , a t h i r d use is i n s o c i a l science courses that do n o t c o n -
xii
Preface
centrate o n p o p u l a t i o n studies. For e x a m p l e , the i n s t r u c t o r

of an i n t r o d u c t o r y course i n m e t h o d o l o g y f o r s o c i o l o g i s t s
m i g h t draw o n this b o o k for s o m e p o r t i o n of the course.
T h e Guide
is designed for self-teaching. M o s t i n s t r u c -
tors w i l l f i n d that they need a m a x i m u m of three h o u r s of

class t i m e to r e v i e w the exercises w i t h students and c l a r i f y
any p o i n t s that are c o n f u s i n g . O u r c l a s s r o o m experience w i t h
these exercises, m a n y of w h i c h appeared i n earlier e d i t i o n s
of the Guide,
and c o m m e n t s w e have received f r o m other
i n s t r u c t o r s b o t h i n the U n i t e d States and i n A s i a have enabled us to i n c o r p o r a t e changes that s h o u l d e n h a n c e t h i s

edition's self-teaching value.
Each chapter explains e l e m e n t a r y p r i n c i p l e s of d e m o graphic measurement,
and each c o n s i s t s of the f o l l o w i n g
parts: (a) d e f i n i t i o n s of measures and e x a m p l e s of t h e i r u s u a l

values, (b) exercises and questions for the s t u d e n t that e m phasize i n t e r p r e t a t i o n rather t h a n c o m p u t a t i o n , and (c) references to other sources that use the measures i n i n t e r e s t i n g
or i m p o r t a n t w a y s .
For m u c h of the d i s c u s s i o n , w e b o r r o w h e a v i l y f r o m
s t a n d a r d references o n d e m o g r a p h i c m e t h o d s ,
George W. Barclay, Techniques
of Population
Analysis
Y o r k : John W i l e y a n d Sons, Inc., 1958); A . J. Jaffe,

of Statistical
Methods
for Demographers
including:
(New
Handbook
(Washington, D . C . :
U . S . Bureau of the C e n s u s , U . S . G o v e r n m e n t P r i n t i n g O f fice, 1951); and H e n r y S. S h r y o c k , Jacob S. Siegel, and A s s o ciates, The Methods
and Materials
of Demography
(Wash-
i n g t o n , D C : U . S . G o v e r n m e n t P r i n t i n g O f f i c e , 1971), 2 v o l umes. ( A condensed e d i t i o n of t h i s last w o r k is a v a i l a b l e i n

one v o l u m e and is s t i l l i n p r i n t f r o m A c a d e m i c Press, 1978.)
O t h e r v a l u a b l e resources have been a m a n u s c r i p t c o p y of the
" M a n u a l of D e m o g r a p h i c Research T e c h n i q u e s , " by D o n a l d
J. Bogue and E v e l y n K i t a g a w a , and various m a n u a l s of the
U n i t e d N a t i o n s . A d d i t i o n a l c i t a t i o n s are p r o v i d e d at appropriate places i n the text, and a c o m p l e t e l i s t of references is
f o u n d at the end of the
Guide.
For data i n o u r tables w e have r e l i e d m o s t h e a v i l y o n

several basic sources, i n c l u d i n g the U n i t e d N a t i o n s
Demo-
Preface
graphic
Yearbook
series and other U N p u b l i c a t i o n s , K e y f i t z
and Flieger's World Population
Growth and Aging:
graphic Trends in the Late Twentieth

Vital Statistics
of the United
States.
Century
Demo-
(1990), a n d
A good s u m m a r y source
of data o n c o u n t r i e s of the w o r l d is the a n n u a l World

tion Data
xiii
Popula-
Sheet p u b l i s h e d by the P o p u l a t i o n R e f e r e n c e B u -
reau.
In preparing a v o l u m e l i k e this, the authors are a l w a y s
indebted to colleagues and students w h o have p a t i e n t l y read
earlier versions and c o n t r i b u t e d to the f i n a l c l a r i t y of the
product through their c o m m e n t s . We are p a r t i c u l a r l y indebted
to R o n a l d Freedman, w h o not o n l y f i r s t suggested the idea
but also p r o v i d e d several of the exercises used here and m a d e
m a n y v a l u a b l e c o m m e n t s . W e have also b e n e f i t e d f r o m c o m m e n t s m a d e a l o n g the w a y b y R e y n o l d s F a r l e y , N a t h a n
K e y f i t z , Susan P a l m o r e , M o n i c a Fong, D a v i d S w a n s o n , D a v o r
Jedlicka, M e a d C a i n , J. S. M a c D o n a l d , R o b e r t R e t h e r f o r d ,
Peter Xenos, Sandra Ward, Robert H e a r n , M a u r e e n St. M i c h e l ,
A n d r e w Kantner, and G r i f f i t h Feeney. W e were v e r y f o r t u nate to have Sandra Ward as our editor for this e d i t i o n . T h a n k s
are also due to C o n n i e K a w a m o t o , Steven Swapp, L o i s Bender,
C l i f f o r d Takara, and R u s s e l l F u j i t a for their assistance i n preparing the Guide
for publication.
F i n a l l y , w e are g r a t e f u l to the m a n y s t u d e n t s at the

U n i v e r s i t y of M i c h i g a n and the U n i v e r s i t y of H a w a i i a n d to
n u m e r o u s participants i n East-West C e n t e r w o r k s h o p s w h o
discovered errors and tried t h e i r best to save us f r o m m a k i n g
s i m p l e matters seem c o m p l e x . N e v e r t h e l e s s , w e m u s t bear
r e s p o n s i b i l i t y for any errors that r e m a i n . W e w o u l d appreciate i t if y o u w o u l d b r i n g t h e m to o u r a t t e n t i o n .
CHAPTER 1
Rates, Ratios, Percentages,

and Probabilities
We c a n measure the o c c u r r e n c e or i n c i d e n c e of an event
(death, f o r example) i n m a n y w a y s . In t h i s chapter w e discuss v a r i o u s types of rates, ratios, percentages, and p r o b a b i l i ties, u s i n g the m e a s u r e m e n t of m o r t a l i t y as an i l l u s t r a t i o n .
W h e n w e as demographers measure an event, w e w a n t
to be precise about:
(a)
the t i m e p e r i o d to w h i c h w e are referring,
(b)
the group of people to w h i c h w e are referring, and
(c)
the type of occurrence w e are m e a s u r i n g .

D i f f e r e n c e s i n the s p e c i f i c i t y of each of these three fac-
tors are responsible f o r the existence of m a n y d i f f e r e n t dem o g r a p h i c measures.
RATIOS.
Y o u are probably already f a m i l i a r w i t h the everyday use of
PROPORTIONS,
ratios, proportions, and percentages. E x a m p l e s of everyday
A N D PERCENTAGES
ratios are:
(a)
" I ' l l give y o u odds of 3 to 1 that R u s s i a w i n s the G o l d

M e d a l f o r g y m n a s t i c s at the next S u m m e r O l y m p i c s "
and
(b)
"Lee's S u p e r m a r k e t is t w i c e as e x p e n s i v e as Fong's."
G e n e r a l l y , a ratio is a single n u m b e r that expresses the
relative s i z e of t w o other n u m b e r s . T h e result of d i v i d i n g a

n u m b e r X by another n u m b e r Y is the ratio of X to Y that is,
;
X
= ratio ofX toY.
Y
M a n y ratios are used i n d e m o g r a p h i c , m e a s u r e m e n t ,
several of w h i c h are d e f i n e d i n T a b l e 1.1. F o r any ratio, w e
Rates, Ratios, Percentages, and Probabilities
s h o u l d s p e c i f y c a r e f u l l y w h a t type of event or p o p u l a t i o n is
the referent. For example, the sex ratio, or n u m b e r of m a l e s
per 100 females, m i g h t refer to:
(a)
the t o t a l p o p u l a t i o n of the U n i t e d States i n 1994,
(b)
persons 15-34 years of age i n B a l i , Indonesia, i n 1987,

or
(c)
l i v e b i r t h s o c c u r r i n g i n H o n g K o n g i n 1985, 1986, a n d
1987.
W e can also use the sex ratio i n m o r t a l i t y a n a l y s i s . F o r
e x a m p l e , w e m i g h t c o m p a r e the n u m b e r of m a l e deaths w i t h
the n u m b e r of f e m a l e deaths f r o m a c e r t a i n disease.
A proportion
is a special type of ratio i n w h i c h the de-
n o m i n a t o r i n c l u d e s the n u m e r a t o r . W e m i g h t , f o r e x a m p l e ,
c a l c u l a t e the p r o p o r t i o n of a l l deaths that o c c u r r e d to males,
as i n the f o l l o w i n g f o r m u l a :
Proportion of deaths
that occurred to males
D + D'
m
deaths to males
deaths to males plus
deaths to females.
Table 1.1. Ratios frequently used in demographic work

Definition
Formula"
Ratio
P+ P
Dependency
ratio
100 x
no. of persons under 15

or over 64
no. of persons 15-64
years old
100 x
no. of males in group i

no. of females in group i
P
50*15
Sex ratio
P/
Population
density
no. of persons in geographic

area i
no. of sq. km. {or miles) of land
area in geographic area i
a.
Child-woman
ratio
5^0
So
P
1,000 X
f
35* 15
30 15
no. of children under

five years
no. of females 15-49
or 15-44 years old
Note: See Appendix 1 for a discussion of the notation system and formulas used in this volume.
a. The symbol ~> stands for infinity. In the formulas throughout this volume it indicates an open-ended age group.
For example, _P refers to the population 65 and over, or 65t.
tf
A percentage
is a s p e c i a l type of p r o p o r t i o n , o n e i n
w h i c h the ratio is m u l t i p l i e d b y a constant, 100, so that the

ratio is expressed per 100. If y o u leaf t h r o u g h the tables i n
this v o l u m e , y o u w i l l see m a n y e x a m p l e s of ratios, proportions, a n d percentages. A l l of these s i m p l e measures are usef u l to the demographer.
RATES
G e n e r a l l y , ratios and percentages are u s e f u l f o r a n a l y z i n g the

c o m p o s i t i o n of a set of events or of a p o p u l a t i o n . Rates, i n
contrast, are used to study the d y n a m i c s of change. A
rate
refers to the o c c u r r e n c e of events over a g i v e n i n t e r v a l i n

t i m e . W e can d e f i n e a rate of i n c i d e n c e i n general t e r m s as
follows:
Rate of incidence
number of events that occur within.

_^
a given time interval
number of members of the population
who were exposed to the risk of the
event during the same time interval.
S p e c i f y i n g the n u m b e r of persons "exposed to r i s k " i n

the d e n o m i n a t o r is i m p o r t a n t . If y o u were s t u d y i n g m o r t a l i t y over a one-year period i n c o u n t r y A , y o u s h o u l d note that
a person w h o died before the year ended was not exposed to
risk for the w h o l e year, n o r was a c h i l d w h o w a s b o r n halfw a y t h r o u g h the year. People w h o m o v e d to c o u n t r y A o n l y
one m o n t h before the year ended w e r e n o t exposed to the
risk of d y i n g i n c o u n t r y A f o r the w h o l e year either.
T h e concept of "person-years l i v e d " is the i d e a l w a y to
s p e c i f y the p o p u l a t i o n exposed to the risk of an event a n d
t h u s the i d e a l d e n o m i n a t o r for a d e m o g r a p h i c rate. It is s i m p l y the product of the n u m b e r of persons m u l t i p l i e d b y the
n u m b e r of years, or f r a c t i o n s of years, that each person l i v e d
i n a g i v e n place. T a b l e 1.2 presents the c a l c u l a t i o n pf personyears l i v e d f o r a h y p o t h e t i c a l s m a l l t o w n . N o t e that the p o p u l a t i o n at the b e g i n n i n g and the e n d of the year, 700, d i f f e r s
f r o m the n u m b e r of person-years l i v e d . T h e e x a m p l e is u n u s u a l because, a m o n g other i t e m s :
(a)
n o net g r o w t h o c c u r r e d i n the t o w n ,
(b)
200 people d i e d o n one day (January 15), and
(c)
100 people arrived f r o m elsewhere o n one day (October 25).

S u c h occurrences w o u l d be h i g h l y u n u s u a l , b u t t h e y
i l l u s t r a t e h o w the n u m b e r of person-years l i v e d c a n be q u i t e
different f r o m the p o p u l a t i o n at e i t h e r the b e g i n n i n g or the
end of a period under study.
T h e c a l c u l a t i o n of a c t u a l person-years l i v e d f o r a real
p o p u l a t i o n of any large size w o u l d be d i f f i c u l t , if n o t i m p o s sible. For t h i s reason, m o s t demographic rates use an approxi-
Table 1.2. Calculation of the number of person-years lived during one year in a hypothetical
small town having a population of 700 persons on January 1 and very erratic demographic
behavior
Number
of
persons
Events and dates
700
Alive on January 1
493
Lived in the town continuously

from January 1 to December 31
Born January 11
Born January 11, died November 9
Died January 15
Born February 21, died April 27
Born March 6, died March 31
Died April 8
Born April 10
Arrived from outside town April 18
Died June 1
Died June 5
Born June 7
Died June 22
Born June 24
Died June 30
Left town August 16
Born August 24
Born September 13, died November 13
Born October 1
Born October 7
Born October 19
Arrived from outside town October 25
1
1
200
1
1
2
94
4
1
1
1
1
1
1
1
1
1
1
2
2
100
Total person-years lived

700
Alive on December 31
Source: Modified from Barclay 11958, 39).
Number
of days
lived
Number of
person-years
lived
179,945
354
302
3,000
65
25
196
24,910
1,032
152
156
207
173
190
181
228
129
61
91
170
146
6,700
493.00
.97
.83
8.22
.18
.07
.54
68.25
2.83
.42
.43
.57
.47
.52
.50
.62
.35
.17
.25
.46
.40
18.36
598.41
m a t i o n of person-years l i v e d i n the d e n o m i n a t o r . W e a s s u m e
that b i r t h s , deaths, a n d m o v e m e n t s i n a n d out of the p o p u l a t i o n o c c u r at u n i f o r m i n t e r v a l s , or " s m o o t h l y , " d u r i n g the
p e r i o d u n d e r study. If t h i s a s s u m p t i o n is true, t h e n the n u m ber of people a l i v e at the m i d d l e of the year (July 1) w i l l e q u a l
the n u m b e r of person-years l i v e d . T h i s p o p u l a t i o n a l i v e at
the m i d d l e of the year is c a l l e d the m i d y e a r or c e n t r a l p o p u l a t i o n , and so a death (or birth) rate w i t h the m i d y e a r p o p u l a t i o n as a d e n o m i n a t o r is k n o w n as a central rate.
If (as w e have assumed) b i r t h s , deaths, a n d m o v e m e n t s
i n and out of the p o p u l a t i o n are e v e n l y d i s t r i b u t e d t h r o u g h out the year:
(a)
for every b i r t h at m i d n i g h t o n January 1, there is one at

m i d n i g h t o n D e c e m b e r 30. T h e average n u m b e r of person-years l i v e d for the t w o b i r t h s is:
(b)
f o r every death at m i d n i g h t o n January 15, there is one

at m i d n i g h t o n D e c e m b e r 16. T h e average n u m b e r of
person-years l i v e d f o r the t w o deaths is:
T h i s is w h y the m i d y e a r p o p u l a t i o n ( w h i c h is o f t e n c a l c u lated by t a k i n g the average of the p o p u l a t i o n at the beginn i n g and the end of the year) is u s u a l l y a good a p p r o x i m a t i o n
of person-years l i v e d . N o t e , however, the s i g n i f i c a n c e of the
a s s u m p t i o n of e v e n l y d i s t r i b u t e d b i r t h s , deaths, and m o v e m e n t s i n and out of the p o p u l a t i o n . In a p o p u l a t i o n subject
to c o n d i t i o n s l i k e the s m a l l t o w n of Table 1.2, the m i d y e a r
p o p u l a t i o n , w h e t h e r a c t u a l l y measured or c a l c u l a t e d as an
average, is not a good, a p p r o x i m a t i o n of the n u m b e r of person-years l i v e d .
W e can f u r t h e r illustrate the errors that m i g h t arise f r o m
u s i n g the m i d y e a r p o p u l a t i o n a p p r o x i m a t i o n f o r person-years
l i v e d w i t h t w o m o r e r e a l i s t i c examples. In the first, a " c o l lege t o w n " w h o s e m a i n i n d u s t r y is higher e d u c a t i o n m i g h t
have h i g h percentages of its p o p u l a t i o n out of t o w n every
January 1st, celebrating N e w Year's D a y at h o m e . E s t i m a t -
ing person-years l i v e d i n s u c h a t o w n by t a k i n g the average

of the p o p u l a t i o n o n successive January firsts w o u l d lead to a
s u b s t a n t i a l u n d e r e n u m e r a t i o n of the true n u m b e r of personyears l i v e d . A second e x a m p l e i n v o l v e s an a g r i c u l t u r a l v i l lage w i t h considerable t e m p o r a r y i n - m i g r a t i o n d u r i n g the
harvest season at the m i d d l e of the year. In t h i s e x a m p l e the
a c t u a l size of the p o p u l a t i o n at m i d y e a r w o u l d be a n overestimate of the n u m b e r of person-years l i v e d d u r i n g the w h o l e
year. A s b o t h examples i l l u s t r a t e , i n c e r t a i n s i t u a t i o n s (espec i a l l y s m a l l populations) and o n certain occasions, use of the
m i d y e a r p o p u l a t i o n , a c t u a l or estimated, can lead to i n a c c u rate a p p r o x i m a t i o n s of person-years l i v e d . In the vast majority of situations, however, the m i d y e a r p o p u l a t i o n is an acceptable a p p r o x i m a t i o n .
To c a l c u l a t e a m i d y e a r p o p u l a t i o n , y o u w o u l d u s u a l l y
take the p o p u l a t i o n o n January 1 of y e a r X , add it to the p o p u l a t i o n o n January 1 of year X + 1, and then d i v i d e by 2. F o r
o u r s m a l l t o w n i n Table 1.2, the m i d y e a r p o p u l a t i o n is [(700
+ 700J/2] = 700. A l t h o u g h the c o n v e n t i o n s used b y v a r i o u s
countries f o r reporting and e s t i m a t i n g t h e i r p o p u l a t i o n s differ, it s h o u l d u s u a l l y be possible to c a l c u l a t e the m i d y e a r
p o p u l a t i o n b y u s i n g the average or s o m e other s i m p l e approach.
EXERCISE i
C o n s t r u c t a s m a l l h y p o t h e t i c a l p o p u l a t i o n , s p e c i f y i n g the
same characteristics and events as are s p e c i f i e d i n T a b l e 1.2.
C a l c u l a t e the m i d y e a r p o p u l a t i o n . C a l c u l a t e the n u m b e r of
person-years l i v e d . A r e they close to the same value? If not,
why
not?
A NOTE OF
Because demographers c o m e f r o m v a r i o u s a c a d e m i c d i s c i -
CAUTION
p l i n e s and for h i s t o r i c a l reasons, the "rates" used b y demographers are not always rates as w e have described t h e m above.
By c o n v e n t i o n , s o m e o r d i n a r y percentage figures are c a l l e d
rates. O n e e x a m p l e of s u c h usage is the " l i t e r a c y rate," w h i c h
is s i m p l y the p e r c e n t a g e of the a d u l t p o p u l a t i o n t h a t is
l i t e r a t e . Y o u m u s t l e a r n h o w to d e t e r m i n e w h e t h e r a rate is
r e a l l y a rate, a s i m p l e percentage, or s o m e t h i n g else. In each
case, the d e f i n i t i o n of the measure s h o u l d be clear e n o u g h to
a l l o w readers to decide w h e t h e r i t is a rate or another type of
measure. M o s t of the rates discussed i n t h i s guide are, f o r t u nately, real rates,- the exceptions are the r e p r o d u c t i o n rates
discussed i n C h a p t e r 3, w h i c h are m o r e l i k e p r o b a b i l i t i e s t h a n
rates.
PROBABILITIES
A s w e have noted, rates refer to the o c c u r r e n c e of events

over a g i v e n i n t e r v a l of t i m e . T h e d e n o m i n a t o r of a rate is,
ideally, the n u m b e r of person-years of exposure a n d m o r e
c o m m o n l y the average or m i d p e r i o d p o p u l a t i o n exposed to
the event i n q u e s t i o n . A probability
is s i m i l a r to a rate, w i t h
one i m p o r t a n t difference: the d e n o m i n a t o r is c o m p o s e d of

a l l those persons i n the g i v e n p o p u l a t i o n at the beginning
the period
of observation.
of
T h u s , if 10 people die i n one year
out of a p o p u l a t i o n that n u m b e r e d 1,000 at the start of the

year, w e say that the p r o b a b i l i t y of d y i n g f o r this group duri n g that year was 10/1000, or 0.01000. N o t e that t h i s is different f r o m the death rate for the same period, w h i c h w o u l d
be (if the deaths were e v e n l y distributed) 10/(1/2(1000+990)]
= 10/995 = 0.01005. For p o p u l a t i o n s e x p e r i e n c i n g o n l y deaths
(and not m i g r a t i o n or births), p r o b a b i l i t i e s of d y i n g w i l l a l w a y s be s m a l l e r t h a n the comparable death rates, because
the numerators w i l l be the same but the d e n o m i n a t o r s w i l l
be larger. We deal w i t h the concept of probabilities m o r e w h e n
w e reach the d i s c u s s i o n of the l i f e table and q
n
i n C h a p t e r 2.
W i t h this brief i n t r o d u c t i o n to concepts used i n the

m e a s u r e m e n t of any d e m o g r a p h i c event, w e n o w t u r n to the
measures used i n s t u d y i n g m o r t a l i t y .
1. The survival ratios used in connection with the life table are sometimes called
survival rates, but not in this Guide.
CHAPTER 2
Mortality
For h i s t o r i c a l reasons w e begin w i t h measures of m o r t a l i t y .
T h r o u g h o u t m o s t of h u m a n h i s t o r y the fate of a p o p u l a t i o n
w h e t h e r i t grew, stagnated, or f a i l e d to s u r v i v e d e p e n d e d
m o r e o n m o r t a l i t y t h a n o n f e r t i l i t y or m i g r a t i o n . U n t i l f o u r
decades ago m o r t a l i t y and its c o n t r o l were the c e n t r a l issue
i n p o p u l a t i o n p o l i c y and of chief d e m o g r a p h i c interest f o r
m o s t of the w o r l d ' s c o u n t r i e s . F e r t i l i t y and m i g r a t i o n gained
the demographic spotlight o n l y recently. C o n s e q u e n t l y , m u c h
of the earliest w o r k o n the d e v e l o p m e n t of d e m o g r a p h i c
measures c o n c e n t r a t e d o n measures of m o r t a l i t y . F o r exa m p l e , w o r k o n the l i f e table (discussed later i n t h i s chapter)
began as early as the m i d - s e v e n t e e n t h century. H e r e w e start
w i t h s i m p l e r measures, the f i r s t b e i n g the crude death rate.
CRUDE
RATE
DEATH
T h e crude death rate (CDR) is defined as the n u m b e r of people

who
die i n a g i v e n year d i v i d e d b y the n u m b e r of people i n
the p o p u l a t i o n i n the m i d d l e of that year. C o n v e n t i o n a l l y ,

we express the rate per 1,000 persons. A s a f o r m u l a , w e have:
1
CDR = 1,000
number of deaths
midyear population
where D = deaths in the year,

P-
midyear population, and
k= 1,000.
1. You should be sensitive to the (act that different constants (100, 1,000, and
100,000 are common constants) are used for different demographic measures. For
example, crude birth rates and crude death rates are usually expressed per 1,000,
but growth rates are expressed as percentages. When calculating a rate, it is always
safest to proceed without the use of a constant until you get the final answer, then
to use the constant to express the rate per thousand, per hundred (percent), or
whatever is the usual constant for that type of rate.
10
Mortality
For example, suppose a very s m a l l , h y p o t h e t i c a l c o u n try
h a d a p o p u l a t i o n of 550 people o n D e c e m b e r 3 1 , 1980,
and a p o p u l a t i o n of 650 o n D e c e m b e r 3 1 , 1981. T h e m i d 1980 p o p u l a t i o n w o u l d be: [(550 + 650)/2] = 600. If 15 deaths

o c c u r r e d i n this s m a l l c o u n t r y i n 1980, the crude death rate
w o u l d be: (15/600 x 1,000) = 25 per 1,000 or, s i m p l y , 25.
In the 1980s, crude death rates for countries w i t h a popul a t i o n of 1 m i l l i o n o r m o r e ranged f r o m just o v e r 2 to m o r e
2
t h a n 23 per 1,000 per a n n u m (see T a b l e 2.1). In other w o r d s ,

for each 1,000 persons exposed to the risk of d y i n g i n the
1980s, between 2 and 24 d i e d each year. T h e c r u d e death rate
t h u s indicates that the r i s k of death was m o r e t h e n 10 t i m e s
higher i n the h i g h - m o r t a l i t y countries than i n the l o w - m o r t a l i t y countries.
EXERCISE i
T h e f o l l o w i n g statements about m o r t a l i t y are a l l inadequate

in some way. In w h a t w a y is each s t a t e m e n t inadequate?
1. T e n people died i n 1991.
2.
T e n people died i n 1991 o u t of a p o p u l a t i o n that n u m -
3.
T e n people d i e d o u t of 1,000 a l i v e o n 3 0 June 1991 i n
bered 1,000 o n 31 D e c e m b e r 1991.

West C o u n t r i d a d .
The
correct answers to the exercises a n d r e v i e w q u e s t i o n s
are f o u n d i n A p p e n d i x 3.
AGE-SPECIFIC
A S i t s n a m e i m p l i e s , the crude death rate is a c r u d e measure.
DEATH RATES
yVe a l l k n o w that a 95-year-old m a n i s m o r e l i k e l y to d i e

than a 20-year-old w o m a n . Soldiers f i g h t i n g o n t h e f r o n t l i n e s
i n a w a r are m o r e l i k e l y to d i e t h a n a student i n a l m o s t a n y
class. In other words, d i f f e r e n t subgroups i n a p o p u l a t i o n are
exposed to different r i s k s of dyingbecause of t h e i r o c c u p a t i o n or t h e i r age or s o m e other characteristic.
Because of these differences i n exposure to the risk of
d y i n g , demographers o f t e n use specific
2. The omitted countries are listed in Appendix 4.
death
rates.
A spe-
Table 2.1. Highest and lowest crude death rates, by region: recent
years
Crude death rates
Region and country
Year or
period
<P
High
Africa
Sierra Leone
Guinea
Mauritius
Tunisia
1985-90
1985-90
1985-90
1985-90
23.4*
22.0'
America, North
Haiti
United States
Costa Rica
Jamaica
Panama
1985-90
1990
1989
1988
1985-90
13.2*
8.6
America, South
Uruguay
Argentina
Venezuela
Chile
1989
1988
1989
1989
Asia (excluding former USSR)

Yemen
Lads
Kuwait
United Arab Emirates
1990
1985-90
1987
1985-90
21.2
16.9'
1990
13.4
1989
1989
1988
12.4
Europe (excluding former USSR)

Hungary
Former German Democratic
Republic
Albania
Spain
Former USSR
Ukraine
Azerbaijan
Oceania'
Papua New Guinea
Australia
New Zealand
er
6.4*
7.3*
3.8
5.2
5.2*
9.6
8.4
4.4
5.8
2.2
3.8'
5.7
8.2
1989
b
10.0
12
6
1985-90
1989
1989
^
Low
11.6*
7.4
8.2
Source: United Nations Statistical Office, Demographic Yearbook 1990 (1992, table 4).
Notes: Countries for which data are known to be incomplete or of unknown
reliability have been omitted. Data for periods before circa 1985 have been omitted.
Countries with populations of fewer than 1 million are excluded (see list of such
countries in Appendix 4).
' Estimates prepared by the Population Division of the United Nations.
a. Only three countries in this region have populations of more than 1 million.
b. Data from Population Reference Bureau, 1992 World Population Data Sheet (1992).
12
Mortality
c i f i c death rate refers o n l y to a subgroup i n a p o p u l a t i o n . T h e

m o s t c o m m o n l y used s p e c i f i c death rates are
death
rates.
age-specific
W e d e f i n e a n age-specific death rate ( A S D R ) i n
the f o l l o w i n g f o r m u l a :
M=
A S D R for age group x to x + n

number of deaths to persons agex, x + n
midyear population o f persons age*, x + n
where D = deaths to persons o f age groupx to x + n,

n
P = midyear population o f age group* to x + n, and
* = 1,000.
Figure 2.1 s h o w s t w o t y p i c a l patterns of age-specific
death rates, one f o r an e c o n o m i c a l l y developed c o u n t r y , the
other f o r a less developed c o u n t r y . In b o t h cases the death
rates are highest for t h e very y o u n g and t h e v e r y o l d . T h i s is
the m o s t c o m m o n pattern f o r age-specific death rates.
L o o k i n g at the m i n i m u m and m a x i m u m figures f o r ages p e c i f i c death rates b y sex s h o w n i n T a b l e 2.2, y o u s h o u l d
Figure 2.1. Age-specific death rates: Guatemala, 1985, and Japan, 1989
1,000 I
0.1
0-4
5-9
10-14 15-19 20-24 24-29 30-34 35-39 40-44 45-49 50-54 55-59 60-4 65-9 70-74 75-79 80*
Age group
Note: The vertical scale of the graph is logarithmic.
Table 2.2. (continued)

Highest ASDR
Ages
30-34
Sex
Rate"
Country and year
Rate*
Country and year
Female
'3.4
2.6
Bangladesh, 1986
Guatemala, 1985
0.3
Male
*5.3
5.3
'3.5
3.1
El Salvador, 1986
Puerto Rico, 1988
Bangladesh, 1986
Guatemala, 1985
0.8
Ireland, 1988
Hong Kong, 1989
Netherlands, 1989
Japan, 1989
*6.4
6.4
*4.9
4.6
El Salvador, 1986
Guatemala, 1985
Bangladesh, 1986
Guatemala, 1985
0.9
'7.1
7.0
'6.6
4.7
11.1
Algeria, 1982
USSR, 1989
Bangladesh, 1986
Guatemala, 1985
Hungary, 1989
1.8
Italy, 1987
Japan, 1989
Norway, 1989
Kuwait, 1986
1.0
Hong Kong, 1989
3.1
1.6
Kuwait, 1986
Japan, 1989
Japan, 1989
5.2
2.6
Japan, 1989
Japan, 1989
8.8
3.8
Japan, 1989
Japan, 1989
12.8
5.7
Japan, 1989
Japan, 1989
19.5
9.5
33.8
17.2
57.3
Japan, 1989
Japan, 1989
Japan, 1989
Japan, 1989
Japan, 1989
Female
35-39
Male
Female
40-44
Male
Female
45^*9
Male
Female
50-54
Male
Female
55-59
Male
Female
60-64
Male
Female
65-69
Male
Female
Male
Female
Male
70-74
75-79
80-84
85+
Lowest ASDR
Female
Male
Female
Male
Female
'7.3
5.8
16.6
10.1
8.8
26.3
'21.4
17.8
34.1
22.4
19.9
60.7
51.8
73.6
59.5
110.9
100.7
75.5
156.3
127.4
276.9
256.3
Bangladesh, 1986
Guatemala, 1985
Hungary, 1989
Bangladesh, 1986
Egypt, 1986
Egypt, 1986
Bangladesh, 1986
Egypt, 1986
Hungary, 1989
Bangladesh, 1986
Egypt, 1986
Egypt, 1986
Egypt, 1986
Egypt, 1986
Egypt, 1986
Korea, Rep. of, 1989
Czechoslovakia, 1989
Guatemala, 1985
Germany, Dem. Rep., 1988
Romania, 1989
Singapore, 1988
Mexico, 1985
0.5
0.7
32.0
88.7
57.2
117.2
95.7
Spain, 1986
Italy, 1987
Switzerland, 1987
Ireland, 1988
Japan, 1989
Kuwait, 1986
Japan, 1989
Hong Kong, 1989
Hong Kong, 1989
Hong Kong, 1989
Hong Kong, 1989
Notes: Many of the rates are estimates that vary in reliability. An asterisk (*) indicates that the data on which the
highest rate is based are incomplete or of unknown reliability, and the highest reliable estimate is also given. Data
are unavailable for the oldest ages in many countries. Countries with populations of under 1 million are excluded; a
list of such countries appears in Appendix 4.
a. Rates per 1,000 population.
Table 2.2. Highest and lowest age-specific death rates, by sex: recent years
Highest ASDR
Ages
Sex
<1
Male
Female
Rate'
* 105.1
98.8
*95.4
91.3
'12.5
7.6
Lowest ASDR
Country and year
Rate*
Algeria, 1982
Egypt, 1986
Algeria, 1982
Egypt, 1986
Algeria, 1982
Egypt, 1986
4.8
Japan, 1989
4.2
Japan, 1989
0.3
Ireland, 1988
Sweden, 1988
Hong Kong, 1989
Norway, 1989
Denmark, 1987
Finland, 1987
France, 1987
Switzerland, 1987
Puerto Rico, 1988
Sweden, 1988
Australia, 1989
Hong Kong, 1989
Japan, 1989
Netherlands, 1989
United Kingdom, 1989
Austria, 1989
Hong Kong, 1989
Japan, 1989
Netherlands, 1989
Denmark, 1987
Germany, Fed. Rep., 1988
Italy, 1987
Switzerland, 1987
France, 1988
Ireland, 1988
Israel, 1988
Sweden, 1988
Austria, 1989
Hong Kong, 1989
Japan, 1989
Netherlands, 1989
Israel, 1988
Japan, 1989
Hong Kong, 1989
1-4
Male
5-9
Male
*2.9
Bangladesh, 1989
0.2
Female
2.0
'2.8
Guatemala, 1985
Bangladesh, 1986
0.1
Male
'1.7
1.3
Bangladesh, 1989
Guatemala, 1985
0.2
Female
*1.1
1.1
'3.6
1.9
'2.3
1.4
*5.1
3.3
'3.1
Bangladesh,
Guatemala,
Iran, 1986
Guatemala,
Bangladesh,
Guatemala,
El Salvador,
Guatemala,
Bangladesh,
0.1
'5.3
4.0
El Salvador, 1986
Guatemala, 1985
10-14
15-19
Male
Female
20-24
Male
Female
25-29
Male
Country and year
1986
1985
0.4
1985
1986
1985
1986
1985
1986
0.2
0.6
0.3
0.7
Italy, 1987
Japan, 1989
Hong Kong, 1989
Italy, 1987
Israel, 1988
Japan, 1989
Hong Kong, 1989
Japan, 1989
Netherlands, 1989

Highest ASDR
Ages
30-34
Sex
Country and year
Rate*
Country and year
Female
3.4
2.6
Bangladesh, 1986
Guatemala, 1985
0.3
Male
*5.3
5.3
"3.5
3.1
El Salvador, 1986
Puerto Rico, 1988
Bangladesh, 1986
Guatemala, 1985
0.8
Ireland, 1988
Hong Kong, 1989
Netherlands, 1989
Japan, 1989
*6.4
6.4
*4.9
4.6
El Salvador,
Guatemala,
Bangladesh,
Guatemala,
1986
1985
1986
1985
0.9
7.1
7.0
*6.6
4.7
11.1
Algeria, 1982
USSR, 1989
Bangladesh, 1986
Guatemala, 1985
Hungary, 1989
1.8
Italy, 1987
Japan, 1989
Norway, 1989
Kuwait, 1986
1.0
Hong Kong, 1989
3.1
Kuwait, 1986
Japan, 1989
Japan, 1989
Female
35-39
Male
Female
40-44
Male
Female
45-49
Male
Female
50-54
Male
Female
55-59
Male
Female
60-64
Male
Female
65-69
Male
Female
Male
Female
Male
70-74
75-79
80-84
85+
Lowest ASDR
Rate"
Female
Male
Female
Male
Female
'7.3
5.8
16.6
'10.1
8.8
26.3
'21.4
17.8
34.1
'22.4
19.9
60.7
51.8
73.6
59.5
'110.9
100.7
75.5
156.3
127.4
276.9
256.3
0.5
0.7
Bangladesh, 1986
1.6
Guatemala, 1985
Hungary, 1989
5.2
Bangladesh, 1986
2.6
Egypt, 1986
Egypt, 1986
8.8
Bangladesh, 1986
3.8
Egypt, 1986
Hungary, 1989
12.8
Bangladesh, 1986
5.7
Egypt, 1986
Egypt, 1986
19.5
Egypt, 1986
9.5
Egypt, 1986
33.8
Egypt, 1986
17.2
Korea, Rep. of, 1989
57.3
Guatemala, 1985
32.0
Germany, Dem. Rep., 1988 88.7
Romania, 1989
57.2
117.2
Singapore, 1988
Mexico, 1985
95.7
Spain, 1986
Italy, 1987
Switzerland, 1987
Ireland, 1988
Japan, 1989
Kuwait, 1986
Japan, 1989
Japan, 1989
Japan, 1989
Japan, 1989
Japan, 1989
Japan, 1989
Japan,
Japan,
Japan,
Japan,
Japan,
1989
1989
1989
1989
1989
Japan, 1989
Hong Kong,
Hong Kong,
Hong Kong,
Hong Kong,
1989
1989
1989
1989
Notes: Many of the rates are estimates that vary in reliability. An asterisk (*) indicates that the data on which the
highest rate is based are incomplete or of unknown reliability, and the highest reliable estimate is also given. Data
are unavailable for the oldest ages in many countries. Countries with populations of under 1 million are excluded; a
list of such countries appears in Appendix 4.
a. Rates per 1,000 population.
Mortality
15
note that data needed to calculate age-specific rates are n o t

available f o r m a n y h i g h - m o r t a l i t y c o u n t r i e s . W e therefore
present t h e range as i t appears f r o m e s t i m a t e d rates reported
i n the U n i t e d N a t i o n s 2990 Demographic
Yearbook.
Higher
A S D R s m a y e x i s t . T h e l o w e s t rates s h o w n are generally f o r

3
European c o u n t r i e s . It is l i k e l y that these l o w e r figures are

really t h e lowest, s i n c e countries w i t h l o w death rates u s u a l l y also have better systems f o r c o l l e c t i n g d e m o g r a p h i c data
o n m o r t a l i t y than do countries w i t h h i g h death rates.
THE EFFECT O F
T h e crude death rate i s a w e i g h t e d s u m of age-specific death
AGE
rates. T a k e the f o l l o w i n g s i m p l e c a l c u l a t i o n s :
ON
COMPOSITION
THE CRUDE
DEATH
RATE
Ages
Number of
persons
in midyear
population
Number of
deaths
in year z
Death rate
in year z
(per 1,000)
2,000
1,000
3,000
40
80
120
20
80
40
0-34
35+
Total, all ages
T h e crude death rate is 4 0 for this h y p o t h e t i c a l p o p u l a t i o n .

It is a w e i g h t e d s u m of t w o age-specific rates, 20 a n d 80. T h e
weights are the proportions of t h e t o t a l p o p u l a t i o n i n each
age group i n t h e m i d y e a r p o p u l a t i o n . T h a t is,
2,000
CDR =
3,000
1,000
x20
40
80
120
3,000
x80
x2oJ + ^ x80
= 40
W e c a n express t h i s basic r e l a t i o n s h i p i n t h e f o l l o w i n g formula:

CDR:
where P = midyear population in age group x to x + n,
n
P = total midyear population,

M = age - specific death rate per 1,000 for age group x to x + n, and
X
L = the sum of the quantity in brackets for all age groups.

3. In many cases the U N estimates are based on civil registers known to be
incomplete or of unknown reliability. The actual ASDRs in such cases are likely
to be even higher.
16
Mortality
T h e fact that the crude death rate is a f u n c t i o n of b o t h

the age-specific death rates a n d the age d i s t r i b u t i o n is d e m onstrated b y the c a l c u l a t i o n s for three h y p o t h e t i c a l p o p u l a tions presented i n Table 2.3. C o u n t r i e s A and B have the same
age-specific death rates, b u t c o u n t r y A's crude death rate is
54 percent higher t h a n c o u n t r y B's. W h y ? Because c o u n t r y A
has a considerably larger p r o p o r t i o n of i t s p o p u l a t i o n i n the
youngest age group, w h i c h is subject to h i g h e r death rates.
C o u n t r i e s B a n d C have the same crude death rates,
but their age-specific death rates are q u i t e different. C o u n t r y
C has a m u c h larger p r o p o r t i o n of i t s p o p u l a t i o n i n the o l d est age group (where w e m i g h t expect to f i n d a h i g h e r death
rate), b u t i n this age group i t s age-specific death rate is o n l y
half of that f o r countries A and B . T h u s , whereas c o u n t r y C
has an older p o p u l a t i o n t h a n either c o u n t r y A o r B, i t s c r u d e
death rate is n o t higher. T h i s example, w h i c h i s designed to
demonstrate the r e l a t i o n s h i p between t h e age d i s t r i b u t i o n
and age-specific rates, does n o t necessarily represent realist i c figures for actual countries.
T w o p o p u l a t i o n s m a y have the same crude death rate
Table 2.3. Age-specific and crude death rates for three hypothetical
populations
Country
Measure
Number of persons in midyear population
for age group:
0-4
5-39
40+
Number of deaths in age group:
0-4
5-39
40+
Age-specific death rate (per 1,000)
for age group:
0-4
5-39
40+
Crude death rate (per 1,000)
1,500
4,000
500
500
5,000
500
500
4,000
1,500
120
40
40
40
50
40
50
20
60
80
10
80
80
10
80
100
5
40
33.3
21.7
21.7
Mortality
17
even t h o u g h one has higher death rates t h a n t h e o t h e r i n

every age group. T h i s result w o u l d occur, f o r e x a m p l e , i f t h e
p o p u l a t i o n w i t h t h e higher age-specific rates were c o n c e n trated i n the age groups between 5 a n d 45, so that m o r e of i t s
people were subject to l o w death rates. It is even p o s s i b l e f o r
one p o p u l a t i o n to have a crude death rate that is l o w e r t h a n
another's a l t h o u g h i t s death rates are h i g h e r at every age.
T h i s paradox is i l l u s t r a t e d i n T a b l e 2.4, w h i c h compares t h e
death rates f o r M a i n e and S o u t h C a r o l i n a i n 1930, a n d i t w i l l
be d e m o n s t r a t e d again w h e n w e d i s c u s s s t a n d a r d i z a t i o n .
S o u t h C a r o l i n a h a d higher A S D R s than M a i n e for every age
group except one. N e v e r t h e l e s s , M a i n e h a d a h i g h e r crude
death rate because a larger p r o p o r t i o n of i t s p o p u l a t i o n w a s
i n the age groups 55 a n d over, w h i c h experience h i g h e r mort a l i t y than m o s t younger groups.
W e have i l l u s t r a t e d the idea that a crude death rate c a n
be s u b d i v i d e d , o r decomposed, i n t o t w o elements: (1) ages p e c i f i c death rates a n d (2) the age d i s t r i b u t i o n , w h i c h deterTable 2.4. Age-specific death rates and populations for Maine and South Carolina: 1930
Ages
0-4
5-9
10-14
15-19
20-24
25-34
35-44
45-54
55-64
65-74
75+
Maine
South Carolina
Percentage
distribution
ASDR
of popu(per 1,000
lation
population) Population
Percentage
ASDR
distribution
(per 1,000
of popupopulation) Population
lation
20.56
0.86
1.40
2.23
3.70
3.91
5.45
10.85
20.36
.52.19
136.45
All ages
Crude death rate
(per 1,000)
13.9
75,037
79,727
74,061
68,683
60,575
105,723
101,192
90,346
72,478
46,614
22,396
9.4
10.0
9.3
8.6
7.6
13.3
12.7
11.3
9.1
5.8
2.8
796,832
99.9
23.92
1.85
1.84
4.26
6.45
8.71
12.42
19.94
33.13
61.47
61.47
205,076
240,750
222,808
211,345
166,354
219,327
191,349
143,509
80,491
40,441
16,723
11.8
13.9
12.8
12.2
9.6
12.6
11.0
8.3
4.6
2.3
1.0
1,738,173
100.1
12.9
Notes: Deaths and populations of unknown ages are excluded. Percentages do not sum exactly to 100.0 because of
rounding.
18
Mortality
m i n e s the p r o p o r t i o n s of t h e p o p u l a t i o n to w h i c h t h e ages p e c i f i c rates apply. W e c a n decompose a crude death rate

i n t o rates that apply to a n y set of characteristics that m i g h t
be h e l p f u l i n a n analysis, a n d w e c a n d i v i d e , o r d i s t r i b u t e ,
the p o p u l a t i o n a c c o r d i n g to those c h a r a c t e r i s t i c s . F o r exa m p l e , i t i s possible to c a l c u l a t e sex-specific death rates a n d
the sex d i s t r i b u t i o n . It i s also possible to c a l c u l a t e age-sexs p e c i f i c death rates (e.g., the death rate f o r m e n or w o m e n of
ages 20-24) a n d the d i s t r i b u t i o n of t h e p o p u l a t i o n b y age a n d
sex. S i n c e m o r t a l i t y rates do v a r y s i g n i f i c a n t l y b y age and b y
sex, a n d because data o n the p o p u l a t i o n d i s t r i b u t i o n a n d o n
deaths are u s u a l l y available b y age a n d sex, i t is c o m m o n f o r
s u c h age-sex-specific death rates to be c a l c u l a t e d . In fact, t h e
approach and logic are quite general. A l s o c a l c u l a t e d are death
rates s p e c i f i c to age, sex, a n d o c c u p a t i o n s i m u l t a n e o u s l y , a l t h o u g h data for these are less c o m m o n l y available. O b v i o u s l y ,
the characteristics f o r w h i c h i t is u s e f u l to decompose a death
rate (or a n y other rate) are u s u a l l y those that m i g h t m a k e a
d i f f e r e n c e i n t h e death rate. T h e r e w o u l d be l i t t l e p o i n t i n
c a l c u l a t i n g death rates s p e c i f i c to eye color, f o r e x a m p l e ,
unless eye c o l o r h a d some bearing o n m o r t a l i t y .
FIRST SET O F
1. In t w o countries, A a n d B, t h e age-specific death rates per
MULTIPLE-CHOICE
i 0 0 0 are as f o l l o w s :
;
Ages
Country A
Country B
0-4
5-24
25-44
45-64
65+
70
5
10
30
80
70
5
10
30
80
W h i c h of the f o l l o w i n g is true?
(a)
T h e crude death rate is h i g h e r i n c o u n t r y A t h a n i n
(b)
T h e crude death rate is higher i n c o u n t r y B t h a n i n c o u n -
(c)
T h e crude death rates are equal i n the t w o c o u n t r i e s .
(d)
T h e crude death rate i n c o u n t r y A m a y be higher, l o w e r ,
country B.
try A .
or the same as i n c o u n t r y B .
Mortality
19
2. T h e crude death rates per 1,000 i n c o u n t r i e s A a n d B are as

f o l l o w s f o r s p e c i f i c areas of the t w o countries:
Areas
Metropolitan areas
Small towns
Rural areas
The
Country A
Country B
15
17
30
14
15
29
crude death rate f o r the w h o l e c o u n t r y is:
(a)
d e f i n i t e l y less i n A t h a n i n B .
(b)
d e f i n i t e l y less i n B than i n A .
(c)
probably h i g h e r i n A t h a n i n B, b u t the reverse is possible.
(d)
p r o b a b l y higher i n B t h a n i n A , b u t the reverse is possible.
STANDARDIZATION
W e have seen that the age c o m p o s i t i o n of the p o p u l a t i o n has

a p r o n o u n c e d effect o n the crude death rate. O t h e r aspects of
p o p u l a t i o n c o m p o s i t i o n m a y also affect the death rate. E x a m p l e s of other variables that o f t e n i n f l u e n c e death rates are:
(a)
urban or r u r a l residence, perhaps because of u n e q u a l

a v a i l a b i l i t y of h e a l t h care f a c i l i t i e s , l i v i n g standards, or
infrastructures;
(b)
d i f f e r e n t o c c u p a t i o n a l c o m p o s i t i o n s (miners or s o l d i e r s
are m o r e subject to risk than are judges or m o s t profess i o n a l workers);
(c)
d i f f e r e n t i n c o m e c o m p o s i t i o n s (the w e a l t h y can a f f o r d
better m e d i c a l care);
(d)
sex ( w o m e n a l m o s t u n i v e r s a l l y have l o w e r death rates

t h a n m e n at m o s t ages); and
(e)
m a r i t a l status (the m a r r i e d u s u a l l y have l o w e r m o r t a l ity t h a n the single, w i d o w e d , or divorced).
S i n c e w e are interested here i n m e a s u r i n g m o r t a l i t y itself

rather t h a n age or o c c u p a t i o n a l c o m p o s i t i o n , h o w do w e remove, or c o n t r o l for, the effects of these other variables ( w h i c h
are said to " c o n f o u n d " the comparison)?
We
c o u l d s i m p l y l o o k at the detailed s c h e d u l e of age-
s p e c i f i c or o c c u p a t i o n - s p e c i f i c or a g e - o c c u p a t i o n - s p e c i f i c
death rates f o r t w o countries and c o m p a r e t h e m . B u t w o u l d
20
Mortality
it not be u s e f u l to have one single measure, s u c h as the c r u d e

death rate, that has s o m e h o w t a k e n i n t o a c c o u n t the effect
of any extraneous variable b e l i e v e d to i n f l u e n c e the crude
death rate? T o o b t a i n this single measure, demographers u s u a l l y use a t e c h n i q u e k n o w n as
standardization.*
L o o k again at the age c o m p o s i t i o n of c o u n t r i e s A a n d B

i n T a b l e 2.3. If they b o t h had the same age c o m p o s i t i o n , i t is
o b v i o u s that t h e i r crude death rates w o u l d be the same
because they have i d e n t i c a l age-specific rates. In standardi z a t i o n the procedure is to apply the same age c o m p o s i t i o n
(or o c c u p a t i o n c o m p o s i t i o n or whatever) to d i f f e r e n t sets of
s p e c i f i c rates and observe w h a t the crude rate w o u l d t h e n be.
T h e age c o m p o s i t i o n used f o r the s t a n d a r d i z a t i o n is c a l l e d
the standard p o p u l a t i o n . T h e rates used are those of the act u a l p o p u l a t i o n s b e i n g studied.
Age standardization is used to answer the question, H o w
w o u l d the crude death rates of t w o p o p u l a t i o n s c o m p a r e i f
they had e x a c t l y the same age d i s t r i b u t i o n (the
"standard"
w e select) but each retained its o w n d i s t i n c t i v e age-specific

death rates? In this w a y w e " h o l d constant," or c o n t r o l for,
the effect of the age d i s t r i b u t i o n , so that any v a r i a t i o n s i n
the total death rates m u s t result f r o m real differences i n ages p e c i f i c m o r t a l i t y rates b e t w e e n the t w o p o p u l a t i o n s .
T h i s same procedure can be applied to a n y rate c o m parison that w e can separate i n t o t w o parts, (1) the effect of
differences i n the d i s t r i b u t i o n of the c h a r a c t e r i s t i c and (2)
the effect of differences i n the c h a r a c t e r i s t i c - s p e c i f i c rates.
For example, w e c o u l d ask: H o w w o u l d p o p u l a t i o n s A and B
c o m p a r e o n the death rate i f they had the same (standard)
d i s t r i b u t i o n by m a r i t a l status and different m a r i t a l statuss p e c i f i c death rates?
T h e s t a n d a r d i z a t i o n t e c h n i q u e also a p p l i e s to m a n y
fields besides demography and to measures other t h a n rates,
s u c h as ratios or percentages. In a study c o m p a r i n g the percentages of people v o t i n g for a c e r t a i n p o l i t i c a l party i n c i t i e s
4. We discuss only the technique of direct standardization in this Guide.
For an
instructive discussion of indirect standardization, see Barclay 11958, 164-66).
Mortality
21
A a n d B, w e m i g h t ask w h e t h e r the d i f f e r e n c e results f r o m

differences between the t w o cities i n the d i s t r i b u t i o n of people
by age a n d i n c o m e . W e c a n apply the age a n d i n c o m e - s p e c i f i c percentages v o t i n g f o r the party i n each c i t y to a standard age a n d i n c o m e d i s t r i b u t i o n to ascertain w h e t h e r t h e
difference b e t w e e n the cities s t i l l persists or is m o d i f i e d .
T o i l l u s t r a t e s t a n d a r d i z a t i o n , w e again b e g i n w i t h a
s i m p l e , h y p o t h e t i c a l e x a m p l e . T w o countries, A and B , have
the f o l l o w i n g age-specific death rates a n d age c o m p o s i t i o n s :
Country A
Country B
Ages
Midyear
population
Death rate
per 1,000
Midyear
population
Death rate
per 1,000
0-44
45+
1,000
4,000
25
40
4,000
1,000
30
45
T h e crude death rates are 3 7 per 1,000 f o r c o u n t r y A a n d 3 3

per 1,000 f o r c o u n t r y B . Suppose b o t h c o u n t r i e s h a d t h e age
c o m p o s i t i o n of c o u n t r y A . T h e n the crude death rates w o u l d
be 37 f o r c o u n t r y A a n d 4 2 f o r c o u n t r y B . (Try to d u p l i c a t e
this result.) In t h i s case w e w o u l d say that c o u n t r y A w a s t h e
"standard p o p u l a t i o n " a n d that 4 2 w a s t h e " s t a n d a r d i z e d
crude death rate" f o r c o u n t r y B. W e can also standardize t h e
crude death rates u s i n g c o u n t r y B as the standard p o p u l a t i o n . In this case the standardized rates are 28 f o r c o u n t r y A
and 33 f o r c o u n t r y B. N o t e that the c h o i c e of standard affects
the absolute values of the standardized death rates. T h e f o l l o w i n g table s u m m a r i z e s the c a l c u l a t i o n s :
Rate
Country A Country B
Unstandardized crude death rate per 1,000

Standardized crude death rate per 1,000
with country A as the standard
with country B as the standard
37
33
37
28
42
33
C o u n t r y A has a higher crude death rate than c o u n t r y B. W h e n

w e standardize o n the age d i s t r i b u t i o n of either c o u n t r y A or
c o u n t r y B , however, c o u n t r y B has a h i g h e r death rate because the age-specific death rates f o r c o u n t r y B are h i g h e r
than those for c o u n t r y A i n every age group.
A s a n e x a m p l e u s i n g a c t u a l data, w e present the c a l c u lations for an age standardization of crude death rates f o r C h i l e
22
Mortality
i n 1989-90 and Japan i n 1989 (Table 2.5). T h e crude death

rates are 5.73 per 1,000 for C h i l e and 6.40 per 1,000 for Japan.
Standardized o n the age d i s t r i b u t i o n of C h i l e , the Japan death
rate w o u l d be o n l y 3.56. Standardized o n the age d i s t r i b u t i o n
of Japan, the C h i l e death rate w o u l d be 9.67. H e n c e , a l t h o u g h
the unstandardized crude death rate f o r Japan is higher t h a n
C h i l e ' s , the standardized rates ( w i t h e i t h e r c o u n t r y as the
standard population) s h o w that C h i l e has higher death rates.
T h e age c o m p o s i t i o n s of the t w o c o u n t r i e s i n 1989-90 were
c o n s p i c u o u s l y different, C h i l e h a v i n g a m u c h younger p o p u lation.
D e a t h rates at d i f f e r e n t ages i n the s a m e p o p u l a t i o n
tend to be h i g h l y correlated. C o u n t r y A , w i t h l o w death rates
at one age, is l i k e l y to have r e l a t i v e l y l o w death rates at a l l
other ages. C o u n t r y B, w i t h higher death rates at one age,
w i l l u s u a l l y also have higher death rates at a l l ages (the A S D R
curves, as s h o w n i n Figure 2.1, w i l l not cross except perhaps
at the oldest ages). W h e n this is true, i t means that any standard p o p u l a t i o n selected w i l l produce the same results: the
standardized rate for c o u n t r y B w i l l exceed the standardized
rate for c o u n t r y A . T h e reason is that w e are m u l t i p l y i n g the
same set of n u m b e r s (the standard) by higher n u m b e r s for
c o u n t r y B t h a n for c o u n t r y A for every age. O r d i n a r i l y , stand a r d i z a t i o n under these c i r c u m s t a n c e s w i l l at least c l a r i f y
the direction
of the difference. It w i l l s h o w that c o u n t r y B
has higher m o r t a l i t y t h a n c o u n t r y A . O f course, the c h o i c e

of the standard p o p u l a t i o n , even i n t h i s n o r m a l case, c o u l d
affect the amount
of the m o r t a l i t y d i f f e r e n c e b e t w e e n c o u n -
tries A and B. Suppose the m o r t a l i t y difference b e t w e e n c o u n tries A and B is e s p e c i a l l y large for ages 4 0 - 4 9 . In that case
the a m o u n t of the difference i n the standardized rates w i l l
depend o n the p r o p o r t i o n of the standard p o p u l a t i o n that is
i n the age group 4 0 - 4 9 .
O c c a s i o n a l l y the s i t u a t i o n is unclear. It m a y be that
p o p u l a t i o n A has higher death rates than p o p u l a t i o n B at s o m e
ages but not at others. In s u c h a case, not o n l y the amount
the d i f f e r e n c e but also the direction
of
after s t a n d a r d i z a t i o n w i l l
depend o n the standard age d i s t r i b u t i o n selected. In situa-
Table 2.5. Age standardization of crude death rates for Japan (1989) and Chile (1989-90)
Age-specific death
rate (per 1,000)
Ages
<1
1-4
5-9
10-14
15-19
20-24
25-29
30-34
35-39
40-44
45-49
50-54
55-59
60-64
65-69
70-74
75-79
80+
All ages
Chile
(1)
Japan
(2)
17.1
0.8
0.3
0.3
0.7
1.1
1.2
1.5
2.0
3.1
4.5
7.1
11.0
15.9
23.3
38.0
62.8
124.5
4.5
0.4
0.2
0.1
0.4
0.5
0.6
0.6
1.0
1.5
2.3
3.9
6.3
9.1
13.7
24.1
42.1
105.5
Source: United Nations Statistical Office, Demographic

Yearbook
a. Age distribution for Chile is for 1990; deaths are for 1989.
Age-specif:ic death
rate (per 1,000)
Chile
(3)
23
89
103
91
94
94
93
80
68
56
48
38
34
29
22
16
11
11
1,000
Japan
(4)
10
45
62
72
81
72
65
64
77
82
75
65
62
53
39
29
24
23
1,000
1990 (1992, tables 7,19).
b. Total is the Chilean crude death rate.

c. Total is the Chilean death rate standardized on Japan's age distribution.
d. Total is the Japanese crude death rate.
e. Total is the Japanese death rate standardized on Chile's age distribution.
Chilean
deaths with
own age
distribution
(for population
of 1,000)'
(5)
(l)x(3)
.39
.07
.03
.03
.07
.10
.11
.12
.14
.17
.22
.27
.37
.46
.51
.61
.69
1.37
5.73
Chilean
deaths with
Japan's age
distribution
(for population
of 1,000]
(6)
(l)x(4)
.17
.04
.02
.02
.06
.08
.08
.10
.15
.25
.34
.46
.68
.84
.91
1.10
1.51
2.86
9.67'
Japanese
deaths with
own age
distribution
(for population
of 1,000)
(7)
(2) x (4)
Japanese
deaths with
Chile's age
distribution
(for population
of 1,000)
(8)
(2) x (3)
.05
.02
.01
.01
.03
.04
.04
.04
.08
.12
.17
.25
.39
.48
.53
.70
1.01
2.43
6.40
.10
.04
.02
.01
.04
.05
.06
.05
.07
.08
.11
.15
.21
.26
.30
.39
.46
1.16
3.56=
24
Mortality
tions like this the process of standardization depends on the

arbitrary choice of a standard and the results are probably
misleading and not very worthwhile. Here is a simple hypothetical case:
Country A
Country B
Ages
Midyear
population
Death rate
per 1,000
Midyear
population
Death rate
per 1,000
0-44
45+
1,000
4,000
35
50
4,000
1,000
25
75
The crude death rates and the standardized rates are summarized in the following table:
Rate
Country A Country B
Unstandardized crude death rate per 1,000

Standardized crude death rate per 1,000
with country A as the standard
with country B as the standard
47
35
47
38
65
35
In this case, country A has a higher crude death rate than

country B and also has a higher standardized death rate if we
standardize on country B's age distribution. If we standardize on country A's age distribution, however, country B has a
higher standardized rate. The choice of the standard population thus reverses the direction of our answer! In cases like
this, techniques other than standardization, such as the life
table, are often used to summarize the underlying mortality
situation. We will shortly examine life-table functions that
might be used.
Fortunately, cases like the one just presented are not
common. For this reason, and because standardization is relatively easy to use, it is widely used by demographers. T h e
formulas for the age standardization of death rates are given
in Table 2.6. A comparison of the formulas for populations A
and B in the last two rows of the table shows what standardization doesnamely, it uses the age composition of the standard population as the weights in obtaining the weighted sum
of age-specific rates that cumulate to form the standardized
crude death rate.
Developing countries often have low crude death rates
(see Table 2.7). Their populations are very youngthat is,
Mortality
25
t h e y have large p r o p o r t i o n s of p e o p l e i n the y o u n g e s t age

groupsas a result of t h e i r recent h i g h b i r t h rates. A g e stand a r d i z a t i o n u s i n g the age c o m p o s i t i o n of t h e U n i t e d States
i n 1980 s h o w s that m o s t d e v e l o p i n g c o u n t r i e s w o u l d h a v e
m u c h h i g h e r c r u d e death rates i f t h e y h a d the U . S . age cornTable 2.6. Formulas for direct age-standardization of the crude death rate for two hypothetical populations
Formula for:
Item
Population A
Population B
Number of people
in age group x,x+n
P!
P =l Px
Total population
JC
Deaths in age group

x,x+n
Deaths in total
population
Death rate in age

group x,x+n
D
M
"
= " *
"
pA
n x
Crude death rate
JC
J^A
Ami
"
n x
y
D
Z j n^x
P ~lA
lA
XU )Ur)
B
/ >n x
X
pB
Death rate standardized

on age distribution of
Population B
= -
"
Death rate standardized

on age distribution of
Population A
XU')U')
y
JC
Table 2.7. Standardized death rates for selected places and years
Standardized death rate
using as the standard:
Crude
death
rate
United
States,
1980
Mexico,
1980
7.78
7.21
16.48
16.26
7.80
6.71
10.50
7.90
15.59
12.38
8.54
5.58
Chile, 1970
Chile, 1980
8.78
6.67
12.98
11.27
7.93
5.09
Mexico, 1970
Mexico, 1983
9.49
5.54
14.40
11.33
8.56
5.44
Venezuela, 1970
Venezuela, 1985
6.46
4.56
12.14
9.52
6.35
4.32
Canada, 1970
Canada, 1980
Canada, 1985
7.32
7.13
7.15
9.39
8.15
7.55
3.99
3.29
2.92
United States, 1970

United States, 1980
United States, 1985
9.43
8.77
8.74
10.37
8.77
8.27
4.54
3.65
3.35
Taiwan, 1970
Taiwan, 1980
Taiwan, 1985
4.90
4.76
4.80
12.69
10.20
9.40
5.27
4.11
3.64
Hong Kong, 1970

Hong Kong, 1980
5.24
4.98
10.32
7.99
4.46
3.18
Japan, 1970
Japan, 1980
6.83
6.19
10.71
7.95
4.15
2.93
Peninsular Malaysia, 1970

6.99
5.55
5.27
12.64
11.54
11.08
6.64
5.20
4.69
Singapore, 1970
Singapore, 1980
Singapore, 1985
5.16
5.18
5.22
12.61
10.99
10.37
5.23
4.23
3.82
Austria, 1970
Austria, 1980
Austria, 1985
13.23
12.32
11.85
12.15
10.32
9.42
4.99
4.06
3.57
Belgium, 1970
Belgium, 1980
Belgium, 1985
12.30
11.54
11.22
11.39
9.83
9.05
4.58
3.80
3.44
Bulgaria, 1970
Bulgaria, 1980
Bulgaria, 1985
9.08
11.05
12.00
10.83
11.51
12.00
4.51
4.49
4.54
Denmark, 1970
Denmark, 1980
Denmark, 1985
9.79
10.92
11.42
9.42
9.05
8.82
3.71
3.44
3.35
France, 1970
France, 1980
10.63
10.15
10.09
8.80
4.10
3.51
Country and year

Mauritius, 1970
Mauritius, 1980
South Africa (white), 1970
South Africa (white), 1985

Standardized death rate
using as the standard:
Country and year
Crude
death
rate
United
States,
1980
Mexico,
1980
France, 1985
10.01
8.17
3.19
Fed. Rep. Germany, 1970

12.12
11.60
11.54
11.80
9.79
8.87
4.80
3.80
3.31
Greece, 1970
Greece, 1980
Greece, 1985
8.76
9.80
10.20
9.04
9.04
8.80
4.09
3.56
3.34
Hungary, 1970
Hungary, 1980
Hungary, 1985
11.63
13.57
11.86
12.50
12.89
12.68
5.28
5.19
5.15
Italy, 1970
Italy, 1980
Italy, 1983
9.68
9.83
9.93
10.27
9.42
9.08
4.38
3.61
3.39
Netherlands, 1970
Netherlands, 1980
Netherlands, 1985
8.41
8.08
8.47
9.51
8.05
7.84
3.66
3.04
2.89
Norway, 1970
Norway, 1980
Norway, 1985
9.99
10.12
10.67
9.15
8.13
7.92
3.54
3.05
2.99
Poland, 1970
Poland, 1980
Poland, 1985
8.20
9.84
10.25
11.98
11.63
11.77
5.11
4.79
4.67
8.42
7.71
7.93
9.97
8.50
7.90
4.25
3.23
3.02
Sweden, 1970
Sweden, 1980
Sweden, 1985
9.95
11.05
11.26
8.74
8.19
7.64
3.34
3.01
2.77
England and Wales, 1970

11.76
11.83
10.72
9.05
4.20
3.32
Australia, 1971
Australia, 1980
Australia, 1985
8.47
7.40
7.53
1081
8.68
8.23
4.38
3.41
3.17
Fiji, 1975
Fiji, 1980
6.89
6.40
13.92
12.39
7.13
6.37
New Zealand, 1970

New Zealand, 1980
New Zealand, 1985
8.78
8.52
8.38
13.92
10.18
9.30
4.93
3.96
3.59
Former USSR, 1979

Former USSR, 1987
10.05
9.85
12.41
11.16
5.83
5.12
Spain, 1970
Spain, 1980
Spain, 1983
Source: Keyfitz and Flieger (1990, 294-583].
28
Mortality
p o s i t i o n . For example, M a u r i t i u s had a crude death rate of

7.2 i n 1980. Standardized on the age c o m p o s i t i o n of the U n i t e d
States, the rate w o u l d be above 16. S i m i l a r results are e v i dent f r o m s t a n d a r d i z i n g the rates f o r M e x i c o , Singapore, a n d
Malaysia.
Y o u s h o u l d m a k e some a d d i t i o n a l c o m p a r i s o n s u s i n g
T a b l e 2.7 to acquire an i n t u i t i v e f e e l i n g f o r the effects of age
s t a n d a r d i z a t i o n . N o t i c e , f o r example, that the amount
of the
differences i n rates is affected by the standard p o p u l a t i o n used.

T h e r e is a difference of 1.01 p o i n t s between the c r u d e death
rates of Japan and Singapore i n 1980, Japan's rate b e i n g higher.
W h e n the rates are standardized o n the age c o m p o s i t i o n of
the U n i t e d States, the difference is 3.04, but in the
direction,
other
Singapore's rate b e i n g higher. W h e n they are stan-
dardized o n the age d i s t r i b u t i o n of M e x i c o , the d i f f e r e n c e is

1.30 points, w i t h Singapore's rate higher.
A s a second example, the trend i n the C D R s can be
d i f f e r e n t f r o m the trend i n the standardized rates. L o o k at
the data f o r D e n m a r k . T h e crude death rate rose f r o m 1970
to 1985, w h i l e the standardized rates f e l l . A g e - s p e c i f i c m o r t a l i t y rates were a c t u a l l y f a l l i n g i n D e n m a r k d u r i n g that period, but t h i s i m p r o v e m e n t i n m o r t a l i t y w a s not r e f l e c t e d i n
the C D R because D e n m a r k ' s p o p u l a t i o n was also aging duri n g the same t i m e period.
S E C O N D SET O F
1. In c o u n t r i e s A and B the age-specific death rates per 1,000
MULTIPLE-CHOICE
are as f o l l o w s :
QUESTIONS
Ages
Country A
Country B
0-4
5-24
25-54
55+
40
20
25
60
29
19
22
58
If the crude death rates for the t w o c o u n t r i e s are standardized o n the same age d i s t r i b u t i o n , w h i c h of the f o l l o w i n g is
true?
(a)
T h e standardized death rate is higher i n c o u n t r y A t h a n

i n c o u n t r y B.
Mortality
(b)
29
T h e standardized death rate is higher i n c o u n t r y B t h a n

in country A .
(c)
T h e standardized death rates i n the t w o c o u n t r i e s are

equal.
(d)
T h e s t a n d a r d i z e d d e a t h rates i n c o u n t r y A m a y be
higher, lower, or e q u a l to those of c o u n t r y B .
2. T h e range of values f o r n a t i o n a l crude death rates i n the

w o r l d today is about:
(a)
10 to 80.
(b)
2 to 25.
(c)
10 to 120.
(d)
2 to 150.
3. D e a t h rates are standardized:

(a)
to e l i m i n a t e the d i f f e r e n t i a l i n f l u e n c e of one or m o r e
variables.
(b)
to o b t a i n an estimate of the ideal rates.
(c)
to d e t e r m i n e the f u t u r e rates that m a y be expected.
(d)
to o b t a i n a correct statement of the a c t u a l or experienced rates.
(e)
to correct for underregistration of the p h e n o m e n o n i n

question.
4. A h i g h sex ratio:
(a)
indicates a h i g h p r o p o r t i o n of m a l e s i n the p o p u l a t i o n .
(b)
indicates a l o w p r o p o r t i o n of m a l e s i n the p o p u l a t i o n .
(c)
indicates a h i g h p r o p o r t i o n of i n f a n t s i n the p o p u l a tion.
(d)
measures the extent of m o r t a l i t y to males i n the p o p u lation.
(e)
EXERCISE 2
is i m m o r a l .
In 1988 c i t y A h a d a crude death rate of 15 per 1,000 a n d c i t y

B a crude death rate of 9. In the same year the crude death
rate of the U n i t e d States w a s about 9.
T h e age-specific death rates of the t w o c i t i e s are standardized o n the age d i s t r i b u t i o n of the U n i t e d States as a
w h o l e i n 1988. C o n s i d e r each of the f o l l o w i n g s i x p o s s i b l e
30
Mortality
results and i n d i c a t e w h a t i n t e r p r e t a t i o n y o u w o u l d give t h e m

in the absence
of any othei
information:
Death rates per 1,000 standardized on the

age distribution of the United States
Results
City A
Crude death rate

Standardized rates:
Case 1
Case 2
Case 3
Case 4
Case 5
Case 6
EXERCISE 3
City B
15
15
15
9
9
12
7
9
15
9
15
10
10
Y o u are g i v e n the f o l l o w i n g data f o r c o u n t r i e s A and B :

Country A
Areas
Metropolitan
areas
Small towns
Rural areas
Country B
Midyear
population
Death rate
per 1,000
500
1,500
8,000
20
35
40
Midyear
Death rate
population per 1,000
6,000
1,500
2,500
25
40
45
C a l c u l a t e the crude death rates f o r each c o u n t r y . A l s o c a l c u late the area-standardized death rates, u s i n g (1) c o u n t r y A as
the standard p o p u l a t i o n and (2) c o u n t r y B as the standard
p o p u l a t i o n . C o m p a r e the answers and interpret t h e m .
THE INFANT
E s t i m a t i n g the n u m b e r of person-years l i v e d f o r c h i l d r e n
MORTALITY RATE
under age 1 is u s u a l l y d i f f i c u l t because the r e q u i s i t e statist i c s are n o t c o l l e c t e d or not p u b l i s h e d even i f c o l l e c t e d . Furt h e r m o r e , for the reasons given i n C h a p t e r 1, the m i d y e a r
p o p u l a t i o n is u s u a l l y a poor e s t i m a t e of the n u m b e r of person-years l i v e d i n the age group under 1. H e n c e , demographers use a special m e t h o d f o r c a l c u l a t i n g m o r t a l i t y f o r c h i l dren under 1 year of age. T h e y c a l l c h i l d r e n under age 1 " i n f a n t s " and c a l c u l a t e the infant
mortality
rate* (IMR)
accord-
ing to the f o l l o w i n g f o r m u l a :
5. Barclay (1958| calls this rate the infant death rate to distinguish it from another
Mortality
31
where D = number of deaths to children under 1 year of age in year z,

B' = number of live births in year z, and
k = 1,000.
T h e i n f a n t m o r t a l i t y rate is t h u s closer to b e i n g a p r o b a b i l i t y
than a rate, since the d e n o m i n a t o r is persons (infants) exposed to death b e g i n n i n g at a c e r t a i n t i m e (birth), rather t h a n
the n u m b e r of person-years l i v e d by i n f a n t s .
A s w e m e n t i o n e d i n C h a p t e r 1, there is a special pattern of m o r t a l i t y d u r i n g the f i r s t year of l i f e . T h i s is i l l u s trated by the data f o r selected c o u n t r i e s and years i n T a b l e
2 . 8 / D e a t h s are not e v e n l y d i s t r i b u t e d t h r o u g h o u t the f i r s t
year of l i f e . Instead, a h i g h p r o p o r t i o n of i n f a n t m o r t a l i t y
o c c u r s i n the f i r s t m o n t h of l i f e . F u r t h e r m o r e , a h i g h proport i o n of the deaths i n the f i r s t m o n t h of l i f e o c c u r s d u r i n g the
first w e e k , and a h i g h p r o p o r t i o n of the deaths i n the f i r s t
w e e k of l i f e o c c u r s d u r i n g the v e r y f i r s t day. M o r t a l i t y of
c h i l d r e n under 28 days of age is generally a l m o s t as h i g h as.
or even higher t h a n m o r t a l i t y i n the next f i v e m o n t h s c o m bined; m o r t a l i t y rates f o r the second half of the f i r s t year are
a l w a y s less t h a n half and u s u a l l y less t h a n o n e - t h i r d of those
for the f i r s t s i x m o n t h s .
Figures f r o m countries w i t h good data u s u a l l y s h o w that
the l o w e r the i n f a n t m o r t a l i t y rate, the higher is the proport i o n of deaths that o c c u r i n the f i r s t m o n t h , the f i r s t w e e k ,
and even the f i r s t h o u r of l i f e . T h i s is so because the causes
type of rate applied to infants in life tables (See Barclay 1958,47 ff., 106 ff., and 138
ff.]. We prefer the present usage, however, to maintain consistency with tables in
the United Nations Demographic
Yearbook
series and other common reference
materials.
6. When constructing life tables, demographers often use the IMR as the value for
,</, the probability of dying between birth and the first birthday. (See the next
section of this.chapter.)
7. Data on the number of days and weeks within infancy are not available for
many developing countries. The examples shown in Table 2.8 are not necessarily
based on good data and hence do not necessarily exhibit the "ideal" patterns
described in the text. In particular, the low values for the death rate in the first day
of life for Egypt, Pakistan, and Albania are suspect, the actual values are probably
higher.
32
Mortality
of very early i n f a n t deaths t e n d to be c o n g e n i t a l m a l f o r m a t i o n , b i r t h injuries, p r e m a t u r i t y , a n d other causes that are

n o t easily prevented by m o d e r n m e d i c a l a n d h e a l t h measures.
C a u s e s of later i n f a n t deaths (such as i n f e c t i o u s diseases o r
poor n u t r i t i o n ) are m o r e susceptible to p r e v e n t i o n or treatment. Hence, whenever public health and individual medical care i m p r o v e , late i n f a n t deaths d i m i n i s h faster t h a n early
deaths, a n d a higher p r o p o r t i o n of a l l i n f a n t deaths are early
deaths.
Three additional problems i n measuring infant mortali t y are caused by the f o l l o w i n g facts: (a) there are seasonal
f l u c t u a t i o n s i n the n u m b e r of births; (b) m a n y babies are b o r n
Table 2.8. Infant mortality rates for selected countries, by age and region: recent years
Age
7-28
days
Age
29^364
days
All ages
<1 year
5.4(10.9)
"69.3 (72.7)
6.4(13.0)
36.1(73.1)
* 101.9 (100.1)
49.4(100.0)
3.4 (7.3)
2.6(36.6)
5.3(11.4)
1.2(16.9)
6.9(14.8)
0.7 (9.9)
31.0(66.5)
2.6(36.6)
46.6(100.0)
7.2(100.0)

Pakistan, 1988"
0.1 (0.1)
Japan, 1989
1.0(21.7)
36.6(34.0)
0.9(19.6)
22.8(21.2)
0.7(15.2)

Albania, 1989
0.9 (2.9)
Netherlands, 1989
1.8(26.5)
3.5(11.4)
2.0(29.4)
3.1(10.1)
0.8(11.8)
23.4(76.0)
2.2(32.4)
30.8(99.9)
6.8(100.1)
Oceania
Australia, 1989
1.4(17.5)
0.9(11.3)
3.3(41.3)
8.0(100.1)
2.3(10.0)
12.9(56.1)
23.0(100.0)
Region, country,
and year
Age
<1 day
Africa
Algeria, 1980'
Egypt, 1987
"26.1 (27.4)
1.5 (3.0)
Americas
Guatemala, 1988
Canada, 1988
Former USSR
2.4(30.0)
Age
1-6
days
7.8(33.9)
48.2(44.8)107.7(100.1)
2.0(43.5)
4.6(100.0)
Notes: Rates are the number of deaths of infants per 1,000 live births. Figures in parentheses are percentages of the
total. The rates for specific ages are based on the same denominator (1,000 live births) as is the total. Consequently,
the sum of the rates for the specific ages equals the total infant mortality rate shown in the last column. Totals may
not equal the sum of constituent rates or percentages because of rounding. Rates are shown only for countries
having at least 1,000 infant deaths in a given year and with populations of 1 million or more. Data from registers
that are incomplete or of unknown completeness are indicated with an asterisk ('). Ranges may not encompass the
actual worldwide range because few countries have the requisite data available.
a. Excludes live-born infants who died before their births were registered. Not included in the calculations of
percentages are 6.6 infant deaths/1,000 births of unknown age.
b. Based on Pakistan's Population Growth Survey.
Mortality
and
33
die i n the same calendar year and are o m i t t e d i n c o u n t s
of the p o p u l a t i o n under age 1 at b o t h the b e g i n n i n g and the

end of the year,- a n d (c) i n m o s t censuses and surveys, i n f a n t s
are o f t e n u n d e r e n u m e r a t e d m o r e than older persons, appare n t l y because m a n y parents do n o t t h i n k of i n f a n t s as persons w h e n asked " H o w m a n y persons l i v e here?"
Infant m o r t a l i t y rates d u r i n g the 1980s ranged f r o m 4.5
per 1,000 b i r t h s (in Japan) to 172 (in A f g h a n i s t a n ) , as s h o w n
i n T a b l e 2.9. I M R s were m u c h h i g h e r i n the past. Rates as
h i g h as 200 were recorded f o r B e l g i u m i n 1900, France i n the
p e r i o d 1 8 5 1 - 1 9 0 3 , a n d S w e d e n i n the p e r i o d 1 7 7 8 - 1 8 3 2
( K e y f i t z and Flieger 1968, 24-39). T h i s means, roughly, that
for every f i v e i n f a n t s b o r n i n B e l g i u m , France, or S w e d e n
d u r i n g those years, one died before its f i r s t birthday.
The
i n f a n t m o r t a l i t y rate c a l c u l a t e d i n the s i m p l e w a y
described above is reliable o n l y w h e n the n u m b e r of b i r t h s

does not change r a p i d l y f r o m one calendar year to the next.
W h e n there are rapid y e a r l y changes i n the n u m b e r of b i r t h s ,
adjusted rates of v a r i o u s k i n d s are needed and can be c a l c u lated; b u t w e delay o u r d i s c u s s i o n of these u n t i l A p p e n d i x 2.
FIRST SET O F
D e t e r m i n e w h e t h e r each of the f o l l o w i n g s t a t e m e n t s is true
TRUE/FALSE
or false:
QUESTIONS
Infant m o r t a l i t y rates are generally h i g h e r i n the less

developed c o u n t r i e s t h a n they are i n the m o r e d e v e l oped countries.
2.
A c c o r d i n g to available data, crude death rates i n the

less developed countries are a l w a y s h i g h e r t h a n those
i n the m o r e developed c o u n t r i e s .
3.
T h e m o r e developed countries probably never had i n fant m o r t a l i t y rates as h i g h as those n o w recorded i n

m a n y less developed c o u n t r i e s .
4.
T h e m i d y e a r p o p u l a t i o n is a l w a y s a good e s t i m a t e of
the person-years l i v e d i n a g i v e n year.
5.
O n the average, age-specific death rates are h i g h o n l y

for persons over age 65.
Table 2.9. Highest and lowest infant mortality rates, by region: latest
available data
Infant mortality rate
(per 1,000 live births)
Region and country
Year
High
Africa
Sierra Leone
Malawi
Mauritius
Tunisia
1985-90
1985-90
1985-90
1985-90
"154.0
"150.0
Americas
Haiti
Bolivia
Canada
UnitedStates
1985-90
1985-90
1988
1990
"97.0
"110.0

Afghanistan
Cambodia
Japan
Hong Kong
1985-90
1985-90
1990
1990
"172.0
"130.0

Romania
1990
Yugoslavia
1990
Sweden
1990
Finland
1989
Oceania
Papua New Guinea
Australia
New Zealand
1985-90
1989
1989
Former USSR'
Turkmenistan
Belarus
1989
1990-91
1990-91
Low
"23.0
"52.0
7.2
t9.1
4.5
t6.1
'26.9
20.2
t5.6
5.8
"59.0
8.0
10.2
23.0
93
20
Sources: Except for the former USSR, all data are from United Nations Statistical
Office, Demographic
Yearbook 1990 (1992, table 15). Data for the former Soviet
republics are from Population Reference Bureau (1992).
Note: Rates are shown only for countries having at least 100 infant deaths in the
specified year and a population of 1 million or more. Data from registers that are
incomplete or of unknown completeness are not included.
* Estimate prepared by Population Division of the United Nations.
t Provisional figure.
a. The U N Demographic
Yearbook reports a figure of 39 deaths per 1,000 births for
the former Soviet republics.
Mortality
T H E LIFE T A B L E
35
Rates and ratios p r o v i d e us w i t h a u s e f u l set of measures f o r

a n s w e r i n g questions about m o r t a l i t y . T h e r e are s t i l l m a n y
questions, however, that w e cannot answer w i t h these measures alone. T o c i t e a f e w examples:
1. O u t of 100 persons i n c o u n t r y A w h o were 20 years o l d
i n 1968, h o w m a n y are l i k e l y to l i v e to age 50?
2.
I m m e d i a t e l y after b i r t h , h o w m a n y years c o u l d a c h i l d
b o r n i n 1950 i n c o u n t r y B expect to live?
3.
A m o n g y o u n g m e n and w o m e n e n t e r i n g the labor f o r c e

at ages 2 0 - 2 4 , w h a t p r o p o r t i o n can be expected to be
a l i v e at age 67 (when, for e x a m p l e , i n the U n i t e d States
they are e n t i t l e d to c o l l e c t s o c i a l s e c u r i t y benefits)? O f
those w h o do begin to c o l l e c t benefits at age 67, h o w
m a n y can be expected to s u r v i v e f o r one year, t w o years,
three years, etc.?
4.
Is there a measure that c a n be used to c o m p a r e the

m o r t a l i t y of m a n y c o u n t r i e s so that differences i n t h e i r
age d i s t r i b u t i o n s w i l l not be d i s t o r t i n g factors a n d so
that an arbitrary c h o i c e of a standard p o p u l a t i o n f o r an
age s t a n d a r d i z a t i o n w i l l not be necessary?
Q u e s t i o n s of this type have i m m e n s e p r a c t i c a l i m p o r -
tance. F o r e x a m p l e , p r o j e c t i o n s of the f u t u r e p o p u l a t i o n
needed to d e t e r m i n e h o w m a n y schools or h o s p i t a l s are req u i r e d depend o n estimates of h o w l o n g people s u r v i v e . In
a d d i t i o n , l i f e i n s u r a n c e c o m p a n i e s need accurate answers to
questions about average l i f e expectancy, f o r w i t h o u t t h e m
they w o u l d not be able to c o n s t r u c t the a c t u a r i a l tables o n
w h i c h they base the p r e m i u m s c u s t o m e r s m u s t pay. S u c h
questions as these can best be answered by life
tables,
al-
t h o u g h the answers are s t i l l a p p r o x i m a t e .

C o n s t r u c t i n g a l i f e table can be a c o m p l e x process. H e r e ,
w e e m p h a s i z e i n t e r p r e t a t i o n rather t h a n c o m p u t a t i o n , beg i n n i n g w i t h a d e s c r i p t i o n of the l i f e table and a f t e r w a r d discussing elementary applications i n demographic analysis.
Instead of the m o r e u s u a l n o t i o n of a p o p u l a t i o n , suppose w e were to d e f i n e a p o p u l a t i o n to be everyone b o r n i n a
c o u n t r y d u r i n g a p a r t i c u l a r year, say 1879. A s demographers,
w e c a l l t h i s group the "1879 b i r t h c o h o r t " f o r that c o u n t r y .
36
Mortality
N o w suppose w e had the death rates for the 1879 b i r t h cohort as i t passed t h r o u g h each age, u n t i l every m e m b e r of the
1879 cohort had died (presumably a l l w o u l d h a v e d i e d b y
now). In t h i s s i t u a t i o n , w e c o u l d e a s i l y a n s w e r q u e s t i o n s
about the s u r v i v a l of m e m b e r s of the c o h o r t f r o m one age to
the next, s i n c e w e w o u l d k n o w t h e i r entire m o r t a l i t y history. F r o m data of this type, w e c o u l d construct w h a t is k n o w n
as a longitudinal,
or generation,
life table,
w h i c h refers to
one b i r t h c o h o r t as i t ages. A generation l i f e table c a n be

c o n s t r u c t e d o n l y after a l l or a l m o s t a l l of the m e m b e r s of the
b i r t h cohort have died. F o r t h i s reason, and because the req u i r e d data are not o f t e n available, generation l i f e tables have
l i m i t e d practical u t i l i t y and are not c o m m o n l y used, a l t h o u g h
h i s t o r i c a l studies of m o r t a l i t y benefit greatly f r o m s u c h tables.
Let us r e t u r n to our m o r e u s u a l d e f i n i t i o n of a p o p u l a t i o n . Suppose w e have a set of age-specific death rates that
represent the i n c i d e n c e of m o r t a l i t y i n each age group f o r a
cross-section of the p o p u l a t i o n over a short p e r i o d of t i m e (a
year, 1990, f o r example). W e a s s u m e that the a g e - s p e c i f i c
m o r t a l i t y experiences d u r i n g 1990 represent the death experience of a w h o l e generation of persons. T h a t is, w e a s s u m e
that a c o h o r t of persons w i l l pass t h r o u g h l i f e e x p e r i e n c i n g
at each age the age-specific death rates f o r 1990. M a k i n g t h i s
a s s u m p t i o n , we can d e t e r m i n e w h a t the n u m b e r of s u r v i vors at any given age w o u l d be out of an i n i t i a l group of births,
according to the given m o r t a l i t y schedule. T h e l i f e table, then,
b e c o m e s a m o d e l of w h a t w o u l d happen to a h y p o t h e t i c a l
b i r t h cohort if the age-specific
were
to remain
experience
constant
of an entire
death
rates for a given
and were to apply

generation.
period
throughout
the
M o r t a l i t y analyses based
o n l i f e tables are n o r m a l l y based o n the a s s u m p t i o n that a

single m o r t a l i t y schedule applies to a hypothetical
persons
group
of
u n t i l a l l the persons have died. In other w o r d s , the
m o r e c o m m o n type of l i f e table is c a l c u l a t e d for a

or hypothetical,
cohort
synthetic,
of b i r t h s .
T h i s m o r e c o m m o n type of l i f e table is c a l l e d a
cross-sectional,
current,
or time-specific
life table.
period,
It answers
the q u e s t i o n , W h a t w o u l d be the m o r t a l i t y h i s t o r y a n d aver-
Mortality
37
age l i f e expectancy of a cohort of people subject t h r o u g h o u t

t h e i r l i f e h i s t o r y to the age-specific death rates of a p a r t i c u l a r
year or p e r i o d of years? T h e p e r i o d l i f e table is a m a t h e m a t i c a l m o d e l of the l i f e h i s t o r y of a h y p o t h e t i c a l cohort. It is a
m o d e l because w e m u s t m a k e s i m p l i f y i n g a s s u m p t i o n s to
c o n s t r u c t the table and because i t refers to a h y p o t h e t i c a l
rather than a real b i r t h cohort.
T h e l i f e table begins w i t h the b i r t h , d u r i n g one year, of
a h y p o t h e t i c a l c o h o r t of persons. T h e n u m b e r of b i r t h s is
u s u a l l y set a r b i t r a r i l y at 100,000. T h i s s t a r t i n g n u m b e r of
8
b i r t h s is c a l l e d the radix of the l i f e table. T h e life-table r e c o r d

c o n t i n u e s u n t i l a l l the m e m b e r s of the c o h o r t have d i e d ,
deaths at each age o c c u r r i n g i n accordance w i t h a m o r t a l i t y
s c h e d u l e that is f i x e d i n advance and does n o t change. N o
factors other t h a n m o r t a l i t y operate to reduce the s i z e of the
s t a r t i n g cohort; that is, the h y p o t h e t i c a l cohort is " c l o s e d "
to m i g r a t i o n of any k i n d . A t each age, except f o r the f i r s t f e w
years of l i f e , the deaths are a s s u m e d to be e v e n l y d i s t r i b u t e d
t h r o u g h o u t the year. H e n c e , half of the deaths to persons
between the ages 15 and 16, say, w o u l d o c c u r by the t i m e the
average person i n the cohort has reached the age of 1 5 V i . M o s t
l i f e tables refer to o n l y one sex, p r i m a r i l y because the death
rates f o r m a l e s a n d females d i f f e r s u b s t a n t i a l l y . L i f e tables
can be s p e c i f i c to any subgroup w i t h i n a p o p u l a t i o n , p r o v i d e d
the necessary A S D R s are available.
T h e c o n v e n t i o n a l l i f e table consists of seven c o l u m n s ,
s i x of w h i c h present w h a t are c a l l e d the life-table
functions.
A brief d e s c r i p t i o n of each c o l u m n f o l l o w s . T o i l l u s t r a t e , w e
use a l i f e table f o r females i n the U n i t e d States f o r the years
1979-81 (Table 2.10).
COLUMN
t: E X A C T A C E : x
E a c h of the life-table f u n c t i o n s refers to a s p e c i f i c age or age

i n t e r v a l . T h e first c o l u m n of the l i f e table s p e c i f i e s the age to
8. Other starting numbers are found in the literature, the most common being 1
and 10,000.
Table 2.10. Complete life table for females: United States, 1979-81
Exact age
in years
Probability
of dying
between
exact
age x and Number of
survivors at
exact
age x+1
exact age x
Number
of deaths Number of
between years lived
between
exact
age x and exact age x
exact
and exact
age x+1
age x+1
<?,
(1)
(2)
(3)
0
1
2
3
4
.01120
.00086
.00056
.00042
.00033
100,000
98,880
98,796
98,740
98,699
1,120
84
56
41
33
99,085
98,838
98,768
98,720
98,682
7,762,496
7,663,411
7,564,573
7,465,805
7,367,085
77.62
77.50
76.57
75.61
74.64
5
6
7
8
9
.00031
.00027
.00024
.00022
.00019
98,666
98,636
98,609
98,585
98,563
30
27
24
22
19
98,651
98,623
98,596
98,575
98,553
7,268,403
7,169,752
7,071,129
6,972,533
6,873,958
73.67
72.69
71.71
70.73
69.74
10
11
12
13
14
.00018
.00018
.00020
.00026
.00033
98,544
98,527
98,509
98,489
98,464
17
18
20
25
32
98,536
98,518
98,499
98,477
98,448
6,775,405
6,676,869
6,578,351
6,479,852
6,381,375
68.75
67.77
66.78
65.79
64.81
15
16
17
18
19
.00040
.00047
.00052
.00055
.00057
98,432
98,392
98,346
98,294
98,240
40
46
52
54
56
98,411
98,369
98,320
98,267
98,212
6,282,927
6,184,516
6,086,147
5,987,827
5,889,560
63.83
62.86
61.89
60.92
59.95
20
21
22
23
24
.00058
.00060
.00062
.00063
.00064
98,184
98,127
98,068
98,007
97,946
57
59
61
61
63
98,156
98,097
98,037
97,977
97,914
5,791,348
5,693,192
5,595,095
5,497,058
5,399,081
58.98
58.02
57.05
56.09
55.12
25
26
27
28
29
.00065
.00066
.00067
.00070
.00072
97,883
97,820
97,755
97,689
97,621
63
65
66
68
70
97,851
97,788
97,722
97,655
97,586
5,301,167
5,203,316
5,105,528
5,007,806
4,910,151
54.16
53.19
52.23
51.26
50.30
30
31
32
33
34
.00075
.00079
.00083
.00089
.00096
97,551
97,477
97,400
97,319
97,233
74
77
81
86
93
97,514
97,439
97,360
97,276
97,186
4,812,565
4,715,051
4,617,612
4,520,252
4,422,976
49.33
48.37
47.41
46.45
45.49
35
36
37
38
39
.00104
.00114
.00125
.00127
.00149
97,140
97,039
96,928
96,807
96,675
101
111
121
132
144
97,890
96,984
96,868
96,741
96,603
4,325,790
4,228,701
4,131,717
4,034,849
3,938,108
44.53
43.58
42.63
41.68
40.74
Expectation
Total
of life
(average
number
number of
of years
years
lived
remaining)
after exact
age x
at exact age x
(4)
T,
(5)
(6)
(7)
Table 2.10.
Exact age
in years
X
(continued)
Probability
of dying
between
exact
age x and Number of
exact
survivors at
age x+1
exact age x
Number
of deaths
between
exact
age x and
exact
age x+1
d
Number of
years lived
between
exact age x
and exact
age x+1
Expectation
Total
of life
number
(average
of years
number of
lived
years
after exact
remaining)
age x
at exact age x
T
(1)
Q,
(2)
(3)
(4)
K
(5)
40
41
42
43
44
.00163
.00180
.00199
.00218
.00239
96,531
96,374
96,200
96,009
95,799
157
174
191
210
229
96,452
96,287
96,104
95,904
95,684
3,841,505
3,745,053
3,648,766
3,552,662
3,456,758
39.80
38.86
37.93
37.00
36.08
45
46
47
48
49
.00262
.00286
.00315
.00347
.00381
95,570
95,320
95,047
94,748
94,419
250
273
299
329
359
95,445
95,184
94,897
95,584
94,239
3,361,074
3,265,629
3,170,445
3,075,548
2,980,964
35.17
34.26
33.36
32.46
31.57
50
51
52
53
54
.00416
.00452
.00490
.00532
.00578
94,060
93,669
93,245
92,788
92,294
391
424
457
494
534
93,864
93,457
93,017
92,541
92,028
2,886,725
2,792,861
2,699,404
2,606,387
2,513,846
30.69
29.82
28.95
28.09
27.24
55
56'
57
58
59
.00627
.00678
.00733
.00796
.00867
91,760
91,185
90,567
89,903
89,187
575
618
664
716
773
91,472
90,876
90,235
89,545
88,800
2,421,818
2,330,346
2,239,470
2,149,235
2,059,690
26.39
25.56
24.73
23.91
23.09
60
61
62
63
64
.00947
.01035
.01129
.01226
.01325
88,414
87,577
86,670
85,691
84,641
837
907
979
1,050
1,121
87,996
87,123
86,181
85,166
84,081
1,970,890
1,882,894
1,795,771
1,709,590
1,624,424
22.29
21.50
20.72
19.95
19.19
65
66
67
68
69
.01427
.01538
.01664
.01811
.01980
83,520
82,328
81,061
79,712
78,269
1,192
1,267
1,349
1,443
1,549
82,923
81,695
80,387
78,990
77,495
1,540,343
1,457,420
1,375,725
1,295,338
1,216,348
18.44
17.70
16.97
16.25
15.54
70
71
72
73
74
.02169
.02375
.02600
.02842
.03106
76,720
75,055'
73,273
71,368
69,340
1,665
1,782
1,905
2,028
2,154
75,887
74,164
72,321
70,354
68,263
1,138,853
1,062,966
988,802
916,481
846,127
14.84
14.16
13.49
12.84
12.20
75
76
77
78
79
.03388
.03704
.04073
,04515
.05033
67,186
64,910
62,506
59,960
57,253
2,276
2,404
2,546
2,707
2,881
66,048
63,707
61,233
58,607
55,812
777,864
711,816
648,109
586,876
528,269
11.58
10.97
10.37
9.79
9.23
(6)
(7)
40
Table 2.10.
Exact age
in years
Mortality
(continued)
Probability
of dying
between
exact
age x and Number of
exact
survivors at
exact age x
age x+1
Number
of deaths
between
exact
age x and
exact
age x+1
Number of
years lived
between
exact age x
and exact
age x+1
Expectation
Total
of life
number
(average
number of
of years
lived
years
after exact
remaining)
age x
at exact age x
<7,
(1]
(2)
13]
(4)
(5)
(6)
80
81
82
83
84
.05622
.06269
.06973
.07722
.08519
54,372
51,315
48,098
44,744
41,289
3,057
3,217
3,354
3,455
3,517
52,844
49,706
46,422
43,106
39,531
472,457
419,613
369,907
323,485
280,469
8.69
8.18
7.69
7.23
6.79
85
86
87
88
89
.09409
.10405
.11420
.12427
.13471
37,772
34,218
30,657
27,156
23,782
3,554
3,561
3,501
3,374
3,204
35,995
32,437
28,907
25,469
22,180
240,938
204,943
172,506
143,599
118,130
6.38
5.99
5.63
5.29
4.97
90
91
92
93
94
.14661
.16024
.17460
.18904
.20348
20,578
17,561
14,747
12,172
9,871
3,017
2,814
2,575
2,301
2,029
19,069
16,154
13,459
11,022
8,867
95,950
76,881
60,727
47,268
36,246
4.66
4.38
4.12
3.88
3.67
95
96
97
98
99
.21823
.23221
.24560
.25834
.27040
7,862
6,147
4,719
3,560
2,641
1,715
1,428
1,159
919
714
7,004
5,433
4,140
3,101
2,283
27,379
20,375
14,962
10,802
7,701
3.48
3.31
3.17
3.03
2.92
100
101
102
103
104
.28176
.29242
.30237
.31163
.32023
1,927
1,384
979
683
470
543
405
296
213
150
1,655
1,182
831
577
394
5,418
3,763
2,581
1,750
1,173
2.81
2.72
2.64
2.56
2.50
105
106
107
108
109
.32817
.33550
.34224
.34843
.35411
320
215
143
94
61
105
72
49
33
22
268
178
119
77
50
779
511
333
214
137
2.44
2.38
2.33
2.28
2.24
110+
1.00000
39
39
87
87
2.23
T,
(7)
Source: Modified from United States, National Center for Health Statistics (1985, table 3).
w h i c h the later c o l u m n s of the table refer. In the l i f e table

the w o r d "age" is used very precisely, a n d t h e p r e c i s i o n is
e m p h a s i z e d by t h e a d d i t i o n of the m o d i f i e r " e x a c t . " W h e n
w e say that a person is exact age 0, w e m e a n that he or she
Mortality
41
was just b o r n . W h e n he or she is exact age 5, he or she has

l i v e d e x a c t l y f i v e f u l l years. In contrast, w h e n w e say i n everyday c o n v e r s a t i o n that s o m e o n e " i s 5 years o l d , " w e m e a n
that the person is b e t w e e n exact age 5 and exact age 6that
is, the person c o m p l e t e d 5 years of l i f e o n h i s or her last b i r t h day.
T h e letter x is used to represent exact age. S o m e of the
life-table f u n c t i o n s refer to the exact a g e x and others refer to
the age i n t e r v a l b e t w e e n exact age x and exact age x + 1.
C O L U M N 2: PROBABILITY O F DYING B E T W E E N E X A C T A G E x
A N D E X A C T A G E X + 1 : fl.
T h e second c o l u m n of the l i f e table [q ] represents the probx
a b i l i t y of d y i n g between exact age x and exact age x + 1. T h i s

c o l u m n s u m m a r i z e s the life-table m o r t a l i t y rates, w h i c h are
p r o b a b i l i t i e s and thus d i f f e r e n t f r o m the age-specific death
rates discussed earlier i n t h i s chapter. T h e q f u n c t i o n is the
x
n u m e r i c a l answer to the q u e s t i o n , A m o n g persons w h o reach

exact age x , w h a t p r o p o r t i o n w i l l die before t h e i r next b i r t h daythat is, w i t h i n one year? T h e t ^ values u s u a l l y are somew h a t l o w e r than the age-specific death rates [MJ w e discussed
earlier, but the rates are c l o s e l y p a r a l l e l . T h e t e c h n i c a l quest i o n of h o w to derive a set of q
values f r o m a set of age-
s p e c i f i c death rates need not c o n c e r n us at t h i s p o i n t . F o r

9
now, just r e m e m b e r that the q values are a set of m o r t a l i t y

x
p r o b a b i l i t i e s f o r the cohort as i t begins each s u c c e s s i v e year

of l i f e .
C O L U M N 3: S U R V I V O R S A T E X A C T A G E x: t
T h e t h i r d c o l u m n of the l i f e table [l ] represents the n u m b e r

x
of people w h o have s u r v i v e d f r o m b i r t h to exact age x . T h e

i n i t i a l cohort, the radix, is 100,000 i n Table 2.10 [l = 100,000).
Q
9. Appendix 2 provides a brief introduction to the relationship between q values

t
and M values.
x
42
Mortality
In the f i r s t year of l i f e , the p r o b a b i l i t y of d y i n g is .01120 (the

v a l u e of q ). C o n s e q u e n t l y , 1,120 persons of the o r i g i n a l
0
100,000 die i n the first year of l i f e (see c o l u m n 4) a n d o n l y

98,880 persons reach exact age 1 (i = 98,880).
x
T h e n u m b e r of s u r v i v o r s to any age [l ) is equal to the

x
product of i , and the v a l u e of the m o r t a l i t y p r o b a b i l i t y f o r

x
the preceding age i n t e r v a l [q _ ], subtracted f r o m the n u m b e r

x
w h o s u r v i v e d to the b e g i n n i n g of the preceding age i n t e r v a l

|/
), In a f o r m u l a :
T o i l l u s t r a t e , w e c a l c u l a t e the v a l u e of f o r exact age 19 f o r

x
the l i f e table i n Table 2.10:
= 98,294 - [(.00055)(98,294)]
=
98,294-54
= 98,240.
If y o u f i n d these f o r m u l a s c o n f u s i n g at first, y o u s h o u l d bear
i n m i n d that they are s i m p l y algebraic statements of the fact
that the n u m b e r of s u r v i v o r s at any exact age consists of those
a l i v e one year earlier m i n u s those w h o d i e d d u r i n g the interv e n i n g year.
T h e m e a n i n g of the l c o l u m n m a y be clearer i f w e rex
fer to its possible use by an i n s u r a n c e c o m p a n y . In T a b l e 2.10,

note that 96,531 people reach age 4 0 and that 157 d i e d u r i n g
t h e i r 4 0 t h year. Suppose that the i n s u r a n c e c o m p a n y desires
to p r o v i d e $1,000 i n t e r m i n s u r a n c e f o r one year's coverage
f o r each of the 96,531 people r e a c h i n g age 40. S i n c e 157 of
the 96,531 are expected to d i e before t h e i r 41st b i r t h d a y ,
$157,000 m u s t be a v a i l a b l e to be p a i d out i n benefits. T h e
p r e m i u m for the i n s u r a n c e is to be p a i d b y 96,531 people;
therefore each m u s t pay:
$157,000
p l u s any charges f o r a d m i n i s t r a t i o n or p r o f i t s f o r the c o m pany.
Mortality
43
C O L U M N 4: NUMBER OF DEATHS BETWEEN E X A C T A G E x

A N D E X A C T A G E x + 1: d
T h e f o u r t h c o l u m n represents the n u m b e r of deaths to the

cohort between exact age x and exact age x + 1. S y m b o l i z e d
as d , it is e q u a l to the n u m b e r s u r v i v i n g to exact age x [t )
x
m u l t i p l i e d by the p r o b a b i l i t y of d y i n g b e t w e e n exact age x

and exact age x + 1:
dMtM
T h e n u m b e r of cohort deaths \d ) is also e q u a l to the differx
ence between the n u m b e r s u r v i v i n g to exact age x a n d the

n u m b e r s u r v i v i n g to exact age x + 1 that is:
;
In T a b l e 2.10, the n u m b e r of deaths i n the f i r s t year of

l i f e is 1,120, w h i c h is the product of 100,000 and .01120. T h e
n u m b e r of deaths at age 84 is 3,517, w h i c h is the p r o d u c t of
41,289 and .08519.
COLUMN
5: Y E A R S LIVED B E T W E E N E X A C T A G E x A N D
EXACT AGE x + 1: L
T h e f i f t h c o l u m n of the l i f e table \L ) represents the n u m b e r

x
of person-years l i v e d by the cohort d u r i n g an age i n t e r v a l .

A l t h o u g h a precise d e t e r m i n a t i o n of L values is n o t u s u a l l y
x
possible, w e can a p p r o x i m a t e the values b y a s s u m i n g that

deaths are evenly distributed throughout the i n t e r v a l between
exact age x and exact age x + 1 except d u r i n g the f i r s t f e w
years of l i f e . M a k i n g this a s s u m p t i o n , w e can e s t i m a t e the
value of L by averaging the n u m b e r of s u r v i v o r s at the bex
g i n n i n g of the age i n t e r v a l [i ] and the n u m b e r of s u r v i v o r s at

x
the end of the i n t e r v a l ( ^ , ) . In other words, i t is u s u a l l y ass u m e d that:

*-
'
T h i s a p p r o x i m a t i o n uses the same logic w e used earlier w h e n

w e e x p l a i n e d h o w the m i d y e a r p o p u l a t i o n is u s e d as a n
44
Mortality
approximation of the number of person-years lived for calculating death rates.

For the first few years of life, the average of l and i
x
x t l
is
not a reasonable approximation of L because deaths are not

x
evenly distributed throughout the year. Instead, they are concentrated at the earlier part of the year, as documented i n
our earlier discussion of the infant mortality rate. For this
reason, values of L for the first few years should be closer to
x
x t l
than to l . In the absence of good data for estimating the

x
relative weighting of l and , it is often assumed that:

x
xtl
L =.3e +J
0
and
L =Ae +.6 .
l
For L and for ages greater than 2, the .5{t + ) approxima2
xtl
tion is used. The formulas above are approximations based

on empirical observations. When data are available on the
mortality of children by the number of months or days since
birth, more refined estimates of L and L, are possible. We do
0
not describe them here, but more sophisticated techniques

are often used for calculating the L values for the first few
x
years of l i f e .
10
It may help you to look at these observations another

way. A l l the persons who survive the year (that is, live from
to t
x
x t l
) live for one year. Therefore the m i n i m u m number
of years lived is equal to t
x t V
For example, all of the 96,374
people who live from age 40 to age 41 i n Table 2.10 contribute one year of life,- hence we begin with a m i n i m u m of 96,374
41-year-olds. In addition, the persons who die during that
year (157 persons during their 40th year) live for some part of
the year. If all of them were to die one second after their 40th
birthday, then we could ignore the addition. O n the other
hand, if all the deaths occurred one second before the 41st
birthday, we could assume that all the decedents lived a f u l l
year. Our assumption is that deaths are likely to be more or
10. More sophisticated techniques were, in fact, used to construct the life table
shown in Table 2.10, which is why the values given f o r L and I, in Table 2.10 are
0
slightly different from what the above formulas would give. More sophisticated
methods also modify the .5 assumption at other ages.
Mortality
45
less evenly spaced throughout the year. If that is so, each

decedent w i l l have lived for an average of one-half year; hence
we add one-half of the deaths i n the 40th year (78) to the
total number alive at age 41 to obtain the total number of
years lived between birthdays 40 and 41. The resulting number is 96,452.
The logic of the life table permits us to make any other
reasonable assumptions about the distribution of deaths during the year. Since we know that most infant deaths occur
early i n the first year of life, data on that first year are used to
obtain an L figure that assumes much less than half a year
x
of life for the infant decedents. When we approach the oldest

ages of the age distribution, there may be similar effects.
C O L U M N 6 : T O T A L Y E A R S L I V E D A F T E R E X A C T A G E x:
The sixth column of the life table gives the number of person-years lived after exact age x. We have already considered
the number of years lived during the 40th year of life using
the L column, which gives these figures for each year of life.
x
The T figure at age 0 is the sum of all L entriesthat is,

x
how many years w i l l be lived i n the first, second, third, etc.,

years of life when all are added together. The T figure for
x
any other age (e.g., exact age 40) is the sum of the years lived
for that age (L ) and all later ages by those survivors still
40
alive at the beginning of the age i n question.

The entries i n the sixth column thus show the number
of person-years that the cohort w i l l live after reaching exact
age x. It is the sum of the values of L for exact age x and all
x
ages greater than x that are presented i n the life table. In a

formula:
i=x
where L.
entry i in the L column, and

x
means "take the sum of the L column starting

x
with entry x and add entries x + 1, x + 2, etc.,

until you have added the last entry ()."
46
Mortality
C O L U M N 7:
E X P E C T A T I O N O F LIFE, OR A V E R A G E N U M B E R O F
Y E A R S L I V E D A F T E R E X A C T A G E x:
e.
The seventh and last column in the life table is the one most
commonly used. It answers the question, If all the persons
reaching any exact age could share equally the total number
of years that all will live from that age onward, how many
years would each live on the average? After having calculated T (the total number of person-years lived after exact
x
age x), and (the number of persons who survived to attain

x
age x), it is easy to determine how long the average person in

the life table lives after exact age x. We simply divide the
entries in the T column by the entries i n the column:
x
In the life table for U.S. females for 1979-81 (Table 2.10),
females of exact age 35 had an expectation of living 44.53
more years on the average. That is, their expected time of
death, on the average, was at exact age 79.53. Females of exact age 0, on the other hand, had an expectation of life of
77.62 years. Expressed informally, this means that women
in the hypothetical cohort who survive the hazards of the
first 35 years exhibit an increase in the average age to which
they w i l l live over the age expected at their birth. The increase is, however, modest: 1.91 years (79.53 - 77.62).
THE
A B R I D G E D LIFE T A B L E
The life table we have just described is known as a
complete
life table because it presents the life-table functions for single

years of age. There are also life tables that present the functions for groups of ages. They may refer, for example, to the
probability of dying between exact age 5 and exact age 10 and
present all the values i n the table only for intervals of 5 years.
In these abridged life tables, the first year of life and the ages
from 1 to 4 are usually presented separately. Later ages are
usually presented in 5- or 10-year intervals. Although the
Mortality
47
calculation of the abridged table is slightly different from that

of the complete table," the interpretation of the values of
the life-table functions is the same as for the complete table.
Only the time interval must be adjusted i n discussions of
values taken from an abridged table. A small number placed
below and to the left of the letter for the life-table function
(e.g., I , d , , or qJ indicates the length of the interval. A n
5
1 0
abridged life table for U.S. females in 1988 is presented i n

Table 2.11.
THIRD SET OF
1. Life expectancy at birth for females i n the United States i n
MULTIPLE-CHOICE
1988
QUESTIONS
( a
was:
about 35 years.
(b) about 55 years.

(c)
about 78 years.
(d) about 100 years.

What is the life expectancy at birth for females i n your own
country?
What is the life expectancy at birth for males i n your own
country?
2. The difference between a generation life table and a period
life table is that:
(a)
the radix is different.
(b) one refers to a true birth cohort and the other does not.
(c)
one uses a different method for calculating q than the

0
other.
(d)
none of the above.
3. Life-table mortality rates (q ) are usually:

x
(a)
about the same as age-specific death rates [M ).

x
(b) higher than age-specific death rates (MJ.

(c)
exactly the same values as age-specific death rates ( M J .
(d) lower than age-specific death rates (M ).

x
4. If country A has a higher life expectancy and also a higher

crude death rate than country B, it is likely that:
11. For example, for calculating ^L^, the formula is (5/2) \ t +
M
? ).
liS
Table 2.11. Abridged life table for females: United
Exact age
in years
X
Size of
interval
in year
n
States, 1988
Probability
of dying
between
exact
age x and
exact
age x+1
Number of
survivors at
exact age x
Number
of deaths
between
exact
age x and
exact
age x+1
(4)
(5)
d.
Number of
years lived
between
exact age x
and exact
age x+1
Total
number
of years
lived
after exact
age x
Expectation
of life
(average
number of
years
remaining)
at exact age x
(6)
T,
(7)
(8)
(1)
(2)
1,
(3)
0
1
5
10
1
4
5
5
.0089
.0018
.0010
.0010
100,000
99,110
98,935
98,834
890
175
101
100
99,243
396,021
494,400
493,954
7,831,495
7,732,252
7,336,231
6,841,831
78.3
78.0
74.2
69.2
15
20
25
30
5
5
5
5
.0024
.0028
.0033
.0041
98,734
98,494
98,222
97,901
240
272
321
405
493,108
491,802
490,324
488,539
6,347,877
5,854,769
5,362,967
4,872,643
64.3
59.4
54.6
49.8
35
40
45
50
5
5
5
5
.0058
.0084
.0135
.0219
97,496
96,929
96,112
94,819
567
817
1,293
2,077
486,163
482,754
477,562
469,225
4,384,104
3,897,941
3,415,187
2,937,625
45.0
40.2
35.5
31.0
55
60
65
70
5
5
5
5
.0347
.0537
.0793
.1210
92,742
89,525
84,715
77,999
3,217
4,810
6,716
9,435
456,141
436,300
407,664
367,619
2,468,400
2,012,259
1,575,959
1,168,295
26.6
22.5
18.8
15.0
75
80
5
5
.1843
.2981
68,564
55,924
12,640
16,671
312,711
239,106
800,676
487,965
11.7
8.7
1.0000
39,253
39,253
248,859
248,859
6.3
85
Source: United States, National Center for Health Statistics (1991, table 6-1).
Mortality
(a)
49
A's population is younger than that of B.
(b) A's population is older than that of B.

(c)
A's population has a higher infant mortality rate.
(d) none of the above is likely.
S E C O N D SET O F
Determine whether each of the following statements is true
TRUE/FALSE
or false:
QUESTIONS
Survival ratios from age 0 to age 1 are higher than other

one-year survival ratios.
2. Life expectancy in the United States is greater for males

than for females.
3. A period life table is a hypothetical model because
mortality rates actually change from one time period
to the next.
4. In a country where mortality rates have remained relatively constant for many years, a generation life table
and a period life table would be almost identical.
APPLICATIONS
The life-table functions provide useful tools for analyzing the
O F T H E LIFE T A B L E
effects of mortality alone because migration is explicitly ex-
TO MORTALITY
eluded and fertility is held constant. The uses of the life table
ANALYSIS
are many and varied, but we concentrate on only three here:

(1) uses of the stationary population concept, (2) survival ratios, and (3) comparisons of life expectancy i n two or more
populations at various ages.
THE STATIONARY POPULATION
CONCEPT
The numbers i n the L column are similar to the midyear

x
population in each age group for a hypothetical Or model population that demographers call the stationary
population.
The
nature of this model population may be understood as resulting from the following process. Suppose that 100,000 persons are born each year and are subject to the mortality rates
shown in Table 2.10. After 40 years the population would
consist of all the age groups shown i n the L column up
x
50
Mortality
through age 40. The persons shown at age 2 would be the

survivors of the 100,000 babies born two years before, the
persons in the 40th year would be the survivors of the 100,000
babies born 40 years previously, etc. After about 110 years,
the whole population structure shown in theL column would
x
have been created. From that time onad infinitumthe

100,000 entering the population at birth would be exactly
balanced by the 100,000 dying at all ages: this is the characteristic that leads to the label "stationary." The size of this
total population would be T
and the L column would give

x
the age distribution of the stationary population.

The stationary population has many of the characteristics of a real population. For example, it has a crude birth
rate, called the life-table
12
birth rate, which is defined as fol-
lows:
b = k^where
= the radix, usually 100,000,
= the first entry of the T column, and
Jc = 1,000.
Note again that T is the total size of the stationary popula0
tion, since it is the sum of all the values in the L column.

x
The stationary population also has a crude death rate,

called the life-table
death rate. The life-table death rate is
equal to:
d = k^T
or, alternatively, to the reciprocal of e multiplied by a con0
stant:
The life-table death rate is the same as the life-table birth

rate, of course, since everyone i n the hypothetical cohort dies
at some age. This hypothetical population is "stationary"
because the number of births and the number of deaths are
12. Fertility measures are discussed in Chapter 3. The crude birth rate is the same
as the crude death rate except that the numerator for the crude birth rate is the
number of live births in a given year.
Mortality
51
equal and therefore the population is neither growing nor

declining in size.
Earlier, we described the life table as the life history of
a cohort of persons born (i.e., reaching exact age zero) during
a single year who move through a series of mortality rates
until all of them have died. Alternatively, we can think of
the life table as describing what happens each year i n a hypothetical stationary population. In talking about a stationary population, demographers look at it the following way:
The
stationary population is a model without immigration or
emigration in which the same age-specific probabilities of death

apply continuously and in which there are the same number of
births and deaths each year (Greville 1946, 21).
Another characteristic of a stationary population is that the

number of persons living in each age group never changes.
The figures in the L column, as we have said, specify the age
x
composition of the stationary population, and this age composition never changes.
The stationary population concept has limited descriptive value because the life-table model is very different from
what happens i n a real population. It is useful for analytic
purposes, however, because it summarizes what the age structure of a population would be if it were subject to the fixed
mortality and birth conditions i n the life table. A comparison of the age composition of females in the United States in
1988 w i t h that of the female stationary population for the
same time period shows, for example, that the stationary
population is older than the actual population (Table 2.12).
This reflects two facts: (1) mortality conditions for American females have improved over time,
13
and (2) crude birth
rates have actually been higher than crude death rates, resulting, i n the absence of migration, i n a growing population
13. Declining mortality does not automatically make a population older or younger;
rather, the effect depends on the age pattern of the mortality changes. Historically,
a decline in mortality has usually been especially important at the youngest ages,
resulting in a younger population by reinforcing the effect of high fertility. Because
infant and childhood mortality is now low in most countries, future mortality
declines are likely to be concentrated at the older ages and will result in an older
population (all other factors being equal).
52
Mortality
and a young age distribution. A similar result would obtain

even if only the latter condition were truethat is, birth rates
had been consistently higher than death rates.
Perhaps the most frequent use of the stationary population concept is for comparing death rates i n a stationary
population with those of an actual population. Table 2.13
presents life-table death rates for white females i n the United
States for various periods between 1900 and 1988 and cornTable 2.12. Age composition of the stationary population and the actual population for U.S.
white females: 1988
Difference
Composition of
actual population
Compositon of
stationary population
Ages
A
(i)
Estimated
between
actual and
stationary
populations'
actual
Percentage
percentage
Percentage
population
distribution
distributions
distribution
on 1 July 1988
of actual
(col. 4 minus
(in thousands)
population
col. 2)
(3)
(4)
(5)
of
A
(2)
99,363
1.3
1,517
1.4
+0.1
1-4
396,632
5.0
5,732
5.4
+0.4
5-9
495,226
6.3
7,063
6.7
+0.4
10-14
494,822
6.3
6,489
6.1
-0.2
15-19
494,000
6.3
7,216
6.8
+0.5
20-24
492,755
6.2
7,866
7.4
+ 1.2
25-29
491,466
6.2
9,067
8.6
+2.4
30-34
489,978
6.2
9,074
8.6
+2.4
35-39
488,011
6.2
8,112
7.7
+ 1.5
40-44
485,110
6.1
7,025
6.6
+0.5
45-49
480,509
6.1
5,697
5.6
-0.5
50-54
472,928
6.0
4,903
4.6
-1.4
55-59
60-64
460,750
5.8
4.7
-1.1
441,931
5.6
4,933
5,132
4.8
-0.8
65-69
414,254
5.2
4,829
4.6
-0.6
70-74
374,756
4.7
4,038
3.8
-0.9
75-79
319,782
4.1
3,227
3.0
-1.1
80-84
245,202
3.1
2,162
2.0
-1.1
85+
253,240
3.2
1,940
1.8
-1.4
7,890,715
99.9
106,023
100.2
0.3
<1
A l l ages
Sources: United States, Bureau of the Census (1990, table 1, p. 411; United States, National Center for Health
Statistics (1991, table 6-1).
Note: Totals may not equal sums because of rounding.
Mortality
53
Table 2.13. Crude death rates and life table death rates for U.S. white
females: selected years, 1900-88
Difference
(life table
Crude death
Life table
rate minus
Years
rate
death rate
crude rate)
1900-02
15.4
19.6
+4.2
1909-11
13.2
18.6
+5.4
1919-21
11.5
17.1
+5.6
1929-31
9.9
16.0
+6.0
1939-41
9.1
14.9
+5.8
1949-51
8.0
13.9
+5.9
1959-61
7.9
13.5
+5.6
1969-71
8.1
13.2
+5.1
1979-81
7.9
12.8
+4.9
1988
8.6
12.7
+4.1
Source: United States, National Center for Health Statistics (1991, tables 1-2, 6-4).
parable figures for the crude death rates. The life-table death
rates are consistently higher than the crude death rates because the age composition of the actual population has been
consistently younger than the age composition of the stationary population.
SURVIVAL
RATIOS
The life table is particularly valuable for making population

projections or for making estimates of the population by age
between census years. If we assume that the mortality conditions of a particular life table w i l l continue in the future,
we can determine what proportion of people in a given age
group w i l l survive from that particular age group to another.
For most age groups i n low-mortality societies such as Japan
or the United States, this is a safe assumption. Death rates at
most ages are so low and so stable that changes are not likely
to be great. Even a considerable percentage change in death
rates that are very low w i l l make little difference in survival
ratios. That is why population projections for countries
like Japan or the United States are not likely to be seriously
in error as a result of assumptions about future mortality
54
Mortality
rates. For example, since 93 percent of the women in the

14
1960 life table were surviving to age 45, projections for women
less than 45 years of age could not be much affected by further reductions in mortality.
The L column specifies the midyear population of the
x
stationary population i n the age interval x to x + 1. If we

want to determine the proportion of persons surviving from
age group x to the later age group x + n, we simply determine:
forward survival ratio from age* to age* + n -
^ -.
tn
On the other hand, if we want to know how many persons

would have been alive n years in the past, we can determine:
reverse survival ratio from age x + n to agex =
L
L x+n
To illustrate forward survival simply, suppose the life

table of U.S. females (Table 2.10) is the latest life table available. Government officials want to know how many females
w i l l be age 6 (i.e., between exact age 6 and exact age 7) i n
1996 if roughly 2 million females were of age 0 (not having
reached their first birthday) i n 1990. This information is
needed, let us say, to determine how many girls w i l l enter
the first grade of elementary school in 1996. Using Table 2.10,
we calculate:
^ = ^ 6 2 3 = 0.9953.
La 99,085
Using the life table, then, we can estimate that 99.53 percent of the girls at age 0 in 1990 w i l l survive to age 6 in 1996.
Multiplying 0.9953 by the number of girls at age 0 i n 1990
(2,000,000) gives us the projected number of females who
w i l l be age 6 in 1996, barring immigration and emigration of
young children: 1,990,600.
We introduced another type of survival ratio earlier, one
14. Experience has shown that such projections may nevertheless be far off the
mark, a result of faulty assumptions about
fertility.
Mortality
55
based on the column of the life table. These survival ratios

x
vary dramatically from country to country and have improved

rapidly in the less developed countries. Compare, for example,
the survival ratios for males i n Sri Lanka (formerly Ceylon)
in 1920, 1946, 1954, 1967, and 1981 with those for white
males i n the United States from 1900 to 1988 (Table 2.14).
Whereas only 67 percent of the Sri Lankans born i n 1920
would have survived to exact age 5 according to that life table,
the figure was almost 96 percent by 1981. The recent ratios
for the United States were even more favorable to survival.
Since these survival ratios depend only on mortality (and
exclude the effects of migration), they show that the mortality conditions in the two countries during this century were
strikingly different and that the Sri Lankan survival ratios
improved significantly i n the years between 1920 and 1981,
as did those of the United States between 1900-02 and 1988.
Table 2.14. Survivors to exact age x {(J of 100,000 male infants ( ):

0
Sri Lanka, 1920-81, and United States, 1900-02 to 1988

Exact age
Country
0
20
50
1920
100,000
67,167
56,681
34,458
19,174
1946
100,000
75,448
70,089
51,963
33,245
1954
100,000
86,948
84,332
76,085
62,541
1967
100,000
92,472
90,584
81,651
66,679
1981
100,000
95,825
94,276
84,819
69,053
1900-02
100,000
80,548
75,984
56,736
38,736
1909-11
100,000
82,718
78,792
60,118
40,264
1919-21
100,000
88,505
84,440
67,553
49,218
1929-31
100,000
91,294
88,220
71,518
50,154
1939-41
100,000
93,624
91,617
78,254
55,776
1949-51
100,000
96,077
94,695
84,158
61,566
1959-61
100,000
96,643
95,491
86,199
64,177
1969-71
100,000
97,395
96,126
86,070
64,318
1979-81
100,000
98,333
97,316
89,007
70,646
1988
100,000
98,676
97,758
89,886
73,517
and year
65
Sri Lanka
United States
Sources: United Nations Statistical Office Demographic

Yearbook 1957 (1957, table
26|, Demographic Yearbook 1974 (1975, table 35), Demographic
Yearbook 1985 (1987,
table 36); United States, National Center for Health Statistics (1991, table 6-4).
Note: For the United States until 1919-21, the < figures include only the states that
voluntarily joined the federal death registration system.
r
56
Mortality
C O M P A R I S O N S O F LIFE E X P E C T A N C I E S IN
DIFFERENT
POPULATIONS
The e column of the life table is particularly useful. In comx
paring the mortality of two countries, we have seen that crude

death rates and even standardized death rates have some
weaknesses. Since the life expectancy figures i n the life table
are derived from a model that excludes migration and holds
fertility constant, the values of the e function are often used
x
to compare the mortality of different countries or the same

country at several points in time.
The values of c , life expectancy at birth, are used espe0
cially often. Although the most accurate comparison of mortality in two countries would involve a detailed analysis of
all the q values or all the e values, the life expectancy at
x
birth is a good summary measure. It has some hazards, because the value of e is disproportionately affected by the
0
infant mortality rate, but infant mortality rates are usually

highly correlated with death rates at other ages. Further, the
value of e has an immediately appealing interpretation: e
0
for a given year of birth measures how long members of cohort born that year can expect to live on the average if mortality conditions remain the same i n the future as they were
during the year of birth.
In the late 1970s and the 1980s, values of life expectancy at birth in countries for which data were available ranged
from 38.1 years to 75.9 years for males (Table 2.15). For females, who usually live longer than males, the values ranged
from 41.2 years to 81.8 years.
Other illustrations of the use of life expectancy figures
are shown in Tables 2.16 and 2.17. From these tables, we can
make the following statements:
1. Males, regardless of color, had lower life expectancies
than females at all ages in the United States i n 1988
(Table 2.16).
2. Within each sex, nonwhites had lower life expectancies than whites at all ages in the United States in 1988
(Table 2.16).
Mortality
57
3. Whether measured by the absolute number of years

gained or proportionate gain, life expectancy at birth
increased substantially between 1850 and 1988 for both
white males and white females. Life expectancy at age
40, however, increased only moderately. Life expectancy at age 70 increased even less (Table 2.17).
You may find other interesting comparisons.
Table 2.15. Examples of high and low values of life expectancy at birth |e ) for males and
0
females: recent years

Males
Females
Year or
Country
Year or
period
Country
period
High e s
High e s
Japan
1989
75.91
Japan
1989
81.77
Hong Kong
1989
74.25
Switzerland
1987-89
80.70
Sweden
1988
74.15
France
1988
80.46
Switzerland
1987-89
73.90
Netherlands
1988-89
80.23
Israel
1988
73.87
Hong Kong
1989
80.05
Malawi
1977
38.12
Malawi
1977
41.16
Sierra Leone
1985-90
39.40*
Afghanistan
1985-90
42.00*
Afghanistan
1985-90
41.00*
Sierra Leone
1985-90
42.60*
Guinea
1985-90
42.00*
Guinea
1985-90
43.00*
Ethiopia
1985-90
42.40*
Ethiopia
1985-90
45.60*
Low e s
Low e s

Yearbook
Note: Countries with populations under 1 million are excluded.
* Estimate prepared by the United Nations Population Division.
1990 (1992, tables 4 , 1 1 ) .
Table 2.16. Life expectancies at selected exact ages for males and females, by color: United
States, 1988
g
Life expectancy at exact age
e x
and color
10
20
30
40
White
72.3
Nonwhite
67.1
63.2
53.6
44.4
35.2
58.6
49.1
40.3
32.0
50
60
70
26.3
18.4
11.8
24.1
17.2
11.6
Males
Females
White
78.9
69.7
59.9
50.2
40.6
31.2
22.6
15.0
Nonwhite
75.1
66.4
56.6
47.1
37.9
29.2
21.3
14.5
Source: United States, National Center for Health Statistics (1991, table 6-1).
58
Mortality
Table 2.17. Life expectancies for white males and females at exact ages 0, 40, and 70: United
States, 1850-1988
White males, by age
Year
40
70
1850
38.3
27.9
1890
42.5
27.4
1900-02
48.2
1901-10
49.3
1919-21
White females, by age

0
40
70
10.2
40.5
29.8
11.3
9.4
44.5
28.8
10.2
27.7
9.0
51.1
29.2
9.6
27.6
8.9
52.5
29.3
9.5
56.3
29.9
9.5
58.5
30.9
9.9
1920-29
57.8
29.4
9.2
60.6
31.0
10.2
1930-39
60.6
29.6
9.3
64.5
32.2
10.2
1939-41
62.8
30.0
9.4
67.3
33.3
10.5
1949-51
66.3
31.2
10.1
72.0
35.6
11.7
1959-61
67.6
31.7
10.3
74.2
37.1
12.4
1969-71
67.9
31.9
10.4
75.5
38.1
13.4
1979-81
70.8
34.0
11.4
78.2
40.2
14.9
1988
72.3
35.2
11.8
78.9
40.6
15.0
Sources: 1850-1929: Dublin, Lotka, and Spiegelman (1949, table 12); 1939-61: Grove and Hetzel (1968, 308), 196971: United States, National Center for Health Statistics (1978b, table 5-11; 1979-81: United States, National Center
for Health Statistics (1985, tables 5, 6); 1988: United States, National Center for Health Statistics (1991, table 6-1).
Note: 1850, 1890: coverage is restricted to Massachusetts; 1900-29: coverage is restricted to death registration
states; 1929-51: coverage is restricted to continental United States.
FOURTHSETOF
MULTIPLE-CHOICE
QUESTIONS
1. The stationary population is a model that:

( ) excludes migration.
a
( b )
holds fertility constant.
(c)
has fixed mortality rates.
|d)
is not very good as a descriptive model and is mainly

useful for analytic purposes.
(e)
Only (a), (b), and (c) are true.
(f)
Answers (a), (b), (c), and (d) are all true.
2. If the death rate of a stationary life-table population is 10,

this implies a life expectancy of about:
(a)
10 years.
(b) 30 years.
(c)
50 years.
Mortality
59
(d) 65 years.
(e)
100 years.
3. In a country with a high life expectancy, the fact that the

actual death rate is lower than the death rate of the stationary population means that:
(a)
the actual population is growing through natural increase.
(b)
the country has a younger actual population than the

stationary population.
(c)
neither of the above is true.
(d) both (a) and (b) are true.

4. Survival ratios may be used for:
(a)
making projections of the future population.
(b) comparing the mortality of several countries or the same

country at different points in time.
(c)
estimating the effect of different levels of q on future

x
population sizes.
(d) only (a) and (b).
(e)
THIRD SET OF
TRUE/FALSE
QUESTIONS
Answers (a), (b), and (c) are all true.

or false:
i
The life-table death rate for females in the United States

is higher than the crude death rate.
2. In a life table the life-table death rate is twice the lifetable birth rate.
3. Life expectancy at age 70 has not increased very much
in the United States i n the past 138 years.
4. If you know the life expectancy at birth for a life table
prepared for the year of your birth, you know how long
you are going to live.
5. Standardized rates are almost always better measures
of mortality than crude rates.
6. q is usually larger than q
0
than q
l0
and q
l0
is usually larger
60
Mortality
ADDITIONAL
For further reading on the materials in this chapter, we rec-
READING
ommend George W. Barclay (1958), Techniques

tion analysis-, A . J. Jaffe (1951), Handbook
ods for demographers-,
of
popula-
of statistical
meth-
Warren S. Thompson and David T.
Lewis (1965), Population
problems,
5th ed. and L. I. Dublin,

;
A . J. Lotka, and M . Spiegelman (1949), Length of life, rev. ed.

For a complete discussion of standardization and the closely
related technique of decomposition, see Prithwis Das Gupta
(1993), Standardization
and decomposition
of rates: A user's
manual.
For data on mortality for many nations and for many
time periods, we found the following sources by Nathan
Keyfitz and Wilhelm Flieger of particular value: (1968), World
population:
An analysis of vital data-, (1990), World
tion growth and aging: Demographic
trends in the late
tieth century-, and (1971), Population:

demography.
populatwen-
Facts and methods
of
In addition, we recommend Samuel H . Preston,
Nathan Keyfitz, and Robert Schoen (1972), Causes of death:

Life tables for national populations-,
United Nations Statis-
tical Office, Department of Economic and Social Affairs,

United Nations
demographic
yearbook,
published annually
(various issues), and United Nations (1993), World
popula-
tion prospects: The 1992 revision.

Illustrations of the wide utility of the measures discussed in this chapter can be found by referring to studies of
mortality in such sources as Richard A. Easterlin (1980), Population and economic
change in developing
countries; Samuel
H . Preston (1978), The effects of infant and child

on fecundity-,
Preston (1976), Mortality
populations;
Preston and Michael R. Haines (1991), Fatal
years: Child mortality
patterns
mortality
in late nineteenth-century
and Tai-Hun K i m (1990), Mortality
in
national
America;
transition in Korea: 1960-
1980. We also recommend two United Nations publications:

(1986a), Consequences
of mortality
and (1986b), Determinants

entials in developing
trends and
of mortality
differentials;
change and
differ-
countries.
Complete bibliographic information about these works

is given in the list of references at the end of the
Guide.
Mortality
More advanced discussion of the material in the
61
Guide
can be found in such books as the following: Donald J. Bogue

et al., eds. (1993), Readings
in population
research
method-
ology, 8 vols.; Mortimer Spiegelman (1968), Introduction
to
demography,
to
rev. ed. Nathan Keyfitz (1968), Introduction

;
the mathematics
Population
of population-, Hugh H . Wolfenden (1954),
statistics
and their compilation,
rev. ed. Henry

;
S. Shryock, Jacob S. Siegel, and Associates (1971), The

ods and materials
(1972), Demographic
of demography,
meth-
2.vols.,- Roland Pressat
analysis: Methods, results,
applications.
Fertility, Natural Increase, and

Reproduction Rates
A t the beginning of the last chapter we noted the importance
of mortality as a determinant of population growth for most
of human history. In more recent history, fertility and fertility control have become dominant in population policy and
demographic interest. To illustrate the facts that prompted
the shift i n attention, in 1985-90 there were annually about
86 million more births than deaths in the world, and the
ratio of births to deaths was more than two to one (United
Nations Population Division 1992, Demographic
Yearbook
1990, table 1). The increases in population due to these "natural" processes of birth and death, known as natural
increase,.
led many concerned nations in the 1960s and 1970s to adopt

national programs for fertility controljust as in the past
they had emphasized death control through campaigns
against smallpox, the plague, malaria, tuberculosis, polio, and
other diseases.
In this chapter we examine the most common measures of fertility and natural increase. In studying these measures, you w i l l note that most of them are rates, and consequently we follow some of the same procedures we use i n
measuring mortality. For example, we talk about crude rates,
specific rates, and standardized rates.
Demographers distinguish between fertility and fecundity. Fertility
refers to actual reproductive performance,
whereas fecundity
refers to the physiological capacity of a
woman, man, or couple to reproduce (United Nations Population Branch 1958, 38; IUSSP 1982, 73 and 78). The definitions are reversed in the Romance languages, sometimes causing confusion at international conferences. In Spanish, for
64
Fertility, Natural Increase, and Reproduction Rates
example, the word for fertility is fecundidad

for fecundity is fertilidad.
and the word
(See Petersen and Petersen 1985
for a multilingual glossary of common demographic terms

in English, French, Spanish, Italian, German, Japanese, C h i nese, and Russian.)
The measurement of fertility poses special problems.
We discuss these first before describing particular rates.
Fertility measures always relate the number of live births to
SPECIAL
PROBLEMS IN
MEASURINC
FERTILITY
specific population base and time reference period. Unfortunately, it is difficult to establish accurate statistical records
on live births because many infants die i n the first few hours
or days of infancy. A definition that describes a live birth
accurately is complex and difficult to establish; and once one
is established, it is difficult to be certain that the definition
is actually used by local registration authorities or by respondents answering questions i n a sample survey. Nevertheless,
the following definition of a live birth has international approval:
A live birth is the complete expulsion or extraction from its
mother of a product of conception, irrespective of the duration
of pregnancy, which, after such separation, breathes or shows
any other evidence of life, such as beating of the heart, pulsation
of the umbilical cord, or definite movement of voluntary muscles,
whether or not the umbilical cord has been cut or the placenta
is attached; each product of such a birth is considered live born
(United Nations Statistical Office 1955, 6).
It is unlikely that this definition is observed everywhere

in all cases. When a child dies before the birth is registered,
neither the birth nor the death may be registered, only one or
the other may be registered, or the birth may be registered as
a stillbirth. The registration system is thus prone to error.
Similarly, when a survey respondent is asked to report in
retrospect on live births, such short-lived children are particularly apt to be omitted.
Fertility measurement also presents special problems
not encountered w i t h mortality measurement because a
65
woman can die only once but she may have no births or more
than one birth. This allows us to consider two approaches to
fertility measurement: the cumulative
fertility
approach and
the vital rates (or yearly-birth-rates) approach. In using the

cumulative fertility approach, we measure the average number of children ever born to women up to some specified age.
In using the vital rates approach, we measure the number of
live births i n a given year as related to the population exposed to the "risk" of giving birth in that year.
The vital rates in fertility measurement are similar to
the mortality rates discussed in the second chapter, but there
are important differences. The population exposed to the risk
of childbearing is not ordinarily decreased by having a birth.
Dying, on the other hand, completely removes a person from
the population exposed to the risk of dying. Moreover, plural
births (e.g., twins or triplets) are possible even though infrequent.
Fertility measurement is also complicated by the fact
that fertility involves two parents, whereas death involves
only one person. The fact that a couple is the "base" is problematic when we want to consider specific rates because we
have to decide whose characteristics to use, the father's or
the mother's.
There are also m i n i m u m and maximum ages at which
men and women are physiologically capable of reproduction.
Yet another problem is that not every woman is truly exposed to the risk of childbearing, for the reason that not every woman in the population is paired with a member of the
opposite sex. In addition, through widowhood, divorce, separation, and the like, individuals may enter or leave a couple
unit at various times i n their lives.
Because of these special problems in measuring fertility, no one measurement system comparable to the life table
has become dominant in fertility studies. Instead, a wide
variety of rates and ratios are currently used, each of which
has advantages and limitations in particular analytic situations.
66
THE CRUDE
The crude birth rate (CBR} is defined as the number of births
BIRTH RATE
i n a given year divided by the number of people in the population in the middle of that year. Again, as for the C D R , the
ideal denominator is the number of person-years lived, which
is often impossible to calculate for a real population. The rate
is usually expressed per 1,000 persons. In a formula, we have:
CBR = 1,000
number of births
midyear population
During the late 1980s the range of crude birth rates for major
countries of the world ranged from 10 to more than 56 births
per 1,000 per annum (Table 3.1). The highest recorded rates
were found in Africa and Asia; the lowest, in Europe and
Japan. Eighty-three percent of the more developed nations
had rates under 15 whereas 73 percent of the less developed
nations had rates of over 30 (Table 3.2). Although the crude
birth rate is not a refined measure of fertility, most other
fertility measures also show this pattern of higher rates in
the developing world.
THE CRUDE RATE
As one might imagine from the recurrent concern about the
OF N A T U R A L
"population explosion," typical values of the crude birth rate
INCREASE
hig
n e r
a n
typical values of the crude death rate. The
crude rate of natural increase measures this gap, as in the

following formula:
,
, _ J number of births - number of deaths
CRNI = 1,000
midyear population
/
n v
crude birth rate - crude death rate.

In the 1985-90 period the population of the world had a crude
rate of natural increase of around 17 per 1,000 (Table 3.3).
The nations with the highest rates (20 to 33) were those i n
the developing regions: almost all of Africa, most of Latin
America except for the temperate region, most of Asia ex-
Table 3.1. Highest and lowest crude birth rates, by region: recent
years
Year or
Region and country
Crude birth rates (per 1,000)
period
High
Malawi
1985-90
56.3*
Uganda
1985-90
52.2*
Mauritius
1985-90
18.6*
Tunisia
1989
25.2
Low
Africa
Central and North America

Nicaragua
1985-90
41.8*
Honduras
1985-90
39.8*
Canada
1988
14.5
United States
1990
16.7t
South America
Bolivia
1985-90
42.8*
Paraguay
1985-90
34.8*
Uruguay
1989
18.0
Argentina
1988
20.7

Yemen
Afghanistan
1990
1985-90
51.2T
49.3*
Japan
1990
9.9t
Hong Kong
1990
11.7t

Albania
1989
24.6
Ireland
1990
15.lt
Italy
1989
9.7
Greece
1989
10.1
Former USSR
1989
Tajikistan
17.7
38
Ukraine
13
Oceania'
Papua New Guinea
1985-90
Australia
1989
14.9
New Zealand
1989
17.5
34.2*

Yearbook 1990 (1992, table 9|.
' Estimate prepared by the United Nations Population Division,
t Provisional figure.
a. Only three countries have populations of more than 1 million.
b. Data from Population Reference Bureau, 1992 World Population Data Sheet |1992|.
Note: Countries with populations of less than one million are excluded.
Table 3.2. Distribution of countries by level of crude birth rate: 1980s

Crude
Asia
birth rate
(per 1,000
Less de-
More de-
(excluding-
World
veloped
veloped
North
Latin
former
population)
total
regions
regions
Africa
America
America
USSR)
Europe
USSR
Oceania
All countries
129
100
29
43
23
34
23
10.0-14.9
24
22
20
15.0-19.9
20.0-24.9
12
11
Less than 10.0
Former
25.0-29.9
30.0^4.9
13
13
35.0^9.9
12
12
40.0-44.9
16
16
45.0^19.9
24
24
21
50.0-54.9
55+
0
Yearbook
0
1990(1992,
table 4).
Notes: Countries with populations under 1 million are excluded. More developed regions, as defined by the United Nations Statistical Office. World Urbanization
Prospects, 1990
(1990) include North America, Japan, Europe, Australia, New Zealand, and the former USSR. Less developed regions include all regions of Africa, Latin America, and Asia except
fapan and all regions of Oceania except Australia and New Zealand.
69
cept East Asia, and all of Oceania except Australia and N e w

Zealand. Europe, the former USSR, North America, Australia, and N e w Zealand had the lowest rates (2 to 7). East Asia
and the Caribbean had intermediate rates. Of course, if continued, any positive rate of natural increase in the absence of
net out-migration would lead to very large populations over
time. Using the compound interest formula and compounding annually, even a yearly natural increase rate of only 5 per
1,000 would quadruple a population in less than 300 years.
A t the high natural increase rate of 30 per 1,000 (or 3 percent), which is found i n many developing countries, a popu-
Table 3.3. Average crude rates of natural increase, by region: 1985-90

Crude rate of
Region
natural increase
Africa, total
30
West Africa
31
East Africa
33
North Africa
27
Central Africa
29
24
Southern Africa
North America
Latin America, total
22
Caribbean
17
Central America
25
South America
20
Asia, total (excluding former USSR)
19
East Asia
13
South Asia
23
Southeast Asia
21
West Asia
27
Former USSR
Oceania, total
11
7
Australia and New Zealand

Melanesia
23
Micronesia
20
Polynesia
29
17
A l l regions
Source: United Nations Demographic
Yearbook 1990 (1992, table 1.)
Note: Many of these rates are estimates and vary in reliability.
70
lation doubles i n only 24 years, triples in 38 years, and quadruples in 47 years (Table 3.4).
The natural increase rates of recent decades are very
high compared with those of previous historical periods. Using the data for all regions shown i n Table 3.5, we have estimated the crude rates of natural increase for periods from
1650 to 1990 to be:
Annual crude rate
of natural increase
(per 1,000)
Years
1650-1750
3.7
1750-1850
4.7
1850-1900
5.4
1900-50
8.4
1950-60
18.3
1960-70
20.2
1970-80
18.5
1980-90
17.4
Table 3.4. Approximate number of years a population takes to

double, triple, and quadruple in size, given specified rates of growth
(based on the compound interest formula of P = P [l+r]")
n
Approximate number of years (n)

that a population takes to
Rate (%) of growth
Double
Triple
Quadruple
per annum (r)
in size
in size
in size
0.50
139
220
278
0.75
93
14
188
1.00
70
111
139
1.25
56
88
112
1.50
47
74
93
1.75
40
63
80
2.00
35
55
70
2.25
31
28
49
45
62
2.50
2.75
26
40
51
3.00
24
38
47
3.25
22
34
43
3.50
21
32
41
3.75
19
30
38
4.00
18
28
35
56
Source: Modified from Marty and Neebe (1966, 1-8).

Note: The percentage growth rate used here is equivalent to one-tenth the growth rate
expressed per 1,000. For example, a growth rate of 0.5 percent is the same as a growth
rate of 5 per 1,000.
Table 3.5. Estimates of midyear populations, by region: selected years, 1650-1990

Population
and region
1650
1750
1850
1900
1950
1960
1970
1980
1990
100
106
111
133
222
279
362
477
642
26
82
166
199
226
252
276
12
16
38
74
166
218
286
363
448
Asia
330
498
801
925
1,377
1,668
2,102
2,583
3,113
Europe*
100
167
284
430
572
639
704
750
787
Europe'
393
425
460
484
498
545
791
Millions of persons
Africa
North America
Central and South America'
Former USSR*
A l l regions
1,262
1,650
180
214
243
266
289
2,516
3,020
3,698
4,448
5,292
Percentage distribution
Africa
18.3
13.4
8.8
8.1
8.8
9.2
9.8
10.7
12.1
North America
0.2
0.3
2.1
5.0
6.6
6.6
6.1
5.7
5.2
Central and South America"
2.2
2.0
3.0
4.5
6.6
7.2
7.7
8.2
8.5
60.6
63.0
63.5
56.1
54.7
55.2
56.8
58.1
58.8
18.3
21.0
22.5
26.1
22.7
21.2
19.0
16.9
14.9
15.6
14.1
12.4
10.9
9.4
7.2
7.1
6.6
6.0
5.5
100.0
99.9
99.9
100.1
100.0
Asia
Europe
Europe'
Former USSR
A l l regions
100.0
100.0
100.0
99.8
Sources: 1650: Carr-Saunders [1936, 42); 1750-1900: Durand (1968, 109|; 1950-90: United Nations Statistical Office, Demographic
a Includes Caribbean beginning in 1950.
b. Includes former USSR, including Asian portions.
c. Excludes former USSR, including Asian portions.
d. Included in Europe totals through 1900
Note: Columns may not total exactly 100 0 percent because of rounding.
Yearbook
1990(1992, table 1, p. 141).
72
We calculated these rates by presuming natural increase

to be constant during each decade and by using the exponential growth formula:
where: P = population at time t,

t
= population at time 0,
i = the growth rate,

t = the number of years, and
e = base of natural logarithms (e = 2.71828282 . . . ) .
For example, the growth rate for 1650-1750 can be calculated by making P = 791, P = 545, and t = 100. Hence:
t
791
= < J
IOOr
545
Solving the equation yields a growth rate of approximately
3.7 per 1,000.
From these estimates it is clear that the rate of growth
has been much higher since 1950 than it ever was previously,
although the rate has declined somewhat i n the past two
decades. In the two decades between 1950 and 1970, death
rates declined to low or moderate levels for much of the
world's population, yet birth rates remained high. It is this
fact that led to concern about the "population explosion"
and to such dramatic (and admittedly unlikely) projections
as the following:
V.
Projection of the post-World War II rate of increase gives a

population of one person per square foot of the land surface of
the earth in less than 800 years. It gives a population of 50 billion
(the highest estimate of the population-carrying capacity of the
globe ever calculated by a responsible scholar) in less than 200
years (Hauser 1960, 7).
Because of projections like this, the crude rate of natural increase has been an important and recurrent measure i n
recent demographic literature. Even though natural increase
rates have declined somewhat i n the last two decades, the
rate of 17.4 from 1980 to 1990 would still lead to a doubling
of the world's population in slightly more than 40 years.
73
THE GENERAL
In the mortality chapter, we discussed the rationale for us-
FERTILITY
ing age-specific death rates or death rates specific for other
RATE
characteristics. Fertility is also highly variable across subgroups of a population, and it is common to calculate agespecific, age-marital-status-specific, and other specific fertility rates.
The frequency of childbirth varies significantly with
the age of the parents, and the age at which maximum fertility occurs may be different for males and females. Furthermore, fertility is higher among couples who have established
some type of regular cohabitation (such as legal marriage or
common-law marriage) than among persons not i n such
unions (single persons, for example). Conventionally, specific
fertility rates are calculated for female parents and not male
parents, and henceforth we w i l l discuss specific birth rates
for females only. Male parallels could be developed i n each
case.
It is rare for a child to be born to a woman less than 15
years old or more than 50 years old. For this reason, one way
to refine the measurement of fertility somewhat is by using
the midyear population of women i n the childbearing years
for the denominator of the rate instead of the total midyear
population of both sexes. The rate so constructed is called
the age-delimited or general fertility
rate (GFR). It is defined
as the number of births i n a given year divided by the midyear population of women in the age groups 15-44 or 15-49,
although the ages 10-49 are sometimes used. In a formula:
GFR = 1,000
= k
B
P
number of births in a given year

midyear population of women
of ages 15 - 44 or 15 - 49
or A:
B
35'15
The purpose of the G F R is to restrict the denominator

to potential mothers, but it is still not restrictive enough.
Rates within five-year age groups may be different for two
74
populations and yet the two populations may have the same
general fertility rate if the age composition of women in the
childbearing years differs for the two. In this sense the GFR
is subject to the same kind of crudeness as the crude birth
rate, although it is a distinct improvement in precision.
In the recent past, general fertility rates for various countries have been in the range of the low 40s to the high 100s
per thousand women of reproductive age. Data for recent years
from table 11 of the United Nations Demographic Yearbook
1990 show that the highest values of the GFR were 188.1 for
Guatemala and 163.9 for Egypt. The lowest values were 39.7
for Japan and 41.7 for Italy. As is true of the crude birth rate,
the highest rates were found in developing countries and the
lowest rates were usually found in Europe.
AGE-SPECIFIC
FERTILITY RATES
Within the age range of 15-49 years, there are marked differences in the fertility of women of various ages. For this reason it is customary to calculate fertility rates for each age or
age group, as in the following formula:
F = Age-specific fertility rate for age group*, x + n
- 1 000
'
^ births to women in age group x, x + n
midyear population of women in age group*, x + n
n u m
5 e r
- L- " '
B
P>
where B = births to women of the age group x, x + n,

Pj = midyear population of women in the group*, x + n, and
X
k = 1,000.
In most analyses, five-year age groups are used to calculate the age-specific rates. Typically, the age-specific rates
are low or moderate in the 15-19 age group, highest in the
20s, and then decline to moderate levels for women in their
30s. Rates after age 39 are usually low. Rates for the mid- to
late-1980s for the United States, Japan, Guatemala, and Costa
Rica are portrayed graphically in Figure 3.1 to illustrate the
75
typical, mountain-shaped patterns of age-specific fertility.

Described more formally, the typical distribution is truncated,
skewed to the left, and leptokurtic (i.e., having a high, narrow concentration about the mean) relative to a normal distribution.
Although the patterns of age-specific rates are reasonably similar for various populations, the absolute levels of
the age-specific rates vary considerably. Table 3.6, which presents recent data on the lowest and highest national age-specific rates by age group, shows that very high age-specific
rates have been recorded among selected groups of women.
An example often cited to illustrate very high fertility
is the schedule of age-specific rates for the ethnic Hutterites
of North America, an Anabaptist religious sect living in small
colonies in the United States and Canada. In the book Man's
Capacity to Reproduce: The Demography of a Unique Population, Eaton and Mayer (1954) report the age-specific fertil-
Figure 3.1. Age-specific fertility rates: Costa Rica, 1984, Guatemala,

1985, Japan, 1989, and United States, 1988
15-19
20-24
25-29
30-34
35^39
40-44
Age group
Source: United Nations Statistical Office, Demographic Yearbook 1990
11992, table 11).
76
ity rates for Hutterite women in the 1936-40 period. Table

3.7 compares the Hutterite rates with the rates for all U.S.
women in 1940. At all ages except for ages 15-19, the
Hutterite rates were dramatically higher than the rates for
all U.S. women. The reason the Hutterite rates were lower
at ages 15-19 is that the Hutterites married at later ages than
the average for U.S. women. These figures mean that during
the peak fertility years, roughly 46 percent of the Hutterite
women gave birth each year (46.2 percent for women of ages
30-34 and 46.6 percent for women of 25-29). Even as late as
ages 35-39, 43 percent gave birth each year.
The Hutterite rates and the ranges found in Table 3.6
are examples of the extremes in age-specific rates. Most of
the rates in any age group are much closer together. In recent
years 93 percent of the countries for which data were available had age-specific rates below 100 at ages 15-19 (Table
Table 3.6. Lowest and highest age-specific fertility rates per 1,000 women: 1980s
Lowest
Women's ages
Rate
Country, year
15-19"
3.5
Japan, 1989
20-24
47.0
Japan, 1989
25-29
30-34
35-39
40-44
84.7
30.9
9.8
1.4
Bulgaria, 1989
Bulgaria, 1989
Bulgaria, 1989
Former German
Democratic Rep., 1988
Japan, Hungary,
Norway, Bulgaria,
Denmark, 1989
45-49
0.1
O.lt
Highest
Rate
137.7*
125.5
307.4*
273.5
333.2
289.5
183.0
122.7*
81.5
212.2*
43.0
Country, year
Honduras, 1981
Guatemala, 1985
Honduras, 1981
Guatemala, 1985
Egypt, 1986
Egypt, 1986
Guatemala, 1985
Dominican Rep., 1980
Guatemala, 1985
Dominican Rep., 1980
Guatemala, 1985
Notes: Countries with populations of less than 1 million are excluded. Of the countries with populations greater
than 1 million, coverage is more complete for countries of the more developed world. For example, in Africa, data
are available for only Botswana, Egypt, Mauritius, and Tunisia, whereas virtually every European nation is included.
* Rate has been calculated from live births recorded in civil registers that are incomplete or of unknown completeness. Such figures are included in the "highest" column only when the rate is higher than any rate from a more
reliable source.
t Based on 30 or fewer live births.
a. Computed using all women under age 20.
b. Computed using all women of ages 45 and above.
77
Table 3.7. Age-specific fertility rates per 1,000 women: Hutterites

and all U.S. women, around 1940
Women's ages
Hutterite women,
U.S. women,
1936-^0
1940-
15-19
259
136
25-29
466
123
30-34
462
83
35-39
431
46
40^14
203
16
45-49
48
Sources: Hutterite women: Eaton and Mayer 11954, table 11); U.S. women: United
States, National Center for Health Statistics 11978a, table 1-6).
a. U.S. rates have been corrected for underregistration of births.
3.8). At ages 20-24, 80 percent of the age-specific rates were

in the range of 50-199. Rates in the remaining age groups
showed similar patterns of concentration in a narrower range
than the ranges given in Table 3.6.
Table 3.8. Distribution of major countries and territories by level of age-specific fertility
rates: recent years
Level of age-specific fertility rates (per 1,000 women)
Women's
0-49
50-99
15-19
37
20
20-24
18
17
14
25-29
37
ages
100-149 150-199 200-249 250-299
300+
All levels
61
61
61
61
Number of countries
30-34
33
15
35-39
38
13
61
40-44
54
61
45-49
61
61
15-19
61
33
101
20-24
30
28
23
98
25-29
10
61
15
101
30-34
54
25
101
35-39
62
21
10
100
40-44
89
10
101
45^19
100
100
Percentage distribution
Source: United Nations Statistical Office, Demographic Yearbook 1990(1992, table 11).
Notes: Percentages may not add exactly to 100 percent because of rounding. Countries with populations under 1
million are excluded. Of the countries with populations greater than 1 million, coverage is more complete for more
developed than for less developed countries. For example, in Africa, data are available for only Botswana, Egypt,
Mauritius, and TUnisia, whereas virtually every European nation is included. To achieve greater coverage, countries
with incomplete registration are included; therefore, actual age-specific rates are somewhat higher than those
summarized in this table.
78
FIRST SET O F
TRUE/FALSE
QUESTIONS

or false:
As a result of postwar progress, only about one-third of
the world's population lives in countries with high rates
of natural increase; the other two-thirds has attained
relatively low rates of natural increase resulting from
low birth rates and low death rates.
2. The majority of countries in the 1980s had crude birth
rates above 30 per thousand per annum.
3. The recent crude rates of natural increase for the population of the world were probably never attained in the
period between 1650 and 1950.
4. It is unlikely that a population would have a crude birth
rate of 40 and a crude death rate of 15 during the same
time period.
5. The lowest birth rates recorded in the late 1980s were
mostly for European nations.
FIRSTSETOF
MULTIPLE-CHOICE
QUESTIONS
1. In two hypothetical countries, A and B, the age-specific

fertility rates for females are as follows:
Ages
Country A
Country B
15-24
25-34
35-44
80
250
100
80
250
100
(a) Country A has a higher general fertility rate than country B.

(b) Country B has a higher general fertility rate than country A.
(c) Country A has the same general fertility rate as country B.
(d) Country A has the same crude birth rate as country B.
(e) The general fertility rate for country A may be the same,
higher, or lower than the general fertility rate for country B.
2. The crude birth rate in the United States is now approximately:
(a) 10 per thousand.
79
(b) 15 per thousand.

(c) 25 per thousand.
(d) 35 per thousand.
What is the crude birth rate in your own country?
3. Characterize as closely as possible the population of the
United States, Canada, and the former USSR.
(a) crude birth rate of 21-44, crude death rate of 20-30.
(b) crude birth rate of 15-20, crude death rate of 5-10.
(c) crude birth rate of 20-30, crude death rate of 15-25.
(d) crude birth rate of 10-25, crude death rate of 5-15.
(e) crude birth rate of 10-16, crude death rate of 5-30.
4. Characterize as closely as possible the populations of the
European nations. (Select from the same answer categories
as for question 3.)
5. A crude rate of natural increase of 30 per thousand leads to
a doubling of the population in approximately:
(a) 15 years.
(b) 25 years.
(c) 50 years.
(d) 75 years.
(e) 100 years.
6. Typically, age-specific fertility rates for women:
(a) are highest at ages 15-24 and lower thereafter.
(b) are highest at ages 20-29 and lower at ages 15-19 and
ages over 30.
(c) are highest at ages 25-34 and lowest at ages 15-24 and
ages over 35.
(d) are fairly constant throughout the childbearing years.
BIRTH RATES
SPECIFIC FOR
CHARACTERISTICS
OTHER THAN A G E
We often need to study birth rates specific for characteristics

other than age. Two important characteristics are marital
d live birth order.
s t a t u s
a n
All societies have forms of culturally sanctioned reproductive units resulting from a religious marriage, legal marriage, consensual union, common-law marriage, "living in"
80
arrangement, or other union. For convenience, we call all of

these institutionalized arrangements "marriage" in the
present discussion. Although marriage is a nearly universal
phenomenon, significant variation exists among norms about
the proper age to marry, about divorce, and about the remarriage of widows. The norms may change over time and the
possibility of adhering to them may be affected by the age
and sex composition of a population. For example, in populations where there is a shortage of eligible males or females,
persons of the opposite sex who wish to marry may find themselves caught in a marriage squeeze. A marriage squeeze is
usually a "by-product of the combination of different-sized
cohorts and the fact that women do not usually marry men
their own age" (Weeks 1992, 320). The squeeze refers to an
imbalance between the number of males and females in the
age groups that usually marry one another: either there are
too many males or too many females relative to the other
sex in the appropriate age groups. (For a more detailed discussion of this phenomenon, see Schoen 1983 and Weeks
1992, 320-21.)
1
Since the proportions married may vary, and since birth

rates generally are much higher for the married than for the
total population of women, demographers often construct
fertility rates specific for marital status as well as for age, so
that one has age-marital-status-specific fertility rates. Five
articles illustrate the use of fertility rates specific for marital
status and age. Freedman and Adlakha (1968), Hirschman
(1986), Ogawa and Retherford (1993), Peng (1993), and
Retherford and Rele (1989) examine respectively the factors
responsible for declining fertility rates in Hong Kong in the
1960s, Peninsular Malaysia from 1970 to 1980, Japan from
1950 to 1990, China from the mid-1960s to the early 1980s,
1. For an informative discussion of the complexity of marital unions in relation to

fertility in the U.S. context, we recommend a series of papers by Ronald Rindfuss
and colleagues (Rindfuss and Parnell 1989, Rindfuss and Jones 1991, and Pagnini
and Rindfuss 1993). Chapters 4-6 in a book by Stycos and Back (1964) published
three decades ago, are still recommended reading on this topic for a look at these
relationships in the Caribbean.
81
and the countries of South Asia from 1960-64 to 1980-84. To

quote Freedman and Adlakha (p. 181):
An important question about such declines in crude birth rates
is whether they result from real declines in the fertility of married
women or from changes in the number of women of childbearing
age or in the proportion who are married in the productive
childbearing years. Are married women having fewer children
or are there simply fewer married women in the important
childbearing years?
Clearly, we cannot review the detailed results of all five studies here. Instead, we present a few examples of the results
showing how changes in the marital status and age compositions of populations can affect fertility:
1. Changes in the marital-status composition of the
populations were among the important causes of the fertility declines in all of the countries studied during the time
periods mentioned above. For example, all of the fertility
decline in Japan between 1980 and 1990 and roughly half of
the decline between 1970 and 1980 were due to changes in
the marital-status composition of the population. Between
1960-64 and 1980-84, the percentages of the fertility declines
due to changes in the marital-status composition were: 51
percent in Bangladesh, 28 percent in India, and 39 percent in
Sri Lanka. In Pakistan from 1960-64 to 1980-84, fertility remained relatively constant, but it would have increased were
it not for changes in the marital-status compositon. In China,
changes in the marital-status composition accounted for
about one-fifth of the declines in rural fertility between the
mid-1960s and the early 1980s. In urban areas of China during the same time period, changes in the maritial-status composition accounted for as much as 40 percent of the decline
(in Shanghai) and as little as 6 percent (in Jiangsu and Hebei
provinces).
2. Changes in the age composition were important in
Hong Kong's fertility decline in the 1961-65 period but not
in 1965-66. In Peninsular Malaysia, changes in the age composition would have led to increased fertility among the
Malays between 1970 and 1980 were it not for the opposite
82
effect of changes in their marital-status composition. Hence,

the crude birth rate for Malays declined only about 1 percentage point in that decade.
Fertility rates specific for live birth order are also useful. The probability of having an additional child is affected
by how many children a woman has already borne. This is
true because she and her husband may use contraception after a certain number of births and because the physiological
capacity to bear children is affected by previous childbirths
as well as by age and other factors. One may calculate the
birth order-specific fertility rate as follows:
number of births of order i

Birth-order-specific
= 1,000
fertility rate
lidyear population of women
mi
of ages 15-44 or 15-49
ff
ff
or k
P'
35 P
15
l
30^15
where B' =births of order i,

30^15
k=
midyear population of women between

ages 15-44 or 15-49, and
1,000.
(Notice that the sum of the birth-order-specific fertility rates

is the GFR.) It is often useful to make the rates specific for
smaller age groups, and we may therefore wish to calculate
age-specific, birth-order-specific fertility rates.
To illustrate the use of rates specific for live birth order, consider the data for the United States summarized in
Table 3.9. The general fertility rate in the United States was
lower in 1988 (67.2) than in 1942 (91.5). During the intervening 46 years, the rate had both increased and decreased from
the 1942 level, reaching a high value of 122.9 in 1957, thereafter declining and then staying fairly steady until 1988. Table
3.9 also shows data for two of the intervening years, 1960
and 1975.
Table 3.9. Birth rates by live birth order and percentage change in rates: United States, selected years, 1942-88
Live births per 1,000 women
15-44 years old
Live birth order
First birth
Second birth
Third birth
Fourth birth
Fifth birth
Sixth and seventh births
Eighth and higher births
All births (GFR)
Percentage change"
1942
1960
1975
1988
1942-60
U)
12)
13)
14)
15)
37.5
22.9
11.9
6.6
4.1
4.6
3.9
91.5
31.1
29.2
22.8
14.6
8.3
7.6
4.3
118.0
28.1
20.9
9.4
3.9
1.7
1.3
0.7
66.0
27.6
22.0
10.9
4.1
1.5
0.9
0.3
67.2
-17.1
27.5
91.6
121.2
102.4
65.2
10.3
29.0
Source: United States, National Center for Health Statistics (1969, table l-8 1990, table 1-8).
a. Column 5 - ((column 2 - column l|/column 1]/ x 100;
Column 6 - ((column 3 - column 2)/column 2)/ x 100;
Column 7 - Ifcolumn 4 - column 3)/column 3)/ x 100,
Column 8 - |(column 3 - column 1 (/column 1 ]/ x 100;
Column 9 - [(column 4 - column l)/column 1]/ x 100.
;
1960-75
(6)
1975-88
1942-75
1942-88
17)
18)
19)
-9.6
-28.4
-58.8
-73.3
-79.5
-82.9
-83.7
-44.1
-1.8
5.3
16.0
5.1
-11.8
-30.8
-57.1
1.8
-25.1
-8.7
-21.0
-40.9
-58.5
-71.7
-82.1
-27.9
-26.4
-3.9
-8.4
-37.9
-63.4
-80.4
-92.3
-26.6
84
Rates for all birth orders but the first were higher in
1960 than in 1942. The 1960 birth rate for the first birth order was low because so many women had already had their
first children in the "baby boom" period of the 1950s. Most
of the difference in the general fertility rates of 1960 and 1942
(118.0 - 91.5 = 26.5) resulted from higher rates for second,
third, and fourth births (25.2 of the 26.5 difference). Much of
the 1960-75 fall (118.0 - 66.7 = 51.3) was due to declines in
the rates for the third and fourth birth orders (total decline of
24.1), but the rates fell for all orders during that period. The
GFRs for 1975 and 1988 were almost the same but rates for
second, third, and fourth births rose and for other birth orders declined between the two years.
;
Fertility rates specific for age, for marital status, or for

live birth order are only three examples of many specific rates
that may be useful in a particular fertility analysis. Demographers may also be interested in the variation in fertility
rates by mothers' parity (i.e., the number of children already
borne), educational attainment, income, size of place of residence, ethnic group, occupation, contraceptive use, and other
social and economic variables. The method of computing rates
specific to other characteristics is similar to that examined
for age, marital status, and live birth order; all you need are
the data.
STANDARDIZED
BIRTH RATES
Because we are interested in measuring fertility itself, we

often want to control for, or eliminate the effect of, other
variables. To do this we may look at a detailed schedule of
specific fertility rates (by age, marital status, or any other
characteristic) and compare two populations. Alternatively,
we may want a single measure that corrects for the effects of
the extraneous variables. One such measure would be a standardized fertility rate, corresponding to the standardized
mortality rates discussed in Chapter 2.
The most common standardized fertility measure in
use is the age-sex adjusted birth rate, which is the crude birth
85
rate standardized for age and sex composition. Because the

procedure used in calculating this rate is similar to that used
for standardizing the crude death rate, we do not repeat that
discussion here. The actual values of age-sex adjusted birth
rates, however, are interesting.
Table 3.10 presents the crude birth rates and standardized birth rates for 37 countries and dates, based on calculations by Keyfitz and Flieger (1990). The standard populations
used are those of the United States and Mexico in 1980. Although many statements can be made about the rates summarized there, we focus on two observations:
1. The rank order of countries from highest to lowest
birth rate is not affected much by the standardizationthat
is, the correlation between the two rank orderings is high
but a few countries do change in rank depending on the standard used. For example, Peninsular Malaysia has the third
highest unstandardized rate. Standardized on the 1980 U.S.
age distribution, it retains this ranking. Standardized on the
1980 Mexican age distribution, however, Peninsular Malaysia has a rank of 5.
2. Even though the rank order of countries is not critically affected by standardization, the amounts of the differences do change substantially. For example, the crude birth
rates are 26.40 for Mauritius in 1980 and 20.96 for China in
1981, a difference of 5.44. Standardized on the age distribution of Mexico in 1980, the rates are 21.88 for Mauritius and
19.48 for China, a difference of 2.40. Hence, a large proportion of the difference in crude rates between Mauritius and
China is due to differences in their age distributions.
Data for the United States between 1940 and 1988 provide an additional illustration of the use of age-sex adjusted
2
2. We can quantify how much of the difference in crude rates is due to differences
in the age distribution and how much is due to differences in the ASFRs using a
technique called decompositon.
For a discussion of decomposition and
standardization techniques, see Das Gupta (1993). For a description of how to
decompose the total fertility rate, discussed in the next section of this chapter,
see Retherford and Rele (1989, 744-45|.
Table 3.10. Crude birth rates and directly standardized birth rates: selected places and dates
Crude
hirth
Crude birth rates standardized using as standard:

United States, 1980
Mexico, 1980
Place, Year
rate
Rate
Rank
Rate
Rank
Guatemala, 1985
Mexico, 1983
Panama, 1980
Venezuela, 1985
Fiji, 1980
Mauritius, 1980
China, 1981
Former USSR, 1987
Chile, 1980
Poland, 1985
Ireland, 1986
Former Yugoslavia, 1985
Bulgaria, 1985
Hong Kong, 1980
New Zealand, 1985
Australia, 1985
Hungary, 1985
United States, 1985
German Dem. Rep., 1985
France, 1985
Spain, 1983
Japan, 1980
Sweden, 1985
Greece, 1985
Portugal, 1985
Canada, 1985
Norway, 1985
Finland, 1985
Singapore, 1985
Belgium, 1984
Austria, 1985
Netherlands, 1985
Italy, 1983
Denmark, 1985
Germany, Fed. Rep., 1985
41.04
34.96
31.34
28.65
29.01
29.67
26.40
20.96
19.92
21.13
18.21
17.35
15.85
13.28
16.98
15.80
15.67
12.23
15.75
13.68
13.93
13.15
12.71
13.51
11.79
11.74
12.85
14.82
12.29
12.81
16.61
11.73
11.57
12.29
10.59
10.51
9.61
48.05
37.46
31.34
30.50
29.59
28.89
24.31
22.68
21.83
20.86
20.18
19.64
17.81
17.70
17.16
16.50
16.34
16.20
15.86
15.64
15.57
15.21
14.92
14.87
14.54
14.53
14.46
14.23
14.23
13.74
13.49
13.25
12.77
12.72
12.60
12.29
10.87
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28.5
28.5
30
31
32
33
34
35
36
37
42.59
33.30
26.62
28.03
26.64
25.79
21.88
19.48
19.80
19.01
18.25
16.55
16.30
17.08
14.65
14.63
14.26
15.17
14.54
14.64
13.63
13.47
12.97
12.68
12.48
13.24
12.98
12.51
12.47
11.84
11.49
11.68
11.48
10.80
11.02
10.64
9.32
1
2
5
3
4
6
7
9
8
10
11
13
14
12
16
18
20
15
19
17
21
22
25
26
28
23
24
27
29
30
32
31
33
35
34
36
37
Source: Keyfitz and Flieger (1990, 294 ff ).
87
birth rates (Table 3.11). The crude birth rate for the United
States increased from 19.4 in 1940 to a peak of 25.0 in 1955,
declined to 14.6 in 1975, then rose to 15.9 in 1980 and again
in 1988. The highest crude birth rates were recorded in the
baby boom period of the 1950s, right after World War II. Although the crude rates in that period were high, they were
not nearly as high as they would have been if the age and sex
composition of the 1940s had still held in the 1950s. For example, if the age and sex composition of the 1955 population
had been the same as the 1940 composition, the birth rate
for 1955 would have been 30.4 instead of the observed value
of 25.0. In fact, the standardized birth rate for every year after 1940 except the early 1980s is higher than the crude rate
when the 1940 composition is used as the standard population. The age and sex structure of the United States since
1940 has generally been less favorable to high crude birth
rates than was the 1940 age and sex structure. The situation
in the 1980s was due to the baby boomers of the 1950s passing through the prime childbearing ages.
We can standardize more refined fertility measures
the general fertility rate, for example, or even age-specific
Table 3.11. Crude birth rates and directly standardized birth rates:
United States, selected years, 1940-88
1940
1945
1950
1955
1960
1965
1970
1975
1980
1985
1988
Crude
birth
rate
Age-sex adjusted birth rate

using 1940 U.S. age and
sex distribution
as standard population
19.4
20.4
24.1
25.0
23.7
19.4
18.4
14.6
15.9
15.8
15.9
19.4
20.9
26.3
30.4
31.2
24.8
21.3
15.3
15.8
15.8
16.5
Source: United States, National Center for Health Statistics (1990, tables 1-2, 1-3, pp.
1,4|.
Note: For years prior to 1960, data are adjusted for underregistration.
Table 3.12. General fertility rates for selected countries, age-standardized by the direct method: late 1980s (standard - Sweden, 1988)
Standard million
for females 15-49,
Sweden, 1988
Ages
15-19
20-24
25-29
30-34
35^9
40-44
45-49
All ages
(1)
135,064
149,706
136,531
138,687
145,961
163,270
130,781
Age-specific fertility rates, by country*

Egypt
(2)
Philippines
(3)
20.1
203.0
333.2
289.5
182.6
69.8
32.7
42.5
162.7
180.4
135.7
96.2
40.5
8.9
Expected number of births
Sweden
(4)
United
States
(5)
Japan
(6)
Egypt
(l)x|2)
11.4
90.7
146.6
100.9
36.8
6.0
0.3
54.8
111.2
113.2
73.7
27.9
4.8
0.2
3.5
47.0
144.8
90.9
19.5
2.4
0.1
1,000,000
Observed GFRs (per 1,000)
163.4
106.8
55.5
60.2
Standardized GFRs (per 1,000)

Sources: United Nations Statistical Office, Demographic Yearbook 1990(1992, tables 7, 11).
Note: Philippine data are from the civil register, which is of unknown completeness,
a. Data for all countries are for 1988 except data for Egypt, which are for 1986.
Philippines
United
States
Japan
U)x(3)
Sweden
(l)x(4)
U)x(5)
U)x(6)
2,715
30,390
45,492
40,150
26,652
11,396
4,277
5,740
24,357
24,630
18,820
14,041
6,612
1,164
1,540
13,578
20,015
13,944
5,371
980
39
7,402
16,647
15,455
10,221
4,072
784
26
473
7,036
19,770
12,607
2,486
392
13
161,072
95,364
55,467
54,607
42,777
161.1
95.4
55.5
54.6
42.7
39.7
89
fertility rates. To illustrate, we have age-standardized recent

general fertility rates for Sweden, Egypt, the Philippines, the
United States, and Japan, using three standard populations:
Sweden in 1988, India in 1990, and the Republic of Korea in
1990. Table 3.12 shows the calculations using the Swedish
population as standard. The comparable calculations for the
India standard and the Korea standard are not shown, but the
observed rates and the standardized values are summarized
in Table 3.13. The rank order of the general fertility rates is
similar for the actual rates and each set of standardized rates:
Egypt always has the highest rate and Japan the lowest, with
the Philippines second highest, and the United States and
Sweden very close. The amounts of the differences, however,
are affected by the standards. To take just one example, the
ratio of the actual GFR for Egypt to that of Japan is 4.1 (163.4/
39.7). Standardized on the age distribution of India, however,
the ratio falls to 3.5 (165.0/47.0). Notice that Japan's rate is
more affected by the standardization (it increases from 39.7
to 47.0 than is Egypt's rate (which increases only from 163.4
to 165.0). What this means is that a smaller proportion of
Japan's women were concentrated in the peak fertility years
than was true in Egypt or India. In other words, Japan's female age distribution was less favorable to high general fertility rates than India's or Egypt's.
Table 3.13. Observed general fertility rates and age-standardized general fertility rates with
Sweden (1988), India (1990), and the Republic of Korea (1990) as standard populations: Egypt,
Philippines, Sweden, United States, and Japan
Age-standardized GFR using as standard population:
Country
GFR
Sweden,
1988
India,
1990
Korea,
1990
Egypt
Philippines
Sweden
United States
Japan
163.4
106.8
55.5
60.2
39.7
161.1
95.4
55.5
54.6
42.7
165.0
102.3
60.9
64.0
47.0
175.5
105.7
64.6
64.6
51.1
Sources: GFRs: United Nations Statistical Office, Demographic Yearbook 1990(1992, table 11). Age-standardized
GFRs use standard age distribution of the countries listed and the ASFRs from Table 3.12 in this chapter.
90
S E C O N D SET O F
MULTIPLE-CHOICE
1. In two hypothetical countries, A and B, the age-specific

fertility rates per 1,000 are as follows:
QUESTIONS
Ages
Country A
Country B
15-24
25-34
35-44
50
100
60
50
100
60
In country A, 60 percent of the population is female and 30

percent of the females are between the ages of 15 and 44. In
country B, 50 percent of the population is female and 35 percent of the females are between the ages of 15 and 44. Assume that births occur only to women between the ages of
15 and 44.
(a) The crude birth rate is higher in country A than in country B.
(b) The crude birth rate is lower in country A than in country B.
(c) The crude birth rate is equal in the two countries.
(d) Any of the above may be true.
2. Using the data in question 1, it is possible to say with
certainty that:
(a) The general fertility rate is higher in country A than in
country B.
(b) The general fertility rate is higher in country B than in
country A.
(c) The general fertility rate is equal in the two countries.
3. Using the data in question 1, it is possible to say with
certainty that:
(a) The age-sex adjusted birth rate is higher in country A
than in country B.
(b) The age-sex adjusted birth rate is higher in country B
than in country A.
(c) The age-sex adjusted birth rates in the two countries
are equal.
91
4. As compared with less developed nations, the age structures of the more developed nations tend to be unusually
favorable to:
(a) high crude birth rates and high crude death rates.
(b) low crude birth rates and high crude death rates.
(c) low crude birth rates and low crude death rates.
(d) high crude birth rates and low crude death rates.
(e) none of the above.
S E C O N D SET OF
TRUE/FALSE
QUESTIONS

or false.
i . The standardization of crude birth rates makes relatively little difference in the rank order of countries for
values of the birth rate, but it does affect the sizes of
rates relative to one another.
2. It would be possible to construct an age-standardized
rate of natural increase.
3. Fertility rates specific for live birth order can be constructed only as period rates and not as cohort rates.
4. It is possible to standardize means, percentages, proportions, and ratios as well as rates.
THETOTAL
FERTILITY RATE
By now you know that the standardization technique is quite

general, and we could apply it to many refined usessuch as
the computation of age- and marital-status-specific fertility
rates standardized on the educational composition of a standard population. Instead, we now turn to a discussion of the
total fertility rate (TFR), which is a standardized rate whose
values are particularly useful in interpreting the cumulative
fertility implied by a given set of age-specific fertility rates.
The total fertility rate is defined as the sum of the agespecific fertility rates for women, when age is given in single
years. We would usually perform the following calculation
to get the total fertility rate:
92
TFR = [sum of the age- specific fertility rates] x 1,000

= ( F > 1,000
j:
where TFR = total fertility rate,

X means one should add up the age- specific rates, and
X
F = the age- specific rates for the age group*, * +1.

x
The total fertility rate is a standardized measure because the

age-specific fertility rate at each age is multiplied by a standard population, usually of 1,000 persons, as above. In other
words, the total fertility rate assumes a "rectangular" age
distribution for the standard population with the same number of persons at each year of age, namely l,000. In practice,
it is usual to sum rates for five-year age groups and to assume that the age-specific rates for each single year are accurately summarized by the average rate for the five-year age
group. The formula then becomes
3
TFR = 5 ,^(1,000)
(see Table 3.14).
The TFR is only one type of standardized rate, but its

use has been particularly widespread because it has a useful
interpretation. The total fertility rate summarizes a hypothetical fertility history analogous to the hypothetical mortality history of a cross-sectional life table. It estimates the
total number of live births 1,000 women would have if they
all lived through their entire reproductive period and were
subject to a given set of age-specific fertility rates. In other
words, the total fertility rate reports the average number of
live births among 1,000 women exposed throughout their
childbearing years to the schedule of age-specific fertility rates
being used, assuming no women died during the childbearing
years.
Actually, age-specific fertility rates change from year
3. The total fertility rate may be expressed either per woman or per 1,000 women.
In other words, the constant k is either 1 or 1,000. In this Guide, we express the
rate per 1,000 women. If sources for the tables give the TFR (or GRR or NRR) per
woman, we have converted the constant to 1,000 women.
93
Table 3.14. Calculation of total fertility rates for the United States:
1957, 1976, and 1988
Age-specific fertility rates
per 1,000 women for:
Women's ages
1957
1976
1988
10-14
15-19
20-24
25-29
30-34
35-39
40-44
45-49
Sum
1.0
96.3
260.6
199.4
118.9
59.9
16.3
1.1
753.5
1.2
52.8
110.3
106.2
53.6
19.0
4.3
0.2
347.6
1.3
53.6
111.5
113.4
73.7
27.9
4.8
0.2
386.4
3,767.5
1,738.0
1,932.0
Sumx5 - T F R
(per 1,000 women)
Source: United States, National Center for Health Statistics (1990, table 1-6, p. 7|.
to year, and it is not likely that the age-specific rates for a

specific calendar year would remain the same throughout
the reproductive years of a woman. Just like the measures
from a cross-sectional life table, the total fertility rate reflects what would happen to a hypothetical or "synthetic"
cohort of women. The rate can be interpreted to reflect completed family size only when we assume that the age-specific fertility rates for women 20-24 years old now will still
be the same when women 15-19 become 20-24 in five years,
and when we also make similar assumptions for the other
age groups.
In the late 1980s the United Nations Statistical Office
(1992, table A-12) estimated total fertility rates as high as
8,490 (Rwanda) and as low as 1,330 (Italy). Higher total fertility rates are found in the less developed areas than in the
more developed areas, just as are higher crude birth rates and
higher general fertility rates.
In fact, all the common measures of fertility we have
discussed thus far are usually highly correlated with one another. Using the fertility measures for 50 nations with reliable data for the 1955-60 period, Bogue and Palmore (1964)
reported the following correlations: (a) .992 between the crude
94
birth rate and the general fertility rate, (b) .980 between the
crude birth rate and the general fertility rate standardized on
the estimated age composition of the world, and (c) .982 between the crude birth rate and the total fertility rate. These
three correlation coefficients summarize only a few of the
relationships, but coefficients above .979 were found between
all the fertility rates we have presented thus far except for
the various age-specific rates. The age pattern of fertility is
more variable within a specific overall fertility level, but even
the lowest correlations were still quite high and can be illustrated by the following two values: the lowest correlation
between the total fertility rate and an age-specific rate was
.711 and the lowest correlation between the standardized GFR
and an age-specific rate was .689. Even the correlations between age-specific rates without being controlled for the overall fertility level were .425 or greater.
Since the various measures of fertility are so highly
correlated, you may well ask why there are so many of them.
Why don't demographers use just one? There are several reasons:
1. The data necessary for calculating any given measure may not be available. For example, for a certain country
you may be able to compute only the crude birth rate because data on the age and sex distribution or on live births by
age of mother are not available.
2. We cannot be certain that the high correlations of
the 1955-60 period have always obtained in the past, and
they may not obtain in the future. Rapid changes in fertility
are occurring in some countries, and the age distribution depends on fertility. Hence, we may get different results in the
future. The articles we referred to earlier when we discussed
the effects of the age and marital status compositions of populations on their fertility rates (Freedman and Adlakha 1968;
Hirschman 1986; Ogawa and Retherford 1993; Peng 1993;
Retherford and Rele 1989) illustrate the types of changes that
can occur and how the different measures help us understand
more comprehensively what has been happening.
3. The values of different measures are highly corre-
95
lated, but the values for specific countries may be deviant. It

may not always be wise to assume that because country A
has a higher crude birth rate than country B, the total fertility rate in country A is also higher than in country B. Further, even if the direction of the difference in two rates is the
same with different measures, the amount of the difference
between the fertility rates of two populations may be different, depending on which measure is used.
4. Finally, an important reason for having a variety of
measures is that each measure answers a somewhat different question about the fertility level.
To cite an example of the last point using the rate we
discussed most recently, we can interpret the total fertility
rate in a way that is not possible with either the crude birth
rate or the general fertility rate. Whereas the total fertility
rate summarizes the fertility data for the same group of
women as the general fertility rate, the TFR takes into account the distribution of births within the childbearing years
and uses the same standard population in every calculation.
It is this feature that allows us to interpret the TFR as the
completed family size for a hypothetical cohort of women.
GROSS A N D NET
REPRODUCTION
R A T E S
Other measures give us yet additional information about the

reproductive behavior of a population. One meaningful question, for example, is whether a given set of fertility rates
implies that the population will grow, exactly replace itself,
or decline. In a way, this is more a question about natural
increase (or "reproduction") than about fertility itself. The
gross and net reproduction rates are often used to provide
partial answers to this type of question.
The gross reproduction rate (GRR) is a standardized rate
similar to the total fertility rate except that it is the sum of
age-specific rates that include only female live births in the
numerators." The formula for the calculation is as follows:
4. No firm standard has been established for expressing the gross reproduciton
rate per woman or per 1,000 women. In this Guide we express it per 1,000 women
96
GRR =
'5 times the sum of five- year age-specific^

x 1,000
^fertility rates including only female births J
5
= 5X5^/0.000).
x
Since the number of female live births by age of mother

may not be known, the proportion of all births that are female is often used as a constant multiplier for the age-specific rates to obtain the data required for the gross reproduction rate. An example of the calculation of the gross reproduction rate for Costa Rica in 1984 using this method is given
in Table 3.15.
Table 3.15. Calculation of the gross reproduction rate for Costa Rica:
1984
Women's
ages
ASFRs
(per 1,000 women)
ASFRs x proportion of
births female (0.4916)
96.0
192.1
181.7
131.0
76.8
27.0
3.1
47.2
94.4
89.3
64.4
37.8
13.3
1.5
Sum x 5 - GRR (per 1,000 women)
1,739.5
15-19
20-24
25-29
30-34
35-39
40-44
45-49
Sum
347.9
Source: United Nations Demographic Yearbook 1990 (1992, table 11, p. 332).
Note: As is often done, the few births to women less than 15 and more than 49 years
old were attributed to women of ages 15-19 and 45-49, respectively.
Note that in the above formula, we multiply the sum

of the ASFRs by 5 because we are dealing with five-year rates;
each woman in the hypothetical cohort of ages 20-24 will
experience 5 ^ 0 for five years. This amounts to the same
thing as summing the single-year ASFRs, as we did when
calculating the TFR.
Like the total fertility rate, the gross reproduction rate,
when expressed per 1,000, can be interpreted as the number
to maintain consistency with the age-specific rates, the crude rate, and the general
fertility rate. It is probably somewhat more common to express the rate per woman.
97
of daughters expected to be born alive to a hypothetical cohort of 1,000 women if no women died during the childbearing
years and if the same schedule of age-specific rates applied
throughout the childbearing years. The advantage of using
only female births in the calculations is that the GRR then
measures the extent to which a hypothetical cohort of women
will replace itself, provided no women die in the childbearing
years.
In 1985 the gross reproduction rate for the world's total
population was estimated to be 1,680 per 1,000 women
(Keyfitz and Flieger 1990, 65). Values were as high as 4,100
(Rwanda) and as low as 630 (Federal Republic of Germany).
For the 1985-90 period the average GRR for the less developed regions was around 1,900, ranging between 1,121 and
3,343, while the average for the more developed regions was
about 936 (see Table 3.16). Whereas 97 percent of the more
developed regions had gross reproduction rates of 1,299 or
less, 82 percent of the less developed regions had GRRs of
1,600 or more and 49 percent of them had rates of at least
2,700 (see Table 3.17).
Of course, the gross reproduction rate measures only
fertility. It makes no allowance for the fact that some women
may die during the childbearing years. For a more accurate
measure of the replacement of women by their daughters in
the hypothetical cohort, we must use the net reproduction
rate.
The net reproduction rate (NRR) is a measure of the
number of daughters who will be born to a hypothetical cohort of women, taking into account the mortality of the
women from the time of their birth. Hence, the net reproduction rate estimates the average number of daughters who
will replace a cohort of 1,000 female infants by the time the
cohort has been subjected to the risk of mortality from ages
0 to 49 and the risks of live birth from ages 15 to 49. Like the
TFR and the GRR, the NRR may be expressed per woman or
per 1,000 women. We start with a hypothetical cohort of 1,000
girls just born. Only a certain proportion of these 1,000 girls
98
will live to reach the childbearing period. Within the

childbearing period mortality will also take its toll, so that a
given woman may bear daughters through age 20 or age 30
say, but not live to age 50. The net reproduction rate provides an estimate of replacement in the hypothetical cohort,
given mortality levels taken from an appropriate life table.
We show how to calculate the net reproduction rate in
Table 3.18. First we enter the age-specific fertility rates, in-
Table 3.16. Estimated crude birth rates and gross reproduction rates
for world regions: 1985-90
Crude
birth
rate
Gross
reproduction
rate
All regions
27.0
1,671
Less developed regions
30.7
1,900
Region
More developed regions
14.8
936
Africa
East Africa
Central Africa
North Africa
Southern Africa
West Africa
44.5
48.1
46.6
36.9
33.7
48.2
3,045
3,343
3,182
2,485
2,193
3,338
Latin America
Caribbean
Central America
South America
27.9
25.2
31.8
26.8
1,657
1,442
1,910
1,593
North America
15.9
921

East Asia
Southeast Asia
South Asia
West Asia
27.5
20.5
29.9
33.9
35.2
1,681
1,121
1,818
2,271
2,456

Eastern Europe
Northern Europe
Southern Europe
Western Europe
12.8
14.9
13.6
11.6
12.2
833
1,023
897
750
770
Former USSR
18.9
1,184
Oceania
19.5
1,228
Source: United Nations Statistical Office (1992, tables A-7, A-13, pp. 150-52, 17476).
Note: Source values for TFR have been multiplied by 0.4873, a rough average
proportion of births that are female, calculated from United Nations Statistical
Office, Demographic Yearbook 1990(1992, table 10).
Table 3.17. Distribution of countries by world region and level of gross reproduction rate: 1985-90
Level
of GRR
Less
More
Asia
Europe
(excluding
(excluding
All
developed
developed
North
Latin
former
former
Former
regions
regions
regions
Africa
America
America
USSR)
USSR)
USSR
Oceania
131
99
32
43
22
34
26
Under 900
21
17
15
900-1,299
Total
21
14
10
1,300-1,599
1,600-1,999
12
12
2,000-2,399
11
11
0
0
2,400-2,699
2,700-2,999
11
11
3,000-3,299
17
17
13
3,300-3,599
16
16
13
3,600+
Source. United Nations Statistical Office (1992, table A-14.)

Note: Countries with populations under 1 million are excluded. More developed regions, as defined by United Nations Statistical Office (1990), include North America, Japan, Europe,
Australia, New Zealand, and the former USSR. Less developed regions include all regions of Africa, Latin America, and Asia except Japan and all regions of Oceania except Australia
and New Zealand. Figures are'from estimates prepared by the United Nations Population Division. Source values for the TFR have been multiplied by 0.4873, a rough average
proportion of births that are female, calculated from United Nations Demographic Yearbook 1990 (1992, table 10).
100
eluding only female live births (column 3). Next, we enter

the values for the number of person-years lived in each age
interval, using the ^L column from an abridged life table for
females in the current period (column 4). Then we take the
product of the L values and the age-specific rates (column 3
multiplied by column 4). The L values already refer to a
five-year period so that we do not need to multiply by 5 to
derive the NRR as we do with the GRR. The NRR is simply
the sum of the products of columns 3 and 4, or 1,165.92 per
1,000 in the present example.
x
Expressed in a formula, the calculation of the net reproduction rate is as follows:

NRR = the sum of the multiplications of (a) each five-year
age-specific fertility rate including only female live
births and (b) the number of person-years lived by
Table 3.18. Calculation of the gross and net reproduction rates and the length of a generation
for the nonwhite population: United States, 1988
Ages
(1)
10-14
15-19
20-24
25-29
30-34
35-39
40-44
45-49
Sum
Midpoint
of age
interval
(2)
Female
births
per 1,000
women per
year*
(3)
Personyears
lived in
age interval (per
female)
- LJ 100,000
(4)
Female
births per
1,000 women
for 5-year
period
- col. (4)
x col. (3)
(5)
Col. (5)
x col. (2)
(6)
12.5
17.5
22.5
27.5
32.5
37.5
42.5
47.5
2.06
45.76
74.23
59.73
38.06
16.34
3.41
0.20
4.90814
4.89889
4.88317
4.85954
4.82719
4.78262
4.72086
4.63498
10.11
224.17
362.48
290.26
183.72
78.15
16.10
0.93
126.375
3,922.975
8,155.800
7,982.150
5,970.900
2,930.625
684.250
44.175
na
239.79
na
1,165.92
29,817.250
Gross reproduction rate = sum of col. (3) x 5 - 1,198.95.

Net reproduction rate = sum of col. (5) - 1,165.92.
Length of a generation - 29, 117.250/1,165.92 = 25.574 years.
Source: United States, National Center for Health Statistics (1990, tables 1 -6, 1 -59, 6-4).
nanot applicable.
a. Calculated by multiplying the proportion female of births in each five-year age group by the age-specific fertility
rate for that age group.
101
the average female in the stationary population for

the age interval corresponding to the fertility rate
f
= XU')|
L^
S'-x
where NRR = net reproduction rate,
means one should sum the products for every age group.
f
x
= the age-specific fertility rates (per 1,000 women ) including only female live births in the numerator, and
5
(AY
the number of person- years lived per woman

in the age interval x referring to the exact age]
at the beginning of the age interval
Keyfitz and Flieger (1990, 65) estimate that the average

NRR in 1985 was 1,430 for the world's total population, 900
for the more developed regions, and 1,590 for the less developed regions. They also cite rates in 1985 as high as 3,250
(Kenya) and as low as 620 (Federal Republic of Germany). It
is difficult to interpret the precise meaning of these net rates
unless we compare them with the gross reproduction rates.
A country may have a low net reproduction rate because fertility rates are low, because mortality rates are high, or both.
To take two examples from a single country, Japan during
1930-34 had a GRR of 2,320 and an NRR of 1,620 per 1,000
women. With these data we can infer that fertility was moderately high and mortality was high. By 1980, however, Japan had a gross reproduction rate of 840 and a net reproduction rate of 830. Both fertility and mortality rates were low.
You may find it useful to interpret other values shown in
Tables 3.19, 3.20, and 3.21 or to refer to a more complete
listing such as that given in Keyfitz and Flieger (1990).
The net reproduction rate is a measure of how many
daughters would replace 1,000 women if age-specific fertility and mortality rates remained constant indefinitely. Con-
5. Using fertility rates per 1,000 women is equivalent to multiplying their sum by
1,000, as we did above for the TFR and the GRR.
Table 3.19. Measures of reproduction: selected countries, recent years

Vital rates
Country and date

Mauritius
1960
1970
1980
Chile
1960
1970
1980
Mexico
1960
1970
1980
Canada
1961
1970
1980
1985
United States
1960
1970
1980
1985
Taiwan
1960
1970
1980
1985
Japan
1960
1970
1980
Peninsular Malaysia
1970
1980
1985
Crude rates
Intrinsic rates
Gross
reproduction rate
Net
reproduction rate
Mean
age at
birth
Life expectancy
at birth"
Natural
increase
Births
Deaths
Births
Deaths
2,800
1,820
1,420
2,370
1,650
1,340
28.22
28.71
27.40
57.7
63.4
66.2
31.40
27.06
20.79
41.19
9.45
10.07
9.78
17.62
10.72
39.31
25.97
26.40
10.70
7.78
7.21
2,300
1,650
1,210
1,870
1,450
1,150
29.30
28.35
27.08
56.7
62.9
69.7
21.80
13.34
5.30
33.53
24.12
16.75
11.73
10.78
11.45
33.79
25.24
21.13
12.12
8.78
6.67
3,140
2,400
2,290
2,540
2,210
2,140
29.06
29.20
28.95
57.2
66.7
68.8
33.10
28.04
27.12
44.11
34.51
33.12
11.01
6.47
5.99
46.14
41.67
34.85
11.32
9.49
6.24
1,860
1,150
850
810
1,790
1,110
830
880
27.79
27.14
26.92
27.41
71.4
72.8
75.1
76.4
21.26
3.93
-6.78
-8.18
27.58
15.28
9.47
8.74
6.32
11.35
16.25
16.92
25.76
17.47
15.42
14.82
7.77
7.32
7.13
7.15
1,780
1,200
890
900
1,710
1,170
870
880
26.37
26.00
25.97
26.31
70.2
71.0
73.9
74.9
20.75
5.90
-5.14
-4.81
27.31
16.72
10.36
10.40
6.56
10.82
15.50
15.21
23.65
18.32
15.92
15.75
9.39
9.43
8.77
8.74
2,810
1,940
1,220
910
2,540
1,860
1,190
890
29.33
27.58
26.17
26.42
64.2
69.1
72.1
73.6
32.31
22.75
6.56
-4.23
38.93
29.19
17.12
11.07
6.72
6.44
10.56
15.30
39.62
27.16
23.35
17.92
6.96
4.90
4.76
4.80
980
1,000
840
920
980
830
27.88
27.83
27.77
67.9
72.2
76.3
-2.89
-0.89
-6.65
12.71
12.89
9.52
15.60
13.78
16.17
17.24
18.53
13.51
7.48
6.83
6.19
2,380
1,880
1,890
2,170
1,790
1,820
29.43
29.30
27.77
65.5
69.1
70.4
26.97
20.11
20.48
34.11
27.28
27.12
7.14
7.17
6.64
32.49
30.29
31.34
6.99
5.55
5.27
Denmark
1960
1970
1980
1985
France
1961
1970
1980
1985
Hungary
1960
1970
1980
1985
Netherlands
1960
1970
1980
1985
Romania
1960
1970
1980
Former USSR
1959
1970
1979
1987
Australia
1960
1971
1980
1985
1,230
960
750
710
1,190
940
740
690
26.95
26.67
26.83
27.73
72.4
73.5
74.3
74.6
6.54
-2.51
-11.14
-13.05
17.22
11.88
7.96
7.23
10.68
14.39
19.11
20.20
16.48
14.37
11.18
10.51
9.40
9.79
10.92
11.42
1,370
1,210
950
890
1,320
1,180
930
870
28.10
27.12
26.83
27.48
70.9
72.4
74.4
75.5
9.97
5.97
-2.66
-4.90
19.39
16.50
11.40
10.20
9.42
10.53
14.07
15.09
18.07
16.75
14.85
13.93
11.32
10.63
10.15
10.01
970
950
940
890
910
910
910
870
25.78
25.47
24.63
25.04
68.1
69.3
69.2
69.2
-3.47
^.57
^.74
-5.65
12.42
12.04
11.85
10.89
15.89
15.60
15.59
16.54
14.62
14.69
13.88
12.23
10.08
11.63
13.57
13.86
1,530
1,260
780
740
1,490
1,230
770
730
29.75
28.16
27.73
28.42
73.4
73.7
76.0
76.5
13.54
7.44
-9.53
-11.21
21.53
17.30
8.28
7.59
7.99
9.86
17.81
18.80
21.07
18.32
12.81
12.29
7.64
8.41
8.08
8.47
1,120
1,400
1,190
1,010
1,310
1,140
26.43
26.63
25.27
65.2
68.1
69.3
0.38
10.29
5.20
15.12
20.50
16.86
14.75
10.20
11.67
18.93
21.09
17.97
9.23
9.54
10.44
1,370
1,180
1,110
1,230
1,270
1,120
1,050
1,180
28.43
27.46
26.49
26.42
68.1
67.7
67.5
69.5
8.41
4.10
1.95
6.28
19.05
16.27
14.83
17.13
10.65
12.17
12.87
10.86
25.21
17.48
18.19
19.92
7.39
8.74
10.05
9.85
1,700
1,400
920
940
1,640
1,360
900
920
27.49
26.92
27.14
27.69
70.8
71.5
74.7
75.5
18.27
11.54
-3.76
-2.97
25.28
20.30
10.92
11.22
7.01
8.76
14.68
14.19
22.62
21.15
15.35
15.67
8.65
8.47
7.40
7.53
Source: Keyfitz and Flieger (1990, 64101).

a. Simple average of male and female values. Intrinsic vital rates, much like the GRR and NRR, refer to what would happen if the ASFRs and ASDRs were to continue
indefinitely into the future.
b. Female population.
Table 3.20. Gross and net reproduction rates: Europe, Great Depression years, post World-War II, and recent past
Middle 1980s
GRR
NRR
GRR
NRR
GRR
NRR
Northwestern and Western Europe

Austria
Denmark
England and Wales
France
Germany'
Netherlands
Norway
Sweden
890
1,040
930
1,100
800
1,310
1,040
820
740
920
810
920
720
1,190
960
730
1,300
1,270
1,330
1,350
1,220
1,480
1,410
1,150
1,240
1,240
1,290
1,320
1,170
1,430
1,370
1,130
714
708
869
888
625
739
814
840
702
697
856
873
604
729
801
828
Southern and Eastern Europe

Greece
Hungary
Italy
Poland
Portugal
Former Yugoslavia
1,870
1,390
1,580
1,710
1,870
2,200
1,250
1,040
1,220
1,240
1,290
1,390
1,090
910
1,300
1,220
1,520
1,280
1,000
860
1,220
1,150
1,350
1,150
806
892
777
1,132
830
1,000
784
867
765
1,100
810
959
E a r l
Middle 1960s
Region and
country
1 9 3 0 s
Sources: Office of Population Research (1950, 172-78, 1968,249-54; United Nations Statistical Office, Demographic Yearbook 1986, (1988, table 22, pp. 548-72).
a. Federal Republic of Germany after World War II.
b. Estimate prepared by the UN Population Division.
Date
1985
1985
1985
1985
1985
1985
1985
1985
1985
1985
1980-85
1985
1985
1980-85
105
Table 3.21. Gross and net reproduction rates, by color: United States, 1905-10 to 1988
NRR
GRR
Year
Total
White
Nonwhite
Total
White
Nonwhite
1905-10
1930-35
1935-40
1946-49
1950-54
1955-59
1960-64
1965-69
1970-74
1975
1980
1985
1988
1,790
1,110
1,100
1,510
1,630
1,800
1,690
1,280
1,030
864
896
898
943
1,740
1,080
1,060
1,480
1,560
1,730
1,620
1,220
980
819
850
853
883
2,240
1,340
1,410
1,780
2,070
2,330
2,160
1,700
1,330
1,120
1,143
1,112
1,208
1,340
980
980
1,420
1,550
1,730
1,620
1,240
1,000
841
876
881
924
1,340
970
960
1,400
1,500
1,670
1,570
1,190
960
800
833
838
868
1,330
1,070
1,140
1,540
1,840
2,110
1,980
1,570
1,250
1,076
1,106
1,082
1,175
Sources: 1905-40: Office of Population Research (1950, 172); 1946-74: Office of Population Research [1979, 353);
1975-88: United, States, National Center for Health Statistics (1990, table 1-4, p. 5|.
Note: Source data for 1905-10 through 1970-74 are multiplied by 1,000.
sequently, rates above 1,000 mean that eventually the population would increase and rates below 1,000 mean that eventually the population would decrease, provided that the agespecific rates remained the same and no migration occurred.
Rates such as 3,127 imply an eventual speedy rate of natural
increase if the age-specific fertility rates do not decline.
THE M E A N LENGTH
OF A G E N E R A T I O N
Another measure of replacement that follows easily from the

calculations performed for the net reproduction rate is the
mean length of a generation. This measure answers the question, On the average, how many years after birth does a
woman replace herself with female children? The mean
length of a generation indicates the speed with which each
woman replaces herself with potential new mothers.
The length of a generation is a weighted sum of the
female births per 1,000 women for each five-year period, all
divided by the net reproduction ratethat is, it is the average age of women at the birth of their children. (More pre-
106
cisely, it is the average age at which the 1,000 women in our

hypothetical cohort give birth. The average age at birth for
women in the actual population is influenced by the age structure of the population.) The weights used are the ages of the
women. We illustrate the computation of this measure in
Table 3.18 using data for the U.S. nonwhite population in
1988. To obtain the mean length of a generation, we multiply the midpoints of each age interval (column 2) by the number of female births per 1,000 women for the five-year period
(column 5) and then divide the result by the net reproduction rate (the sum of column 5).
From the calculation procedure, you can see that the
length of a generation is affected by two components: (1) the
overall fertility and mortality levels and (2) the proportion of
fertility that occurs in each age group. This is true because
every age-specific female birth rate is affected by the overall
fertility and mortality level (e.g., columns 3 and 4 would have
lower entries in general if the overall level of fertility were
lower and mortality were higher) and also because higher
age-specific fertility rates at younger ages lead to a lower value
for the mean length of the generation than do higher ASFRs
at older ages.
6
For the countries and years listed in Table 3.19, values

of the mean length of a generation have ranged between 24.63
(Hungary, 1980) and 29.75 (Netherlands, 1960). This means
that, in the absence of migration and of changes in age-specific fertility and mortality rates, the average woman will
replace herself with daughters in no fewer than 24 and no
more than 30 years.
6. Technically there is a slight difference between the length of a generation (T)
and the mean age at childbirth (called the mean age of the net maternity function
by Keyfitz and Flieger 1990]. The former is the average time between the births of
members of one cohort and the births of their daughters, whereas the latter is
"the average age |u) at which women bear their girl children. Generally, Tis slightly
smaller than \i in increasing populations, while in decreasing populations . . . the
opposite holds true" (Keyfitz and Flieger 1990, 3031). The values we cite are
actually p., not T.
107
The length of a generation is important because it affects the growth rate of a population independently of the
number of children born as measured by the net reproduction rate. The net reproduction rate tells us how much a population is growing per generation. It does not tell us how long
the generation is. The more rapidly a generation replaces itself, the more rapidly it will add new members to the population (at whatever rate per generation prevails).
An example should help you to understand how to use
the net reproduction rate in combination with the length of
generation. The United States has a shorter generation length
(as shown by the mean age at childbirth) than Western Europe because the average age at marriage is earlier and
childbearing occurs at younger ages in the United States (Table
3.19). Therefore, even if the NRRs were the same in the
United States and Western Europe, the U.S. population growth
rate would be higher because the cycle of reproduction is
repeated more rapidly.
For these reasons the age pattern of fertility decline in
countries with high fertility is important. If the net reproduction rate falls by 10 percent as a result of fertility declines
among older women, it will have less effect than an equal
decline among younger women. To illustrate, women in India have their children at early ages (as compared, for example, with Chinese women in Singapore or Malaysia or
Taiwan). This means that the population growth rate for India is likely to be higher even if the total number of children
born per woman is no greater there than in other populations where childbearing takes place at older ages.
Changing the age at which women bear children can,
in itself, have an effect on the growth rate. Information on
the length of generation should therefore be important to
policymakers and family planning program administrators.
They should know that, in the long run, preventing births
among young women is more important for reducing population growth than preventing births among older women.
108
EXERCISE
Alter the fertility rates in column 3 of Table 3.18 in such a

way as to retain the same sum (i.e., keep the same gross reproduction rate). Do this by increasing the rates for younger
women and decreasing the rates for older women. What effect does this have on the net reproduction rate? What effect
does this have on the mean length of a generation?
EXERCISE
How would you interpret the following information for various countries? Assume that there is no net migration affecting the age structure in any of these countries.
Country
A
B
C
D
E
F
THIRD SET O F
TRUE/FALSE
QUESTIONS
Gross
reproduction
rate
Net
reproduction
rate
Crude
birth
rate
Crude
death
rate
1,000
1,000
3,000
3,000
1,500
3,000
985
985
2,950
1,000
1,485
1,500
14
17
Not available
45
Not available
45
14
6
Not available
45
Not available
22

or false.
1. The net reproduction rate can never be higher than the
gross reproduction rate.
2. If the gross reproduction rate declines in any given year,
it inevitably means that at least a minority of the
women in the childbearing years will end up with fewer
children than they would have had prior to the decline.
3. Regardless of which fertility measure we use, we will
find that fertility is higher in most of the less developed areas of the world than in the more developed areas.
4. For all practical purposes, the gross reproduction rate is
equal to the product of the total fertility rate times the
proportion of live births that are female.
5. A gross reproduction rate of 1,500 is very high.
109
6. A total fertility rate of 2,350 is very high.

7. For most of human history, it is likely that net reproduction rates close to 1,000 were common.
THIRD SET OF
MULTIPLE-CHOICE
QUESTIONS
1. The net reproduction rate in the United States is now approximately:

(a) 500-1,000 per 1,000 women.
(b) 1,000-1,500 per 1,000 women.
(c) 2,500-3,500 per 1,000 women.
(d) 4,000-5,000 per 1,000 women.
What is the net reproduction rate in your own country?
2. A net reproduction rate of more than 1,000 means that:
(a) a population will certainly increase in the future.
(b) a population will certainly decrease in the future.
(c) a population will eventually increase if age-specific fertility and mortality rates remain fixed and there is no
migration.
(d) a population will eventually decrease if age-specific fertility and mortality rates remain fixed and there is no
migration.
(e) a population will remain at about the same size if agespecific fertility and mortality rates remain fixed and
there is no migration.
3. The American Hutterites had a gross reproduction rate of
4,000 and a net reproduction rate of 3,660 during one period.
This indicates that:
(a) both mortality and fertility were very high.
(b) both mortality and fertility were very low.
(c) fertility was very high and mortality was moderately
low.
(d) mortality was moderately high and fertility was very
low.
(e) mortality was very low and fertility was only moderately high.
110
4. The net reproduction rate is a measure of the:

(a) annual excess of births over deaths.
(b) annual rate at which women are replacing themselves
on the basis of prevailing fertility and mortality, assuming no migration.
(c) decennial growth rate of the population.
(d) per generation growth rate assuming current age-specific fertility and mortality rates and no net migration.
5. In two countries, A and B, the age-specific fertility rates
per 1,000 women are as follows:
Age group
10-19
20-29
30-39
40-49
Country A
Country B
25
100
50
25
25
100
50
25
In country A, 60 percent of the population is female whereas

only 50 percent of the population is female in country B. In
country A, 35 percent of the females are between the ages of
10 and 49, whereas 40 percent of the females are between the
ages of 10 and 49 in country B.
(a) Country A has a higher gross reproduction rate than
country B.
(b) Country B has a higher gross reproduction rate than
country A.
(c) Country A has the same gross reproduction rate as country B.
(d) Country A has the same net reproduction rate as country B.
(e) The crude birth rates are the same in both countries.
(f) The general fertility rates are the same in both countries.
(g) Two of the above are correct.
(h) Three of the above are correct.
6. Populations with net reproduction rates of 1,000 per 1,000
women:
(a) invariably have low age-specific fertility rates.
111
(b) have low crude birth rates but may have high age-specific fertility rates.
(c) have declining age-specific fertility rates.
(d) may have either high or low age-specific fertility rates.
(e) invariably have low crude birth rates.
CENSUS MEASURES
OF FERTILITY
Thus far we have discussed an interrelated system of measures that usually require census data for the denominators
and vital statistics data for the numerators. In many countries, vital registration systems either do not exist or are inaccurate: they underregister the number of vital events and
often misclassify the characteristics (e.g., age, place of residence) of the persons who gave birth or died. In countries
where this is true, other measures of fertility based on census information have been used as substitutes for the measures we have already discussed. Such census measures have
an advantage over vital-statistics measures because censuses
usually collect much more information than do birth certificates on many characteristics of individualssuch as income,
education, rural-urban residencewhich are important because of their effects on fertility. Therefore, census measures
allow a much more thorough study of differential fertility.
Most nations have had at least one census in the last
10 years, and the data collected can be used to calculate various indirect measures of fertility. The most common indirect measures are the following:
1. the ratio of children 0-4 years old to women of ages 1549 or 15-44 years
2. the ratio of children 5-9 years old to women of ages 1549 or 15-44 years
3. the percentage of the total population 0-4 years old
7
7. Sample surveys are sometimes used to collect data for both the numerators and
the denominators, and a census can collect the information for both numerators
and denominators. It is difficult, however, to obtain accurate reporting on births
in a census, given the levels of training and supervision normally employed.
112
6. the number of children ever born to women, by fiveyear age groups of the women
7. the number of women's own (as opposed to adopted)
children under age 5, by five-year age groups of women
We discuss only the first and sixth measures here because
the problems of interpretation and use are similar for the
other measures.
The ratio of children 0-4 years old to women of ages
15-44 or 15-49 is often called the child-woman ratio (CWR).
It can be expressed algebraically as follows:
CWR=*-^V
or k-^y
pf
pf
30M5
35M5
where P = population 0 - 4 years old

5
,P/
(or
P/ )
J5
number of women 15 - 44 (15 - 49) years old, and
it = 1,000.
The child-woman ratio is based not on births, but on
the survivors of births occurring in the last five years. One
drawback of this measure is that the deaths of children in
those five years are not accounted for and although the deaths
of women in the childbearing years compensate partly for
the deaths to children, the ratio understates fertility. Moreover, because the ratio deals with survivors instead of actual
births, two populations may have the same fertilty rates but
the child-woman ratios will not reflect this fact if one population has higher child mortality rates than the other. The
population with the higher death rates will have a lower childwoman ratio.
;
Using the child-woman ratio poses several other problems. One is that, if the ratio has surviving children of ages
0-4 in the numerator, it measures past fertilityfertility 2.5
years, on the average, before the census date. Another is that
young children are more likely than older people to be
underenumerated in a census. For that reason demographers
sometimes select children 5-9 years old for the numerator,
but this aggravates the problem of measuring current fertil-
113
ity because the ratio then refers, on the average, to fertility

7.5 years before the census date.
Even with all these problems, any reasonable measure
is better than none, and the child-woman ratio has often been
used when vital registration data are lacking for a country or
for subdivisions of a country. Palmore (1978), for example,
summarized the child-woman ratios for most nations and
territories, using data from the 1970 round of censuses. He
found that the the child-woman ratio ranged from 313 (in
Sweden) to 928 (in Western Samoa) per 1,000 women and
that the ratio was well correlated with more direct measures
of fertility in countries with reliable data. For 56 nations with
reliable data the child-woman ratio had the following correlations with direct measures of fertility: .961 with the crude
birth rate, .975 with the general fertility rate, and .970 with
the total fertility rate. On the basis of this information,
Palmore developed a series of equations for determining the
level of direct measures of fertility using such census meaTable 3.22. Percentage ever married and number of children ever born for women of ages
40-44 and 30-34: United States, selected years 1940-1990
Children ever born
Percentage
ever married
Percentage
childless among
those ever married
Per 1,000
women
Per 1,000
ever-married
women
40-44
1990
1980
1970
1960
1950
1940
92
95
95
94
92
90
11.3
6.6
8.6
14.1
20.0
17.4
2,045
2,988
2,952
2,409
2,170
2,490
2,167
3,105
3,096
2,564
2,364
2,754
30-34
1990
1980
1970
1960
1950
1940
82
90
93
93
91
85
16.8
13.7
8.3
10.4
17.3
23.3
1,589
1,826
2,640
2,445
1,871
1,678
1,788
1,970
2,804
2,627
2,059
1,964
Age of women
in given year
Sources: United States, Bureau of the Census (1966, table 1, pp. 11-12; 1979, tables 6, 7, pp. 32-34; 1982:, table 12,
pp. 41-43|; Bachu (1991, table 1, p. 17|.
114
sures as the child-woman ratio and other selected facts about

each population. This material is beyond the scope of the
present Guide, but you may want to refer to those techniques.
Data on children ever born (CEB) are collected in fewer
censuses than the data required for calculating the childwoman ratio. For measures of this type, the census must
contain a question for each woman asking her how many
live births she has ever had. This information can then be
tabulated by the woman's age, yielding measures of the cumulative fertility of women up to specified points in their
childbearing years. Like the child-woman ratio, statistics on
children ever born measure past fertility and are subject to
the additional problem that children who die young may not
be remembered. Nevertheless, this type of data has been used
widely. Tables 3.22 and 3.23 provide examples from the
United States and Table 3.24 provides an example from Indo-
Table 3.23. Number of children ever born per 1,000 women and per 1,000 ever-married
women, by age: United States, selected years, 1940-90
Ages
All women
15-44
15-19
20-24
25-29
30-34
35-39
40-44
45-49
ver-married women
15-44
15-19
20-24
25-29
30-34
35-39
40-44
45-49
1940
1950
1960
1970
1980
1990
1,238
68
522
1,132
1,678
2,145
2,490
2,740
1,395
105
738
1,436
1,871
2,061
2,170
2,292
1,746
127
1,032
2,006
2,445
2,523
2,409
2,245
1,918"
206
736
1,790
2,640
3,015
2,952
2,707
1,506'
179
554
1,177
1,826
2,457
2,988
3,091
1,248
101
574
1,089
1,589
1,909
2,045
u
1,904
572
987
1,463
1,964
2,414
2,754
2,998
1,859
604
1,082
1,654
2,059
2,247
2,364
2,492
2,314
792
1,441
2,241
2,627
2,686
2,564
2,402
2,357
633
1,064
1,978
2,804
3,167
3,096
2,840
1,965
628
930
1,397
1,970
2,572
3,105
3,185
1,757
718
993
1,329
1,788
2,048
2,167
u
Sources: United States, Bureau of the Census 11966, table 1, p. 12; 1979, table 6, pp. 32-33; 1982, table 12, pp. 41-44,
Bachu (1991, table 1, p. 17|
udata unavailable.
a. Numbers for ages 18-44.
b. Numbers for ages 18 and 19.
115
Table 3.24. Average number of children ever born per woman among ever-married women of
ages 45-49, by province and urban/rural residence: Indonesia, 1980 and 1990
Urban
Rural
Urban and Rural
Province
1980
1990
1980
1990
1980
1990
Aceh
North Sumatra
West Sumatra
Riau
Jambi
South Sumatra
Bengkulu
Lampung
Jakarta
West Java
Central Java
Yogyakarta
East Java
Bali
West Nusa Tenggara
East Nusa Tenggara
East Timor
West Kalimantan
Central Kalimantan
South Kalimantan
East Kalimantan
North Sulawesi
Central Sulawesi
South Sulawesi
Sulawesi Tenggara
Maluku
Irian Jaya
Indonesia
5.66
6.61
6.91
6.72
6.40
6.71
6.97
6.47
5.46
6.20
5.08
5.04
4.58
5.51
6.19
6.62
u
6.22
6.02
6.19
5.87
6.08
6.31
5.99
6.69
6.58
5.80
5.55
5.32
5.86
5.62
5.35
5.73
5.85
6.21
5.73
4.78
5.48
4.79
4.28
4.17
4.50
4.94
5.77
4.63
5.34
5.35
5.27
5.24
4.36
5.58
5.39
5.59
5.24
5.51
5.00
5.21
6.92
6.38
6.48
5.93
6.34
6.58
6.09
6.29
5.89
5.25
4.99
4.56
5.06
6.76
5.63
u
6.16
5.94
5.27
5.32
6.87
6.32
5.51
5.92
6.38
4.09
5.46
5.17
6.30
5.86
5.55
5.31
5.77
5.91
5.77
u
5.41
4.83
4.30
4.05
4.36
6.40
5.48
4.18
5.52
5.34
5.00
4.98
5.26
5.91
5.16
5.63
5.56
4.44
5.03
5.24
6.84
6.43
6.53
5.99
6.43
6.61
6.14
5.50
5.95
5.22
5.00
4.57
5.12
6.68
5.67
u
6.15
5.95
5.44
5.54
6.74
6.32
5.58
5.97
6.40
4.33
5.48
5.19
6.15
5.82
5.49
5.39
5.79
5.96
5.77
4.78
5.43
4.82
4.29
4.08
4.40
6.32
5.50
4.20
5.49
5.34
5.07
5.09
5.06
5.86
5.21
5.62
5.50
4.65
5.02
Source: Indonesia, Central Bureau of Statistics (1993, table 8, p. 24).

udata unavailable.
nesia. One reason for the wide use of this measure is related
to the notion of cohort fertility, a concept we discuss next.
COHORT
FERTILITY
MEASURES
When we discussed the life table in the mortality chapter,

we pointed out that there were two types of life table, the
period or cross-sectional life table and the generation or longitudinal life table. A similar distinction can be made among
fertility measures. Thus far we have discussed mostly.what
are known as period or calendar-year fertility rates. When
116
we discussed the total fertility rate, however, we introduced

the idea of a cohort, albeit a hypothetical or synthetic cohort. We can also construct fertility rates for real cohorts,
and measures so constructed are called cohort fertility measures.
Two types of cohorts commonly used in fertility measurement are marriage cohorts and birth cohorts. If we discuss data for a birth cohort, we refer to the fertility rates for
a group of women all born in the same year or group of years.
For example, we might talk about the 1935-39 birth cohort
of women. If discussing marriage cohorts, we refer to the
fertility rates for a group of women all married in the same
year or group of years (e.g., women of the 1940-44 marriage
cohort). Usually, use of the word "cohort" by itself refers to a
birth cohort, and we devote our discussion here to data for
birth cohorts of women.
One rationale for using birth cohorts to measure fertility hinges on the fact that childbearing in a particular year is
determined in part by how many children women have had
in preceding years, and that number, in turn, is determined
in part by their age. Another rationale for using birth-cohort
measures is based on the argument that cultural ideas about
family size may change over time. Furthermore, other
changes in a society may occur that lead to new patterns of
childbearing in successive generations. Examples of such
changes are a war that disrupts family formation during one
generation (e.g., World War II) and the development of new
methods for controling fertility that were not previously available (e.g., the birth control pill, which was introduced into
many societies in the 1960s).
The fertility of a population may be influenced by both
cohort effects and period effects. For example, an economic
depression may affect many cohorts simultaneously (although
at different stages of their reproductive histories), causing a
low level of period fertility during the depression. Once the
depression has passed, period fertility may rise, and women
who postponed having children during the depression may
117
make up for the postponement. Some cohorts, however, will

have reached the end of their reproductive years by the time
the depression has ended and will no longer be able to bear
children. Such a depression-induced fall in fertility is an example of a period effect.
A cohort effect may also be the product of a depression.
Children born during a depression may tend to be conservative about their own fertility, preferring the certainty of being able to provide for a few children to the risks of having
many children when the economic weather might worsen
once more. In spite of economic good times, they may still
have small completed families. Another cohort living through
the same economic good times may take advantage of the
prosperous conditions to have larger families. The resulting
overall period fertility could be high, low, or average, but it
would be composed of cohorts having various patterns and
levels of fertility.
There is no guarantee that measures of period fertility
and measures of cohort fertility referring to the same time
span will show the same trends. An interesting example was
presented by Barclay, using data for Taiwan in the 1933-52
period (Barclay 1958, 184-88). During that period, the total
fertility rate changed as follows:
Period
1933-42
1938-47
1948-52
Total fertility rate

7,400
6,850
6,250
These period rates clearly show declining fertility. Cohort

measures for the same time period, however, show a different pattern:
Birth cohort
of women
Average number
of children ever born
1888-92
1893-97
1898-1902
1903-07
6.90
6.90
7.25
7.35
The cohort measures clearly indicate rising fertility!

How is this possible? The answer lies in recognizing that
118
both sets of measures are correct but refer to different groups

of women.
How discrepancies like the one for Taiwan occur is clarified by a simple artificial example. Suppose we have the following age-specific rates:
Age-specific fertility rates at ages:
Birth cohort
15-24
25-34
35-44
1891-1900
1901-10
1911-20
1921-30
50
70
110
90
300
300
300
300
200
180
180
220
If we assume that all the births occurred to women 15-44

years old and that there was no mortality, we can make the
following statements:
1. In 1935 the women born in 1911-20were 15-24 years
of age the women of the 1901-10 cohort were 25-34; and the
women of 1891-1900 cohort were 35-44. Hence, the three
figures on the left-to-right upward diagonal (110, 300, 200)
represent the period fertility for the year 1935. The total fertility rate for that year was 6,100. Similarly, the total fertility rate for 1945 was 5,700. These two rates indicate a decrease in fertility.
;
2. Although the period total fertility rates declined between 1935 and 1945, the cohort rates were successively
higher:
1891-1900
1901-10
1911-20
1921-30
5,500
5,500
5,900
6,100
This example demonstrates that it is possible to have

period rates that change in one direction and, at the same
time, cohort cumulative rates that change in the opposite
direction. Such a paradox results from differences in the timing of births for the separate cohorts, which can produce unusually low or high points while the basic cohort trend is in
a direction not indicated by the period rates.
8. Since we are using 10-year age groups, the total fertility rate is the sum of the
age-specific rates multiplied by 10.
119
In the example just given, the age pattern of childbearing

changed during successive birth cohorts, and the change produced a discrepancy between cohort rates and period rates.
Such discrepancies are quite possible if there have been shifts
in the ages at which women marry, or if women plan and
control their fertility and there have been outside causes (the
economic depression mentioned earlier, for example) that lead
to either a postponment or an acceleration of childbearing.
That is, women born between 1891 and 1920 had an increasing proportion of their children in the earlier childbearing
years. Women born after 1920 began having more children in
their later childbearing years.
F O U R T H SET OF
MULTIPLE-CHOICE
QUESTIONS
1. Cohort fertility analyses:

(a) have essentially the same use as the net reproduction
rate.
(b) have essentially the same use as the gross reproduction
rate.
(c) have the advantage of linking current and future fertility rates to past fertility histories of each cohort.
(d) are useful only for populations in which contraception
is not widely used.
(e) refer to the experience of Roman military cohorts.
2. The chief difficulty with the net reproduction rate as a
predictive device for population growth is that it:
(a) excludes the influences of fertility.
(b) makes inadequate allowance for mortality.
(c) is based on the rates of a single year.
(d) overlooks the type of culture possessed by the population.
(e) includes only survivors of births in some past period.
3. Period birth rates and cohort birth rates may exhibit large
differences under which of the following conditions?
(a) When most couples plan their fertility.
(b) When the mean age at marriage is increasing.
(c) When the mean age at marriage is decreasing.
120
(d) Two of the above.

(e) Includes (a), (b), and (c) above.
4. Assuming there were no deaths to children or women in
the past five years, one-fifth of the child-woman ratio should
be approximately equal to:
(a) the average general fertility rate for the past five years.
(b) the total fertility rate for 7.5 years ago.
(c) the crude birth rate for some indeterminate past period.
(d) the average gross reproduction rate for the past five
years.
(e) the average net reproduction rate for the past five years.
ANALYSIS OF
BIRTH I N T E R V A L S
A final type of fertility analysis that is now prominent in the

demographic literature is measuring the length of time between each birth and the next (a birth interval). As stated by
Rindfuss, Palmore, and Bumpass (1982, 5), these measures
are important because
the fertility process is itself a sequential and time-dependent
process. Birth interval analysis allows more precision in
investigating many fundamental questions; it allows the
assessment of the effects of intermediate variables, like
contraceptive use or lactation, and the explication of the effects
of various socioeconomic variables in terms of intermediate
variables.
Methods for properly analyzing birth intervals are still under

development because three complex methodological problems are associated with birth interval analysis.
The first problem has to do with the quality of the data
available for studying birth intervals. Misdating of births or
failure to remember their occurrence can reduce the data
quality.
The second problem has to do with what has become
known as censoring. Censoring occurs when birth intervals
are incomplete, or "open." When a sample survey is conducted, for example, many respondents have not yet completed all of their birth intervals. Some of the open intervals
121
will eventually be closed by a subsequent birth, but when

that will happen is unknown. If you analyze only the open
intervals or only the closed intervals, an analytic bias results
because open intervals tend to be longer than closed intervals. The open intervals tend to be longer partly because some
of the open intervals will never be closed by another birth.
The usual solution to this problem has been to use life table
techniques, with the next birth treated as a "death" and the
initial cohort consisting of women who have had the immediately preceding birth. That is, women of parity n who have
not yet given birth to a child of order n + 1 are like persons of
age x who have not yet died.
We provide an illustration of this life-table approach in
Table 3.25, which is adapted from an article by Bumpass et
al. (1982) that analyzes data from the 1973 National Demographic Survey of the Philippines and the 1974 Korean World
Fertility Survey. In the top panel of the table the proportion
of women who gave birth is tabulated by whether or not contraception had been used during the interval and by the duration of the interval. These figures are the equivalent of q
values. For example, the figure. 14 in the first cell of the table
is the probability that a Korean woman who used contraception would have a second birth within the first 20 months
after having her first birth.
Comparing women who used contraception with those
who did not (i.e., by looking at the proportionate reduction
in birth probabilities in the second panel of the table), we
can make several observations:
1. In the second interval for Koreans and in all intervals shown for Filipinas, the effect of having used contraception is usually distinctly lower in the first duration segment
(0-20 months) than in the second duration segment. Proximate determinants other than contraceptive use were probably the dominant influences in this duration segment. Lactation (and its effect on fecundability) and lower coital frequency immediately after the birth of a child are likely candidates.
2. After the first duration segment, the proportionate
x
Table 3.25. Birth probabilities within successive birth intervals 2, 3, and 4-8, by duration of interval and contraceptive use status: Philippines
and Republic of Korea, 1973-74
Duration of interval and whether
contraception used or not
Interval 2
Korea
Proportion giving birth during interval segment

<21 months
Yes
No
21-26 months
Yes
No
27^2 months
Yes
No
33-44 months
Yes
No
Proportionate reduction in birth
probability due to contraceptive use" (|P -P J/PJ
<21 months
21-26 months
27-32 months
33-44 months
no
Interval 3
Philippines
Korea
Intervals 4-8
Philippines
Korea
Philippines
.14
.23
.28
.43
.03
.13
.23
.32
.02
.09
.19
.26
.21
.41
.20
.40
.08
.31
.14
.36
.04
.19
.11
.29
.28
.42
.21
.35
.15
.44
.17
.35
.07
.31
.11
.28
.45
.59
.48
.50
.36
.67
.36
.48
.12
.46
.20
.42
.39
.48
.34
.22
.35
.50
.38
.03
.74
. 73
.66
.47
.29
.60
.53
.26
.84
. 79
.77
.75
.29
.62
.61
.52
Number of cases
< 21 months
Yes
348
141
577
216
1,928
740
No
1,724
1,779
1,308
1,608
2,917
5,491
21-26 months
Yes
288
96
526
156
1,862
569
No
1,304
1,005
1,114
1,081
2,622
4,001
27-32 months
Yes
202
62
439
118
1,685
441
No
743
570
751
677
2,073
2,716
33-44 months
Yes
132
40
312
77
1,395
325
No
404
351
399
411
1,361
1,840
Source: Bumpass et al. (1982, 248].

Note: The first interval is between marriage and the first birth, the second interval is between the first and second births, etc.
a. Proportions in the second panel calculated from unrounded figures.
124
reduction in birth probabilities due to contraceptive use declines after the second duration segment (21-26 months).
Contraceptive use for spacing births is probably the explanation for this effect.
3. Among Korean women at higher parities, there is an
impressive reduction in fertility with contraceptive use.
The third problem in analyzing birth intervals has to
do with the selectivity of the birth intervals available for
analysis. Selectivity is particularly evident with data from
sample surveys because such surveys typically restrict the
respondents to specific age ranges and restrict them by marital status or other criteria.
A hypothetical survey conducted in 1990 illustrates the
problem. Table 3.26 shows the birth intervals that would be
available for analysis in a survey of all ever-married women
between ages 15 and 49. The horizontal dimension of the
table indicates the ages of the women at the time of the interview; the vertical dimension represents the ages of the
women at the start of a birth interval. Each cell of the table
represents the year in which a birth interval began. (The years
are shown with the leading 19s omitted; for example, "46" is
1946.} Diagonals from the top left to the bottom right of the
table represent the birth intervals begun in a given year.
The solid triangle encloses the intervals actually available for analysis. Notice, first, that intervals begun at age 15
could have been initiated in any year between 1956 and 1990,
whereas intervals begun at age 49 could start only in 1990.
The time periods represented are different for various ages at
initiation. Second, by comparing birth cohorts (each column
in the table) you will notice that the cohorts vary considerably in the possible ages at the beginning of the interval. At
the extremes, women of age 49 could have initiated an interval at any age from 15 to 49. Women of age 15 could have
initiated an interval only at 15. Third, look at the time periods during which the birth intervals were begun. Intervals
begun before 1966, for example, had to be initiated at age 25
or younger. Intervals begun before 1959 had to be initiated at
age 18 or younger.
125
The three points made above clearly illustrate biases

introduced by selectivity. If a woman's age at the beginning
of an interval, her birth cohort, and the time period were all
unrelated to her fertility, these biases could be ignored. Unfortunately, all three of these variables are known to be highly
related to fertility.
The principal question introduced by selectivity biases
is, Of those birth intervals available for analysis, which should
be analyzed? There is no single solution, and a full treatment of the selectivity issue is beyond the scope of this Guide.
If you are interested in pursuing the matter further, you should
read Rodriguez and Hobcraft (1980) and Rindfuss, Palmore,
and Bumpass (1982).
FIFTH SET O F
MULTIPLE-CHOICE
QUESTIONS
1. Censoring refers to the fact that:

(a) women often forget the exact dates of birth of their children.
(b) at the time of data collection, some women have not
completed childbearing.
(c) not all women in the population are interviewed.
(d) data for some countries are suppressed by the government.
(e) Both (b) and (c) above are correct.
2. Selectivity biases in survey data:
(a) are important only in the analysis of birth intervals.
(b) may affect studies of the intervals between marriages,
geographic movements, or job changes.
(c) arise because women often forget the exact ages of birth
of their children.
(d) arise because the survey is. cross-sectional rather than
a complete longitudinal study of each birth cohort of
women.
(e) Both (a) and (c) above are correct.
(f) Both (b) and (d) above are correct.
3. Open birth intervals are likely:
(a) to be longer than closed birth intervals.
Table 3.26. Year in which any birth interval had to begin, given a woman's age at the
place in 1990: all intervals
Age at
beginj g f
n
Age at time of interview
interval
15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
15
16
90 89 88 87 86 85 84 83 82 81 80 79 78 77 76 75 74 73 72 71 70 69 68 67 66 65
\ 9 0 89 88 87 86 85 84 83 82 81 80 79 78 77 76 75 74 73 72 71 70 69 68 67 66
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
32
34
36
37
38
39
40
41
42
43
44
45
46
47
48
49
\ 9 0 89 88 87 86 85 84 83 82 81 80 79 78 77 76 75 74 73 72 71 70 69 68 67
\ 9 0 89 88 87 86 85 84 83 82 81 80 79 78 77 76 75 74 73 72 71 70 69 68
\ 9 0 89 88 87 86 85 84 83 82 81 80 79 78 77 76 75 74 73 72 71 70 69
\ 9 0 89 88 87 86 85
\ 9 0 89 88 87 86
\90 89 88 87
\90 89 88
\ S 0 89
\50
\9
84 83 82 81 80 79 78 77 76 75
85 84 83 82 81 80 79 78 77 76
86 85 84 83 82 81 80 79 78 77
87 86 85 84 83 82 81 80 79 78
88 87 86 85 84 83 82 81 80 79
89 88 87 86 85 84 83 82 81 80
0 89 88 87 86 85 84 83 82 81
\ 9 0 89 88 87 86 85 84 83 82
\ 9 0 89 88 87 86 85 84 83
\ 9 0 89 88 87 86 85 84
\ 9 0 89 88 87 86 85
\ 9 0 89 88 87 86
\ 9 0 89 88 87
\ 9 0 89 88
\ 9 0 89
\90
\
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
\
73 72
74 73
75 74
76 75
77 76
78 77
79 78
80 79
81 80
82 81
83 82
84 83
85 84
86 85
87 86
88 87
89 88
90 89
N. 90
\
71 70
72 71
73 72
74 73
75 74
76 75
77 76
78 77
79 78
80 79
81 80
82 81
83 82
84 83
85 84
86 85
87 86
88 87
89 88
90 89
\90
\^
beginning of the interval and her age at the time of being interviewed, for a survey taking
41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69
64 63 62 61 60 59 58
65 64 63 62 61 60 59
66 65 64 63 62 61 60
67 66 65 64 63 62 61
68 67 66 65 64 63 62
69 68 67 66 65 64 63
70 69 68 67 66 65 64
71 70 69 68 67 66 65
72 71 70 69 68 67 66
73 72 71 70 69 68 67
74 73 72 71 70 69 68
75 74 73 72 71 70 69
76 75 74 73 72 71 70
77 76 75 74 73 72 71
78 77 76 75 74 73 72
79 78 77 76 75 74 73
80 79 78 77 76 75 74
81 80 79 78 77 76 75
82 81 80 79 78 77 76
83 82 81 80 79 78 77
84 83 82 81 80 79 78
85 84 83 82 81 80 79
86 85 84 83 82 81 80
87 86 85 84 83 82 81
88 87 86 85 84 83 82
89 88 87 86 85 84 83
90 89 88 87 86 85 84
90 89 88 87 86 85
90 89 88 87 86
90 89 88 87
90 89 88
90 89
90
57 56 55 54 53 52
58 57 56 55 54 53
59 58 57 56 55 54
60 59 58 57 56 55
61 60 59
62 61 60
63 62 61
64 63 62
65 64 63
66 65 64
67 66 65
68 67 66
69 68 67
70 69 68
71 70 69
72 71 70
73 72 71
74 73 72
75 74 73
76 75 74
77 76 75
78 77 76
79 78 77
80 79 78
81 80 79
82 81 80
83 82 81
84 83 82
85 84 83
86 85 84
87 86 85
88 87 86
89 88 87
89 88
89
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
51 50 49 48 47
52 51 50 49 48
53 52 51 50 49
54 53 52 51 50
55 54 53 52 51
56 55 54 53 52
57 56 55 54 53
58 57 56 55 54
59 58 57 56 55
60 59 58 57 56
61 60 59 58 57
62 61 60 59 58
62 62 61 60 59
64 63 62 61 60
65 64 63 62 61
66 65 64 63 62
67 66 65 64 63
68 67 66 65 64
69 68 67 66 65
70 69 68 67 66
71 70 69 68 67
72 71 70 69 68
73 72 71 70 69
74 73 72 71 70
75 74 73 72 71
76 75 74 73 72
77 76 75 74 73
78 77 76 75 74
79 78 77 76 75
80 79 78 77 76
81 80 79 78 77
82 81 80 79 78
83 82 81 80 79
84 83 82 81 80
85 84 83 82 81
46 45 44 43 42 41
47 46 45 44 43 42
48 47 46 45 44 43
49 48 47 46 45 44
50 49 48 47 46 45
51 50 49 48 47 46
52 51 50 49 48 47
53 52 51 50 49 48
54 53 52 51 50 49
55 54 53 52 51 50
56 55 54 53 52 51
57 56 55 54 53 52
58 57 56 55 54 53
59 58 57 56 55 54
60 59 58 57 56 55
61 60 59 58 57 56
62 61 60 59 58 57
63 62 61 60 59 58
64 63 62 61 60 59
65 64 63 62 61 60
66 65 64 63 62 61
67 66 65 64 63 62
68 67 66 65 64 63
69 68 67 66 65 64
70 69 68 67 66 65
71 70 69 68 67 66
72 71 70 69 68 67
73 72 71 70 69 68
74 73 72 71 70 69
75 74 73 72 71 70
76 75 74 73 72 71
77 76 75 74 73 72
78 77 76 75 74 73
79 78 77 76 75 74
80 79 78 77 76 75
40 39 38
41 40 39
42 41 40
43 42 41
44 43 42
45 44 43
46 45 44
47 46 45
48 47 46
49 48 47
50 49 48
51 50 49
52 51 50
53 52 51
54 53 52
55 54 53
56 55 54
57 56 55
58 57 56
59 58 57
60 59 58
61 60 59
62 61 60
63 62 61
64 63 62
65 64 63
66 65 64
67 66 65
68 67 66
69 68 67
70 69 68
71 70 69
72 71 70
73 72 71
74 73 72
37 36
38 37
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
128
(b) to be shorter than closed birth intervals.

(c) to be about the same length as closed birth intervals.
4. Selectivity biases refer to biases introduced by selectivity
on:
(a) age at the initiation of a birth interval.
(b) time period.
(c) birth cohort.
(d) all of the above.
ADDITIONAL
READING
The literature on fertility analysis is growing rapidly. Consequently, the works mentioned here are necessarily selective
and do not adequately reflect the diversity of the literature.
Further discussion of the methods described in this chapter
can be found in several of the sources listed in Chapter 2. In
particular, you may want to consult Bogue et al. (1993).
Most books of readings on population have one or more
chapters on fertility, as do most demography textbooks. New
methods for constructing fertility rates from deficient data
are being developed regularly. Bogue and Palmore (1964) and
Palmore (1978), mentioned earlier in this chapter, are among
the early illustrations. Prominent methods include the following: the "own-children method" developed principally by
Lee-Jay Cho and colleagues (see Cho, Retherford, and Choe
1986 for a thorough description of the method and Rao et al.
1993 for one example of the method's use); the "Brass methods" developed by William Brass and others (Natarajan and
Singh 1988 discuss the methods and use them to estimate
district-level fertility for India in 1980); and the regression
techniques developed by James Palmore (1978; for an illustration of which, see Swamy et al. 1993), and by J. R. Rele
(1967,1987), Subbiah Gunasekaran and James Palmore (1984),
and others (see the Bogue et al. compendium mentioned above
for many of the others).
Additional materials on the use of both fertility and
mortality measures can be found through judicious sampling
129
of appropriate journals, some of which are listed below. The

list is far from exhaustive. In particular, we include only a
small, illustrative sample of the journals published outside
of the United States and Europe.
Asian and Pacific Population Forum, published from
1974 through 1993 in Honolulu by the East-West Center under
three titles, the other two being the Asian and Pacific Census Newsletter and the Asian and Pacific Census Forum; in
English.
Majalah Demografi Indonesia, published in Jakarta by
Universitas Indonesia, Lembaga Demografi; most articles in
Indonesian, but some in English with Indonesian abstracts.
Demography, a journal of the Population Association
of America, published in Providence, Rhode Island, by the
Department of Sociology, Brown University; in English.
Family Planning Perspectives, published in New York
by The Alan Guttmacher Institute; in English.
Genus, published in Rome by Comitato italiano per lo
studio dei problemi della popolazione in Italian.
International Family Planning Perspectives, published
in New York by The Alan Guttmacher Institute; in English.
Journal of BioSocial Science, published in Oxford, England, by Blackwell Scientific Publications for the Galton
Foundation; in English.
Philippine Population Journal, published in Manila by
the Commission on Population and the University of the
Philippines Demographic Research and Development Foundation, Inc.; in English.
Pogon sahoe nonjip (Journal of population, health and
social welfare), published in Seoul by Hanguk Ingu Pogon
Yonguwon,- in Korean with English abstracts.
Population, published in Paris by the Institut National
d'Etudes Demographiques; in French.
Population and Development Review, published in
New York by The Population Council; in English.
Population and Environment, published in New York
by Human Sciences Press; in English.
;
130
Population Index, published by the Population Association of America in cooperation with the Office of Population Research, Princeton University in Princeton, New Jersey; in English. Although mainly an index, this publication
also contains one or more articles in each issue.
Population Studies, the journal of the International
Union for the Scientific Study of Population, published in
London by London University Press; in English.
Social Biology, published in Madison, Wisconsin, by
the Society for the Study of Social Biology; in English.
Studies in Family Planning, published in New York by
The Population Council; in English.
Theoretical Population Biology, published in New York
by Academic Press; in English.
Many journals that do not focus specifically on population studies also carry articles of interest.
Notations and Formulas

Many notation systems are found in the demographic literature. Although the meaning of the symbols used is normally
clear from the context, variability in notation can result in
confusion. In this Guide we have endeavored to use a consistent system of notation based upon that of the life table. You
should be aware that in some cases our system is different
from that found elsewhere in the literature. We believe that
the consistency we have introduced makes up for this difference.
In demography, we use letters to stand for a number of
events or persons. Thus, the capital letter B is used to represent the number of births, D to represent the number of
deaths, and P to represent the number of people in the population. (The letter often, but not always, represents the first
letter of the word for the concept we are symbolizing.) Lowercase letters are also used; for example, d stands for the number of deaths in a life-table population.
Subscripts and superscripts are common in demographic
notation. Perhaps the most common is the subscript x, which
usually follows the letter it modifies and stands for exact age
at the beginning of an age interval. D refers to all deaths to
persons x to x + 1 years of age, whereas D refers to deaths to
all persons who became 18 on their last birthday (which is
the same as saying all persons of exact age 18 to exact age
19). Another common subscript is n, which often precedes
the letter it modifies and refers to the size of an age interval:
Pf refers to all females of ages x to x + n. Thus P refers to
all females of ages 20 through 24that is, exact age 20 to
exact age 25, an age interval of five years. If n = 1, the n is
often not written: D = ,D .
x
18
18
132
In this volume we use superscripts mainly to designate

gender P refers to the female population, P to the male
population. Superscripts (or subscripts) may also be used to
refer to a date or period of time: P would be the 1975 population count or estimate.
F
1975
Two other symbols need to be mentioned here. One is

k, which refers to a constant by which many demographic
measures are multiplied to make them easier to understand.
For example, the crude birth rate in a country may be 0.012
per year. Although this is an accurate demographic description, it is harder for many people to understand than the same
rate multiplied by a constant, k = 1,000. The crude birth rate
is then 12, or 12 per 1,000, and that is how it is usually expressed. Similarly, the population growth rate is often expressed as a percentage; in this case, k = 100. The other symbol is the Greek capital letter sigma, S, which is a summation sign. It is used in demography to indicate that the expression following it is to be summed. For example, the notation
1-49
In
means: Take the sum of the population at each age from 0

through 49that is, P + P, + P + . . . + P + P .
Bearing these conventions and rules in mind, we present
a list of concepts defined in this volume and the formulas
used to describe them algebraically.
0
Concept
48
A9
Formula
MORTALITY
Crude death rate (CDR)
Age-specific death rate (ASDR)

(for exact ages x to x + n)
Age-standardized death rate
(for population B with A as standard)
Concept
Formula
Infant mortality rate (IMR)
IMR = k
133
B
LIFE TABLE
Probability of dying between exact ages x and x + n
Number of deaths between exact ages x and x + n
Survivors to exact age x
Number of years lived between exact ages x and x + n
Total years lived after exact age x
Expectation of life after exact age x
RELATIONSHIPS AMONG COLUMNS OF THE LIFE TABLE
d
q = - -
Probability of dying between exact ages x and x + n
Number of deaths between exact ages x and x + n
Number of years lived between

exact ages x and x + n
d =l -
x
xtn
L.={t
2 '
(except at youngest and oldest ages)
Total number of years lived after exact age x
x=i
where refers to the

last (open) age interval
Expectation of life, or average number of years
lived after exact age x
Life-table birth rate (b) = life-table death rate (d) = S-
j
e
'
134
Concept
Formula
FERTILITY, N A T U R A L INCREASE, A N D R E P R O D U C T I O N RATES
Crude birth rate (CBR)
CBR = k
P
Crude rate of natural increase (CRNI)
General fertility rate (GFR)
GFR = - ^ - r * or
pf
-^k
pf
Age-specific fertility rate (ASFR)
F = " 'k
(for exact ages x t o x + n)

Birth-order-specific fertility rate
Total fertility rate (TFR)
Gross reproduction rate (GRR)
pi
F = " 'k
"*
p f
1
TFR = ^ F
GRR = nXF/
X
Net reproduction rate (NRR)
Child-woman ratio (CWR)

where: B = births
D = deaths
P = population
P = female population
1
x = exact age
n = size of age interval
i = order of birth, and
k = a constant
NRR = X ( F / ) |
CWR =
-^V*
or
pf
30* 15
pf
35'15
APPENDIX 2
Relationship between q and M Values

x
THE GENERAL CASE
Constructing a life table for a real population requires determining values of the q function from observed values of agex
specific death rates, which are symbolized by M in our notax
tion. The q values differ from the age-specific death rates

x
(MJ that we have discussed earlier in the following ways:

1. In the q values the denominator includes members
x
of only one (hypothetical! birth cohort, whereas in the agespecific death rate (MJ the denominator includes members
of more than one (real) birth cohort. For example, the persons in the real population age group of exact age 4 to exact
age 5 in midyear 1989 would include some persons born in
1984 and some born in 1985. Hence, the denominator for the
age-specific death rate includes parts of both the 1984 and
1985 birth cohorts.
2. Similarly, in the q values the numerator includes
x
members of only one birth cohort, whereas in the age-specific death rate (MJ the numerator includes members of more
than one birth cohort. For example, the persons who died at
age 4 in 1989 would include some persons born in 1984 and
some persons born in 1985.
3. For the denominator of the age^specific death rate
(MJ we use the midyear population as an estimate of the
number of person-years lived. The midyear population is a
biased estimate of the number of persons exposed to the risk
of dying at the beginning of the year (as opposed to exposure,
which, as we have seen, the midyear population does estimate satisfactorily in the absence of unusual conditions) to
the extent that it excludes persons who died during the first
half of the year. Further, for life table purposes it is biased
because it includes persons who migrated into the popula-
136

x
tion during the yearand migration is expressly omitted in

the life table calculations.
Age-specific death rates [MJ usually overestimate the

probabilities
of dying during a given exact age interval (<jj
because they exclude persons dying in the first half of the

year from the denominator and because they refer on the
average to persons x + Vi years old instead of x years old.
Rather complex methods have been developed for calculating q values from M values, most of which are well
x
beyond the scope of this Guide. An approximate value for q
can be found for ages over 4, however, on the assumption

that:
M
This is the most common formula for calculating values for

q (except for the first few years of life), although various
x
methods of adjusting the values of M are often used before

x
the basic formula is applied.
INFANT
MORTALITY
For the younger ages, particularly infancy (exact age 0 to ext_ age 1), the determination of q is especially problematic.
ac
Often, the infant mortality rate (IMR) is used directly as the

value of q in a life table. The defect in the infant mortality
0
rate is the same in principle as that for death rates at other

agesthat is, more than one cohort is involved in the numeratorbut it is more serious because births (the denominator of the IMR) may fluctuate rather dramatically from year
to year. It is sometimes possible to obtain a satisfactory degree of precision by averaging over several years. For example,
we might calculate the infant mortality rate for 1985-87 by
dividing the number of deaths to infants in years 1985, 1986,
1. Of course, deaths of persons who migrate into the population during the year
are also included in the numerator of death rates. If we assume that migration is
smoothly distributed over the year and that migrants experience the same ASDRs
as others, then including migrants in M does not pose a problem.
2. Forage groups larger than one year the corresponding formula is q - \2n( MJ]/
n
\2
n( MJ\.
n

x
137
and 1987 by the number of births occurring in those same

years: This provides more accuracy than the usual method
of calculating the infant mortality rate because the deaths in
the numerator are better matched with the proper set of births
in the denominator.
Another method of calculating q involves either clasQ
sifying infant deaths by year of birth or determining the number of deaths of children by age (in months) at death. With
such data it is possible to construct a rate so that the numerator and denominator both refer to the same cohort: infant deaths in year z to infants born in year z are divided by
the number of births in year z infant deaths in year z to
;
infants born in year z - 1 are divided by the number of births

in year z - 1 the two are then added together. If the requisite
;
data are not available, there are also methods for estimating
this type of rate, but we do not discuss them here. If you are
interested in them, you can refer to the more advanced sources
cited in the concluding section of Chapter 2. The more precise techniques for estimating infant mortality rates discussed
in those sources are similar to those used in the life table,
and consequently no further adjustment is required; that is,
the adjusted IMR simulates a cohort approach and can be
used directly as q . For ages 1 to 4, similar techniques are
Q
often employed, although the use of the simple formula given

above for ages over 4 is often used to convert values of M,
through M into the appropriate q values.
4
APPENDIX 3
Answers to Selected Exercises

CHAPTER
EXERCISE
i (page 10)
1. Where did the deaths occur? How many people were

exposed to the risk of dying?
2. The population size is given only for the end of the
year. This is usually an inadequate measure of the number of persons exposed to the risk of dying during the
year.
3. When did they die?
FIRST SET O F M U L T I P L E - C H O I C E
QUESTIONS
(page 18)
1. (d). Cannot tell without knowledge of the age structure.

2. (c). Cannot tell for certain without knowledge of the
age structure.
S E C O N D SET O F M U L T I P L E - C H O I C E
QUESTIONS
(page 28)
1. (a). The age-specific rates for country A are higher in

every age group.
2. (b). See Table 2.1.
3. (a).
4. (a). See Table 1.1.
EXERCISE
2 (page 29)
Case 1: The differentials in the crude death rates are not a

result of differences in age distributions. City B had about
the same average mortality levels as the United States, but
140
City A had substantially higher mortality levels than either

City B or the United States.
Case 2: City A and City B both had substantially higher
mortality levels than the United States when age differentials are taken into account, and the two cities were very
similar in their mortality levels. City B probably had a younger
age distribution than either the United States or City A, and
this accounts for its lower crude death rate.
Case 3: The mortality rates of the two cities were on the
average closely similar to that of the United States. Since
City B apparently had an age structure similar to that of the
United States, this mortality similarity to the United States
is reflected in either the crude or the standardized rate comparisons. City A must have had an older population than either City B or the United States, however, because its higher
mortality, as reflected in comparisons of the crude death rate,
disappears in the age-standardized comparisons.
Case 4: When age is taken into account and standardization
is used, it appears that City A had mortality levels like those
of the United States but City B had substantially higher mortality rates than either City A or the United States. This reverses the comparative mortality levels as measured by the
crude death rates. Therefore, it is likely that City A had an
old population, which gave it a high crude death rate despite
low age-specific mortality rates. By the same logic, City B
must have had a very young population, which gave it a relatively low crude death rate even though on the average its
age-specific death rates were high.
Case 5: Once age is controlled by standardization, it appears
that City B had slightly higher mortality levels and City A
even higher mortality levels as compared with the United
States. City B must have had a somewhat younger population than the United States, or at least one that had somewhat less concentration in higher mortality age groups, be-
141
cause initially the crude death rate was equal to that of the
United States, and standardization makes it a little higher.
City A must have had a significantly older population than
either City B or the United States because the overall mortality differential as compared with the United States or City
B is reduced (but not eliminated) when an age adjustment is
made.
Case 6: Age differentials obscure the probable mortality differentials among the three populations. The fact that City A
had lower mortality levels than either City B or the United
States (in age-standardized comparisons) must be obscured
in the crude rate comparisons by the fact that City A's age
structure must have been very different from that of the other
populations. Presumably it has an old age structure because
standardization reduces its rate by more than 50 percent,
whereas it only slightly increases the rate for City B.
EXERCISE 3
(page 30)
Rate
Country A Country B
Crude death rate

Death rate standardized on distribution for
Country A
Country B
38.25
32.25
38.25
27.25
43.25
32.25
Country B's lower crude death rate results from the fact that
a large part of its population lives in the metropolitan areas,
where death rates are relatively low. Country A initially has
a high crude death rate, despite its low mortality within each
type of region, because its population is concentrated in the
high-mortality rural areas.
FIRST SET O F T R U E / F A L S E
QUESTIONS
(page 33)
1. True. See Table 2.8.

2. False, because of the younger age structure in the developing countries.
3. False.
142
4. False. It is not a good estimate if events have occurred

unevenly throughout the year.
5. False. Death rates are highest at the extreme ages
that is, among both the very young and the old.
THIRD SET O F M U L T I P L E - C H O I C E Q U E S T I O N S
(page 47)
1. (c).
2. (b).
3. (d). The reason is that a probability has all persons at
the start of a period in the denominator, whereas a death
rate has the total number of person-years livedwhich,
in the absence of migration, must be less than those at
the start unless there is no mortality at all.
4. (b).
S E C O N D SET O F T R U E / F A L S E Q U E S T I O N S
(page 49)
1. False. Death rates at these ages are relatively high, and

therefore survival ratios are low.
2. False.
3. True.
4. True.
F O U R T H SET O F M U L T I P L E - C H O I C E Q U E S T I O N S
(page 58)
1. (f).
2. (e). The reason is that the life-table death rate =
and e = Tjt
0
tjT
= 1 (life-table death rate).
3. (d).
4. (e), not (d), because (c) is true for the reason that q dex
termines L .
x
THIRD SET OF T R U E / F A L S E QUESTIONS
1. True. See Table 2.13.

2. False. They are equal.
3. True.
(page 59)
143
4. False. Life tables refer to groups, not individuals, and

they refer to real groups (not hypothetical groups) only
if mortality rates are not changing or the life table in
question is a generational life table.
5. Arguable. Crude death rates do measure the actual rate
of mortality of the population as it is at a particular
time. They do not measure mortality independently of
the effect of the age distribution. Therefore, the standardized rates are better for comparing the underlying
population trends. The crude rates are better for measuring the rate at which the population is dying without reference to whether the age distribution has affected it.
6. True.
CHAPTER 3
FIRST SET O F T R U E / F A L S E Q U E S T I O N S
(page 78)
1. False. The division is closer to half-and-half.

2. True. At least for the countries included in Table 3.2.
3. True.
4. False. This would result in a crude rate of natural increase of 25 per 1,000, and numerous countries have
rates that high or higher. The continent of Africa as a
whole was recently estimated to have a CBR of 43 and
a CDR of 13, giving it a CRNI of 30 (Population Reference Bureau, 1992).
5. True.
FIRST SET O F M U L T I P L E - C H O I C E Q U E S T I O N S
1. (e).
2. (b).
3. (b).
4. (d).
5. (b).
6. (b).
(page 78)
144
S E C O N D SET OF MULTIPLE-CHOICE
QUESTIONS
(page 90)
1. (d). The two countries have almost identical percentages of childbearing-age women in the population and
identical ASFRs. We need to know more about the age
distribution of the women within the age group 15-44.
2. (d). We need to know more about the age distribution
of the women within the age group 15-44.
3. (c).
4. (b).
S E C O N D SET OF TRUE-FALSE Q U E S T I O N S
(page 91)
1. True.
2. True.
3. False.
4. True.
EXERCISE
i (page 108)
The exact results will depend on the particular changes made.

However, any shift that raises birth rates at younger ages and
also makes an equal reduction in birth rates at older ages
should have the effect of both (a) decreasing the length of a
generation because the average age of mothers at the birth of
their children will be less and (b) increasing the net reproduction rate because mortality will be less at younger ages.
In a population such as that of the United States, this shift
will not be of great importance since mortality is low at all
ages within the reproductive span and the net and gross reproduction rates are nearly the same.
EXERCISE
2 (page 108)
Country A: Both fertility and mortality are low because both

the net and gross reproduction rates are low, and the difference between them is small. If the current age-specific birth
145
and death rates continue indefinitely, the population size will

decline slowly, eventually stabilizing at the rate of 15 per
thousand per generation. Since the birth and death rates are
equal and the birth rate is low, it appears likely (but not certain) that the age structure is not far from that required for
this permanent condition.
Country B: The statements made about country A with respect to the net and gross reproduction rate apply here too.
However, the fact that the crude birth rate is much higher
than the crude death rate, with a substantial rate of present
natural increase, makes it probable that the age structure is
young (probably as a result of higher past fertility). Therefore, the attainment of the slow growth decline will take a
long time, even if the age-specific vital rates continue at their
present level.
Country C: The high gross reproduction rate indicates very
high fertility rates. That there is little difference between
the net and gross rates means that mortality is very low. This
is a population that will grow rapidly (195 percent per generation) if its current vital rates continue indefinitely. Unless it has an unusually old age structure, the crude birth
rate is likely to be very high and the crude death rate very
low at present. This would be characteristic of a population
like the American Hutterites.
Country D: This is a country in which fertility and mortality are both very high. The large gross reproduction rate indicates that fertility is high. The fact that the net reproduction
rate is so much lower means that mortality must be very
high. The net reproduction rate of 1,000 means that for the
long run, mortality is sufficiently high to offset completely
the high fertility. Over the long run these rates imply a stationary population. That birth and death rates are currently
equal at a high level suggests that a stationary condition is
already closely approximated.
146
Country E: Fertility rates are moderately high and mortality

rates low. This inference follows from the fact that the gross
reproduction rate is substantially above 1,000 (although there
are many higher rates) and the net reproduction rate differs
from it only a little. In the long run if the age-specific vital
rates remain at their current levels, this population will grow
at the rate of about 48 percent per generation. This is a situation rather similar to that in the United States during the
period following World War II.
Country
F: This country has high fertility and moderate
mortality because its gross reproduction rate is very high and

the net reproduction rate differs from the gross rate moderately. (If mortality were very low, the net reproduction rate
would differ very little from the gross rate. If mortality were
very high, the net reproduction rate would be 1,000 or less.
Should these vital rates continue indefinitely, the growth rate
per generation would be about 50 percent. The actual crude
birth rate and death rates given are consistent with a situation in which fertility remains high but mortality has fallen
from previously higher levels. This situation is probably representative of a country like Pakistan.
THIRD SET OF T R U E / F A L S E Q U E S T I O N S
(page 108)
1. True, although the two rates can be equal if there is no

mortality before the end of the childbearing ages.
2. False. The GRR refers to a hypothetical
real one.
3. True.
4. True.
5. False.
6. False.
7. True.
cohort, not a
THIRD
SET O F M U L T I P L E - C H O I C E Q U E S T I O N S
147
(poge 109)
1. (a). See Table 3.21.

2. (c).
3. (c).
4. (d).
5. (c)., assuming that the sex ratios at birth by age are equal.
The TFRs are certainly equal.
6. (d).
FOURTH SET O F MULTIPLE-CHOICE QUESTIONS
(page 119)
1. (c).
2. (c).
3. (e).
4. (a).
FIFTH SET O F M U L T I P L E - C H O I C E Q U E S T I O N S
1. ( H
2. (f).
3. (a).
4. (d).
(page 125)
Countries with Populations of Fewer than

1 Million: 1990 Estimates or Latest
Census
Most of the countries listed below are not included in the
tables in the Guide
because demographic measures for small
populations are subject to rather great variability. Figures with

an asterisk (*) are U N estimates.
Country
Population
Africa
British Indian Ocean Territory
Cape Verde
Comoros
Djibouti
Equatorial Guinea
Gambia
Guinea-Bissau
Reunion
Sao Tome and Principe
Seychelles
St. Helena
Swaziland
Western Sahara
3,000'
370,000
551,000'
409,000 *
348,999
.861,000'
965,000 *
599,000'
121,000'
67,000
7,000'
768,000
179,000*
North America
Anguilla
Antigua and Barbuda
Aruba
Bahamas
Barbados
Belize
Bermuda
British Virgin Islands
Cayman Islands
Dominica
Greenland
Grenada
Guadeloupe
8,000'
77,000'
60,000 *
253,000
255,000'
188,000
61,000
13,000
27,000
83,000 *
57,000'
85,000'
344,000'
150
Countries with Populations of Fewer than 1 Million
Country
Martinique
Montserrat
Netherlands Antilles
St. Kitts and Nevis
Saint Lucia
St. Perrre and Miquelon
St. Vincent and the Grenadines
Turks and Caicos Islands
U.S. Virgin Islands
Population
341,000 *
13,000*
189,000*
44,000 *
151,000*
7,000 *
116,000 *
10,000 *
117,000 '
South America
Falkland Islands
French Guiana
Guyana
Suriname
2,000 *
99,000'
796,000 *
422,000'
Asia
Bahrain
Brunei Darussalam
Cyprus
East Timor
Macau
Maldives
Qatar
503,000 *
266,000 *
702,000
737,000 *
479,999 *
215,000*
368,000 *
Europe
Andorra
Channel Islands
Faeroe Islands
Gibraltar
Holy See
Iceland
Isle of Man
Liechtenstein
Luxembourg
Malta
Monaco
San Marino
52,000
300,000'
48,000 *
31,000'
1,000 *
255,000
64,000 *
29,999 *
373,000 *
354,000
29,000 *
24,000"
Oceania
American Samoa
Cocos Islands
Cook Islands
Fiji
French Polynesia
Guam
Johnston Island
Kiribati
Marshall Islands
Federated States of Micronesia
39,000
555
18,000'
765,000 *
206,000 *
119,000'
1,007
66,000 *
40,000 *
99,000 *
Countries with Populations of Fewer than 1 Million
Country
Nauru
New Caledonia
Niue
Northern Mariana Islands
Palau
Pitcairn
Western Samoa
Solomon Islands
Tokelau
Tonga
Tuvalu
Vanuatu
Wake Island
Wallis and Futuna Islands
151
Population
10,000*
168,000*
3,000 *
26,000'
18,000*
52
164,000
321,000*
1,552
95,000 *
10,000*
147,000
1,647
18,000'
Source: United Nations Statistical Office, 1990 Demographic Yearbook (1992, table 3).
Bachu, Amara. 1991. Fertility of American women: june, 1990. Current

Population Reports, Series P-20, No. 454. Washington, D.C.: U.S. Dept.
of Commerce, Economics and Statistics Administration, Bureau of the
Census.
Barclay, George W. 1958. Techniques of population analysis. New York:
John Wiley and Sons.
Bogue, Donald J. 1969. Principles of demography. New York: John Wiley
and Sons.
Bogue, Donald J., Eduardo E. Arriaga, Douglas L. Anderton, and George
W. Rumsey, eds. 1993. Readings in population research methodology. 8
vols. Chicago: Social Development Center, University of Chicago
(published for the United Nations Population Fund).
Bogue, Donald J., and Evelyn Kitagawa. n.d. Manual of demographic
research techniques. Unpublished manuscript, Community and Family
Study Center, University of Chicago.
Bogue, Donald J., and James A. Palmore. 1964. Some empirical and
analytic relations among demographic fertility measures, with regression
models for fertility estimation. Demography 1(1): 316-38.
Bumpass, Larry L., Ronald R. Rindfuss, James A. Palmore, Mercedes
Concepcibn, and Byoung Mohk Choi. 1982. Intermediate variables and
educational differentials in fertility in Korea and the Philippines.
Demography 19(2): 241-60.
Carr-Saunders, A. M . 1936. World population. Fair Lawn, New Jersey:
Oxford University Press.
Cho, Lee-Jay, Robert D. Retherford, and Minja Kim Choe. 1986. The
own-children method of fertility estimation. Honolulu: East-West
Center.
Das Gupta, Prithwis. 1993. Standardization and decomposition of rates:
A user's manual. U.S. Bureau of the Census, Current Population Reports,
Series P23, No. 186. Washington, D.C.: U.S. Government Printing Office.
Dublin, L. I., A. J. Lotka, and M. Spiegelman. 1949. Length of life. Rev.
ed. New York: Ronald Press.
Durand, John D. 1968. The modern expansion of world population. In
Charles B. Nam, ed., Population and society: A textbook of readings, pp.
108-20. Boston: Houghton-Mifflin.
Easterlin, Richard, ed. 1980. Population and economic change in
developing countries. Chicago: University of Chicago Press.
154
References
Eaton, Joseph W., and Albert J. Mayer. 1954. Man's capacity to

reproduce: The demography of a unique population. Glencoe, Illinois:
The Free Press.
Freedman, Ronald, and Axjun L. Adlakha. 1968. Recent fertility declines
in Hong Kong: The role of the changing age structure. Population Studies
22(2): 181-98.
Greville, Thomas N . E. 1946. United States life tables and actuarial
tables, 1939-1941. Washington, D.C.: U.S. Bureau of the Census.
Grove, Robert D., and Alice M . Hetzel. 1968. Vital statistics rates in the
United States 1940-1960. Washington, D.C.: U.S. Government Printing
Office.
Gunasekaran, Subbiah, and James A. Palmore. 1984. Regression
estimates of the gross reproduction rate using moments of the female age
distribution. Asian and Pacific Census Forum 10(4): 5-10.
Hauser, Philip M . 1960. Population perspectives. New Brunswick, New
Jersey: Rutgers University Press.
Hirschman, Charles. 1986. The recent rise in Malay fertility: A new
trend or a temporary lull in a fertility transition! Demography 23(2):
161-84.
Hollman, Frederick W.. 1990. United States population estimates, by
age, sex, race, and Hispanic origin: 1980-1988. Current Population
Reports, Series P-25, No. 1045. Washington, D.C.: U.S. Dept. of
Commerce, Bureau of the Census.
Indonesia. Central Bureau of Statistics. 1993. Summary of the 1990
population census results. (In Indonesian.) Jakarta.
International Union for the Scientific Study of Population (IUSSP). 1982.
Multilingual demographic dictionary. 2d ed., English section. Liege,
Belgium: Ordina Editions.
Jaffe, A. J. 1951. Handbook of statistical methods for demographers.
Washington, D C : U.S. Bureau of the Census.
Keyfitz, Nathan. 1968. Introduction to the mathematics
Reading, Massachusetts: Addison-Wesley.
of population.
Keyfitz, Nathan, and Wilhelm Flieger. 1968. World population: An

analysis of vital data. Chicago: University of Chicago Press.
. 1971. Population: Facts and methods of demography. San
Francisco: W. H. Freeman and Co.
. 1990. World population growth and aging: Demographic trends
in the late twentieth century. Chicago: University of Chicago Press.
Kim, Tai-Hun. 1990. Mortality transition in Korea: 1960-1980. Seoul:
Population and Development Studies Center, Seoul National University.
Marty, Robert, and David J. Neebe. 1966. Compound interest tables: For
long-term planning in forestry. U.S. Department of Agriculture Forest
References
155
Service, Agriculture Handbook No. 311. Washington, D.C.: U.S.

Government Printing Office.
Natarajan, K. S., and Phool Singh. 1988. Fertility in India: An analysis of
1981 census data. Occasional Paper No. 13 of 1988. New Delhi:
Demography Division, Office of the Registrar General, India.
Ogawa, Naohiro, and Robert D. Retherford. 1993. The resumption of
fertility decline in Japan: 1973-92. Population and Development Review
19(4): 703-41.
Office of Population Research, Princeton University. 1950. Population
Index 16(2), April.
. 1968. Population /ndex34(2), April-June.
. 1979. Population Index 45 (2), April.
. 1981. Population Index 47(2), Summer.
Pagnini, D. L., and Ronald R. Rindfuss. 1993. The divorce of marriage and
childbearing: Changing attitudes and behavior in the United States.
Population and Development Review 19(2): 331-47.
Palmore, James A. 1978. Regression estimates of changes in fertility,
1955-1960 to 1965-1975, for most major nations and territories. Papers
of the East-West Population Institute, No. 58. Honolulu: East-West
Center.
Peng Xizhe. 1993. Regional differentials in China's fertility transition. In
Richard Leete and Iqbal Alam, eds , Tie revolution in Asian fertility:
Dimensions, causes, and implications. Singapore: Oxford University
Press.
Petersen, William, and Renee Petersen. 1985. Dictionary of demography:
Multilingual glossary. Westport, Connecticut: Greenwood Press.
Population Reference Bureau. 1992. J 992 world population data sheet.
Washington, D.C.: Population Reference Bureau.
Pressat, Roland. 1972. Demographic analysis: Methods,
applications. New York: Aldine-Atherton.
results,
Preston, Samuel H. 1976. Mortality patterns in national patterns. New

York: Academic Press.
Preston, Samuel H , ed. 1978. The effects of infant and child
on fecundity. New York: Academic Press.
mortality
Preston, Samuel H., and Michael R. Haines. 1991. Fatal years: Child
mortality in late nineteenth-century America. Princeton, New Jersey:
Princeton University Press.
Preston, Samuel H , Nathan Keyfitz, and Robert Schoen. 1972. Causes of
death: Life tables for national populations. New York: Seminar Press.
Rao, N . Rama, Robert D. Retherford, P. D. Joshi, and Norman Y. Luther.
1993. Census-based estimates of fertility by rural-urban residence and
156
References
religion: Gujarat, Haryana, Himachal Pradesh, fammu and Kashmir,

and Uttar Pradesh. Occasional Paper No. 3 of 1993. New Delhi: Office of
the Registrar General and Census Commissioner.
Rele, J. R. 1967. Fertility analysis through extension of stable population
concepts. Ph.D. dissertation, Institute of International Studies,
University of California, Berkeley.
. 1987. Fertility levels and trends in India, 1951-1981. Population
and Development Review 13(3): 513-30.
Retherford, Robert D., and J. R. Rele. 1989. A decomposition of recent
fertility changes in South Asia. Population and Development Review
15(4): 739^t7.
Rindfuss, Ronald R., and J. A. Jones. 1991. One parent or two? The
intertwining of American marriage and fertility patterns. Sociological
Forum 6(2): 311-26.
Rindfuss, Ronald R., James A. Palmore, and Larry L. Bumpass. 1982.
Selectivity and the analysis of birth intervals using survey data. Asian
and Pacific Census Forum 8(3): 5ff.
Rindfuss, Ronald R., and Alan M . Parnell. 1989. The varying connection
between marital status and childbearing in the United States. Population
and Development Review 15(3): 447-70.
Rodriguez, German, and John N . Hobcraft. 1980. Illustrative analysis:
Life table analysis of birth intervals in Colombia. World Fertility Survey
Scientific Reports, No. 16. London: World Fertility Survey.
Schoen, R. 1983. Measuring the tightness of a marriage squeeze.
Demography 20(11: 61-78.
Shryock, Henry, Jacob S. Siegel, and Associates. 1971. The methods and
materials of demography. 2 vols. Washington, D.C.: U.S. Bureau of the
Census. Condensed edition available from Academic Press, New York,
1978.
Spiegelman, Mortimer. 1968. Introduction to demography. Rev. ed.
Cambridge, Massachusetts: Harvard University Press.
Stycos, J. Mayone, and Kurt W. Back. 1968. The control of human
fertility in Jamaica. Ithaca, New York: Cornell University Press.
Swamy, V. S., A. K. Saxena, James A. Palmore, Vinod Mishra, J. R. Rele,
and Norman Y. Luther. 1992. Evaluating the sample registration system
using indirect estimates of fertility and mortality. Occasional Paper No.
3 of 1992. Delhi: Office of the Registrar General and Census
Commissioner.
Thompson, Warren S., and David T. Lewis. 1965. Population
5th ed. New York: McGraw-Hill.
problems.
United Nations. 1986a. Consequences of mortality trends and

differentials. Population Studies, 95. New York: Department of
International Economic and Social Affairs, United Nations.
References
157
United Nations. 1986b. Determinants of mortality change and

differentials in developing countries. Population Studies, 94. New York:
Department of International Economic and Social Affairs, United
Nations.
United Nations. 1993. World population prospects: The 1992 revision.
New York: Department for Economic and Social Information and Policy
Analysis, United Nations.
United Nations Population Branch. 1958. Multilingual demographic
dictionary. English Section. Population Studies, No. 29. New York:
Department of Economic and Social Affairs, United Nations.
United Nations Statistical Office. 1949-. Demographic yearbook.
Annual. New York: Department of Economic and Social Affairs, United
Nations.
. 1955. Principles for a vital statistics system. Statistical Papers,
Series M , No. 19. New York: United Nations.
. 1990. World urbanization prospects, 1990. New York: United
Nations.
. 1992. World population prospects: The 1992 revision. Annex
tables. New York: United Nations.
United States. Bureau of the Census. 1966. Fertility of the population:
June 1964 and March 1962. Current Population Reports, Series P-20, No.
147. Washington, D.C.: U.S. Government Printing Office.
. 1973. Women by number of children ever born: 1970 census of
population. Subject Report, Final Report PC(2)-3A. Washington, D.C.:
U.S. Dept. of Commerce, Bureau of the Census.
. 1979. Fertility of American women: fune 1978. Current
of Commerce, Bureau of the Census.
. 1982. Fertility of American women: fune, 1980. Current
of Commerce, Bureau of the Census.
United States. National Center for Health Statistics. 1969. Vital
statistics of the United States. Vol. 1. Washington, D.C.: U.S.
Government Printing Office.
. 1978a. Vital statistics of the United States, 1975. Vol. 1.
Washington, D.C.: U.S. Government Printing Office.
. 1978b. Vital statistics of the United States, 1976. Vol. 2, section
5. Washington, D.C.: U.S. Government Printing Office.
. 1985. U.S. decennial life tables for 1979-81. Vol. 1, No. 1, United
States life tables. Hyattsville, Maryland.
. 1990. Vital statistics of the United States, 1988. Vol. 1, Natality.
Hyattsville, Maryland.
158
References
. 1991. Vital statistics of the United States, 1988. Vol. 2,

Mortality, part A. Hyattsville, Maryland.
Weeks, John R. 1992. Population: An introduction to concepts and
issues. 5th ed. Belmont, California: Wadsworth.
Wolfenden, Hugh H. 1954. Population statistics and their
Rev. ed. Chicago: University of Chicago Press.
compilation.
INDEX
[Note: page numbers in boldface type indicate the location of figures and
tables.)
age
at birth [see mean age at childbirth)
exact, 37, 40-41
at last birthday, 41
age at marriage
effect on fertility measures, 119
age composition
effect of fertility and mortality, 5 In, 51-52
effect on crude birth rate, 81-82, 84-89 (see also standardization of
birth rates)
effect on crude death rate, 15-18, 16, 17, 52-53 (see also
standardization of death rates)
effect on crude rates, exercises on, 91
effect on fertility measures, exercises on, 110
effect on general fertility rate, 73-74, 89
in life table (see life table, functions, L )
rectangular, 92
of stationary population, 51-52, 5 2
and voting, 20-21
age distribution (see age composition)
age-specific birth rates [see age-specific fertility rates)
age-specific death rates (ASDRs, M ), 10-15, 12
as component of crude death rate, 15-18
correlation among, 22
exercises on, 18
formula for, 12, 132
patterns of, 12
and probability of death, 135-36
for selected countries, 12, 13-14
age-specific fertility rates (ASFRs, F ), 74-77
correlations among, 94
correlation with total fertility rate, 94
exercises on, 78-79, 90
formula for, 74, .134
of Hutterites, 75-76, 77
pattern of, 74-75, 75
and cohort-period comparisons, 119
and growth rate, 107
for selected countries, 76
world distribution by level, 76-77, 77
ASDR [see age-specific death rates)
ASFR (see age-specific fertility rates)
average age at childbirth (see mean age at childbirth)
r
160
Index
birth
age at (see mean age at childbirth)
live, defined, 64
birth cohort
defined, 35
hypothetical (see hypothetical cohort)
synthetic (see hypothetical cohort)
birth intervals, analysis of, 120-25
life-table approach, 121-24, 122-23
open intervals, 120-21
exercise on, 125-28
problems with
censoring, 120-24
exercises on, 125
data quality, 120
selectivity, 124-25, 126-27
exercises on, 125, 128
birth-order-specific fertility rates, 82-84, 83
and general fertility rate, 82-84, 83
birth probabilities
and contraceptive use, 121-24, 122-23
birth rates (see also fertility rates)
age-specific (see age-specific fertility rates)
birth-order-specific (see birth-order-specific fertility rates)
crude (see crude birth rate)
intrinsic, for selected countries, 102-3
in life table, 50
formula for, 133
marital-status specific, 80-82
specific, 79-84
standardized (see standardization of birth rates)
births
smooth distribution assumption, 5
timing of, and fertility measures, 118
causes of death
in infancy, 31-32
CBR (see crude birth rate)
CDR (see crude death rate)
censoring
of birth intervals, 120-24
defined, 120
exercise on, 125
census measures of fertility, 111-15
central population, 5
childbearing
average age at (see mean age at childbirth)
risk of, 65
childlessness in United States, 113
children ever born (CEB), 65, 114-15
in Indonesia, 115
Index
161
in United States, 113, 114

child-woman ratio (CWR), 112-14
correlations with crude birth rate and total fertility rate, 113
exercises on, 120
formula for, 2, 112, 134
problems with, 112-13
range of, 113
closed population (see under migration assumptions)
cohort
birth (see birth cohort)
fertility of, 115-19 (see also children ever born)
exercises on, 119-20
and period fertility, 117-19
hypothetical (see hypothetical cohort)
life tables for, 35-36
marriage, 116
synthetic (see hypothetical cohort)
cohort effects on fertility, 116-17
composition
age (see age composition)
controlling for (see standardization)
and crude death rate, 19
confounding factors (see standardization)
constants, use of
in demographic measures, 9n, 132
in gross reproduction rate, 95n
in net reproduction rate, 97, lOln
in percentages, 3
in total fertility rate, 92n
contraceptive use and birth probabilities, 121-24, 122-23
crude birth rate (CBR), 50n, 66
correlations
with child-woman ratio, 113
with general fertility rate, 93-94
with total fertility rate, 94
for selected countries, 66, 67, 68, 86, 102-3
standardized (see standardization of birth rates)
for world and regions, 98
crude death rate (CDR), 9-10
components of, 15-18, 16, 17 (see also standardization of death rates)
decomposition of, 15-18
effects of age composition, 15-18, 16, 17, 52-53 (see aiso standardization of death rates)
effects of compositional factors, 19
exercises on, 18-19
formula for, 9, 15, 132
and life-table death rate, 52-53, 53
for selected countries, 10, 11, 102-3
standardized (see standardization of death rates)
162
Index
crude rate of natural increase (CRNI), 66, 69-72 (see also natural
increase)
exercises on, 78-79
historical levels, 70
for selected regions, 66, 69, 69
time needed for population to double, triple, and quadruple, 69-70, 70
cumulative fertility (see children ever born)
d (deaths in life table) (see under life table, functions)
data quality, 15n, 32-33, 64, 120
death
causes of, in infancy, 31-32
risk of, 3, 10, 31, 65, 135
death rates
age-specific (see age-specific death rates)
crude [see crude death rate)
for infants [see infant mortality rate)
intrinsic, for selected countries, 102-3
in life table (see life table, death rate)
standardized (see standardization of death rates)
deaths
in life table, 43
smooth distribution assumption, 5, 37, 43, 44-45
decomposition, 85n
of crude death rate, 15-18
denominator (see also midyear population; person-years lived; see also
specific measures)
of age-specific death rates, 135-36
from census data, 111
of child-woman ratio, 112
of general fertility rate, 73
ideal, 3, 7, 66, 135
of infant mortality rate, 31, 136-37
of probabilities, 7
of proportions, 2
of q (probability of dying), 135-36
of rates, 3-6
density, formula for, 2
dependency ratio, formula for, 2
doubling time, 69-70, 70, 72
exercise on, 79
x
e (see life expectancy)

events
basic elements of, 1
exercise on, 10
rate of occurrence (see rates)
smooth distribution assumption, 5
expectation of life (see life expectancy)
exposure to risk, 3-6, 7 (see also denominator; person-years lived)
x
Index
163
of childbearing, 65
of dying, 3, 10, 31, 65, 135
F (see age-specific fertility rates)
factors, confounding (see standardization)
fecundity
contrasted with fertility, 63-64
fertility, 63-95, 111-25
age range of, 73
of cohort (see cohort, fertility of)
cohort effects on, 116-19
cumulative (see children ever born)
effect on age composition, 51n, 51-52
factors affecting, 73
and fecundity contrasted, 63-64
measurement problems, 64-65
measures of
based on census, 111-15 (see also children ever born child-woman
ratio)
correlations among, 93-94
data sources for, 111, 11 In
estimation techniques, 113-14
indirect (see child-woman ratio)
period, 115 (see also age-specific fertility rates; crude birth rate;
general fertility rate)
and timing of births, 118-19
period effects on, 116-19
and population growth, 63
fertility decline
age pattern of, and growth, 107
components of, 81-82
fertility rates (see also birth rates; crude birth rate; general fertility rate;
total fertility rate)
age-specific (see age-specific fertility rates)
birth-order-specific, 82-84, 83
marital-status-specific, 80-82
specific, 79-84
formulas for demographic measures, 132-34 (see also specific measures)
t
general fertility rate (GFR), 73-74

and birth-order-specific fertility rates, 82-84, 83
correlation with crude birth rate, 93-94
effects of age composition, 73-74, 89
range of, 74
standardized, 87-89, 88, 89
generation, length of [see mean length of generation)
GFR (see general fertility rate)
gross reproduction rate (GRR), 95-97
calculation of, 96, 100
164
Index
exercises on, 108-10

interpretation of, 96-97
for selected countries, 102-4, 104
for United States, 105
for world and regions, 97, 98
world distribution by level, 99
growth (see population growth)
Hutterites
fertility rates of, 75-76, 77
hypothetical cohort
defined, 36
in fertility analysis (see gross reproduction rate; net reproduction rate,total fertility rate)
in life table, 36-37 (see also life table)
hypothetical population, 15-16, 49-51 (see also stationary population)
IMR (see infant mortality rate|
incidence rates (see rates)
income composition
and voting, 20-21
infant, defined, 30
infant death rate, 30n
infant mortality rate (IMR), 30-34, 136-37
by age, 3 In, 31-32
exercises on, 33
levels of, 33, 34
measurement problems, 32-33, 136-37
for selected countries, 33, 34
intermediate variables, 120
intrinsic rates
t (survivors in life table) (see under life table, functions)
L (person-years lived in life table) (see under life table, functions)
length of generation (see mean length of generation)
life expectancy (ej, 46
change over time, 57, 58
comparing, 56-57
problems in, 56
differentials in, 56, 57, 58
interpretation, 56
range of, 56, 57
for selected countries, 102-3
life table, 35-57
abridged, 46-47
definition of, 46
t
Index
165
definition of, 46
example of, 48
applications of (see life expectancy, comparing; stationary population;
survival ratios)
and birth interval analysis, 121-24
birth rate, 50
formula for, 133
cohort, 36, 115
complete, 37-46
definition of, 46
example of, 38-40
cross-sectional (see life table, period)
current (see life table, period)
death rate, 50
compared with crude death rate, 52^53, 53
formula for, 133
deaths [d ), 43
x

exercises on, 47-49, 58-59
functions
d (deaths), 43
e , 46 (see also life expectancy)
L (person-years lived), 43-45
formulas for, 43, 44, 133
at youngest ages, 44
l (survivors), 41-42
formula for, 42
q (probability of dying), 41 (see also probability of dying)
formula for, 133
T (total person-years lived), 45
generational, 36, 115
interpretations of, 37, 51
life expectancy (see life expectancy)
longitudinal, 36, 115
migration assumption, 37, 49, 56
period, 36-37, 115
basic assumption, 36
defined, 36
person-years lived in [L T ), 43-45
formulas for, 43-45, 133
probability of dying (see probability of dying)
radix, 37, 37n, 41
defined, 37
survivors [( ), 41-42
time-specific (see life table, period)
x
M [see age-specific death rates)

marital status composition, 79-82
effect on fertility, 81
x
166
Index
percent ever married in United States, 113

marital-status-specific fertility rates, 80-82
marriage
norms regarding, 80
variations in, 79-80
marriage cohort, 116
marriage squeeze, 80
mean age at childbirth (u), 102-3, 106n (see also mean length of
generation)
mean length of generation (7), 102-3, 105-7
calculation of, 100
effect on growth, 106-7
factors affecting, 106
and mean age at childbirth, 106n
and migration, 106
and net reproduction rate, 107
midyear population (see also denominator; see also specific measures)
in age-specific death rates, 135-36
as approximation for person-years lived, 5-6, 7, 30, 66, 135
as estimate of exposure to risk, 135
in general fertility rate, 73
and migration, 135-36, 136n
in stationary population, 54
migration
and estimating person-years lived, 5-6
migration assumptions
closed population
and generation length, 106
and life table, 37, 49, 56
and midyear population, 135-36, 136n
and population growth, 69, 105
and probability of dying, 7
and projections, 54
and stationary population, 51
and survival ratios, 55
smooth distribution, 5
models, 37 (see also gross reproduction rate; life table; net reproduction
rate; total fertility rate)
mortality, 9-61 (see also age-specific death rates; crude death rate death
rates; life table; standardization of death rates)
differentials, 56, 57, 58
effect on age composition, 51, 51n
exercises on, 10
and population growth, 9
u (mu) (see mean age at childbirth)
;
natural increase (NI), 66, 69, 69-72

defined, 63
intrinsic rate of, for selected countries, 102-3
for world and regions, 66, 69, 69
net reproduction rate (NRR), 97-98, 100-105
Index
calculation of, 98, 100, 100

effects of fertility and mortality, 101
formula for, 100-101, 134
implications, 101, 105
interpretation of, 97-98
range of, 101
for United States, 105
notation, 131-32, 132-34
in life table, 41, 133
occupational composition
order-specific fertility rates (see birth-order-specific fertility rates)
percentages, 3
called rates, 6-7
period effect on fertility, 116-19
period fertility
and cohort fertility, 117-18
person-years lived, 3-6, 4
for ages under one, 30, 44
approximations for, 4-6
calculation of, 4
exercises on, 6
ideal denominator for rates, 3, 7, 66, 135
in life table (L TJ, 43-45
formulas for, 43-45, 133
population
central, 5
closed (see under migration assumptions)
hypothetical, 15-16, 49-51 (see also hypothetical cohort; stationary
population)
midyear (see midyear population)
rectangular, 92
standard (see standard population)
stationary (see stationary population)
population density
formula for, 2
population explosion, 72
population growth (see also crude rate of natural increase; natural
increase)
exponential, 72
and fertility, 63, 107
and generation length, 106-7
geometric, 70
and migration, 69, 105
and mortality, 9
rates, for world, 70, 72
population projections, 53-54, 54n, 72
j7
167
168
Index
population size
of countries with fewer than one million people, 149-51
of selected regions, 1650-1990, 71
probabilities, 7
probability of dying [q ], 41
derived from age-specific death rates, 135-36, 136n
formula for, 133
and IMR, 31, 31n, 136-37
and migration, 7
projections of population, 53-54, 54n, 72
proportions, 2-3
x
q [see probability of dying)

quadrupling time, 69-70, 70
x
radix (of life table), 37, 37n, 41

defined, 37
rates, 3-7
age-specific [see age-specific death rates; age-specific fertility rates)
birth [see birth rates,- fertility rates)
central, 5
death [see death rates)
denominators for, 3-7 [see also denominator)
fertility [see fertility rates)
infant mortality rate (IMR) [see infant mortality rate)
literacy, 6
percentages called rates, 6-7
reproduction [see gross reproduction rate; net reproduction rate)
specific, 10, 12, 73 [see also age-specific death rates; age-specific
fertility rates,- birth-order-specific fertility rates; maritalstatus-specific fertility rates)
standardized [see standardization)
survival, 7n
ratios, 1-3
examples and formulas, 2
survival [see survival ratios)
rectangular population, 92
registration system and coverage of births, 64, 111, 113
reproduction rates [see gross reproduction rate; net reproduction rate)
residence composition
risk [see exposure to risk)
SDR (standardized death rate) [see standardization of death rates)
selectivity
and analysis of birth intervals, 124-25, 126-27
sex composition
and fertility measures, exercises on, 110
sex ratio, formula for, 2
Index
169
standardization
standardization of birth rates, 84-89 (see also gross reproduction rate;
net reproduction rate; total fertility rate)
for age and sex, 84-87
effects on comparisons, 85
exercises on, 91
exercises on, 90-91
for selected countries, 85, 86
for United States, 85, 87, 87
standardization of death rates, 19-30
for age, example of, 21-22, 23
common factors controlled for, 19
exercises on, 28-30
for selected countries, 24-28, 26-27
standardization of general fertility rate, 87-89, 88, 89
standard population, 20
choice and effects of, 21, 22-24, 28, 85, 89
stationary population, 49-53 (see also life table)
age structure of, 51-52, 52
defined, 51
exercises on, 58-59
and migration, 51
surveys
and censoring, 120-24
completeness of, 33, 64
and selectivity, 124
survival rates, 7n
survival ratios, 7n, 53-55
change in, over time, 55, 55
exercises on, 59
formula for, 54
reverse, 54
survivors in life table {l ), 41-42
synthetic cohort (see hypothetical cohort)
x
T (length of generation) (see mean length of generation)

T (total person-years lived in life table) (see under life table, functions)
total fertility rate (TFR), 91-95
calculation of, 93
constant used in calculation of, 92n
correlations
with ASFRs, 94
with child-woman ratio, 113
with crude birth rate, 94
defined, 91
exercises on, 108
formula for, 92, 118n, 134
interpretation of, 92, 95
for periods and cohorts contrasted, 117-18
range of, 93
tripling time, 69-70, 70
x
Measuring
Mortality,
Fertility, and Natural
A Self-Teaching Guide to Elementary
Increase
Measures
FIFTH E D I T I O N
James A . Palmore and Robert W. Gardner
This newly revised edition of Palmore and Gardner's
popular introductory textbook presents elementary
measures used in demographic analysis, beginning w i t h
rates, ratios, percentages, and probabilities and proceeding
to the crude death rate and age-specific death rates,
standardized rates, the infant mortality rate, the life table,
the crude birth rate and age-specific fertility rates, the
general fertility rate, total fertility rate, gross and net
reproduction rates, period and cohort fertility measures,
and the analysis of birth intervals. Written i n a direct,
conversational style, it includes numerous examples and
illustrations that have been updated with data from the
1990 round of censuses. A t the end of each section are
exercises and quizzes designed to test students' understanding of the material presented. Four appendixes and
recommendations for further reading provide readers w i t h
additional useful information. Includes an index.
EAST-WEST
CENTER
ISBN 0-86638-165-1
$15.00

MeasuringMortalityFertilityAndNaturalIncrease5thEdition1994 (Pdfa)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

MeasuringMortalityFertilityAndNaturalIncrease5thEdition1994 (Pdfa)

Uploaded by

Copyright:

Available Formats

Measuring Mortality,

T h e U . S . Congress established the East-West C e n t e r i n 1960

Measuring Mortality, Fertility, and

Measuring Mortality, Fertility, and

]ames A . Palmore and Robert W. Gardner

fames A. Palmore was a senior fellow in the East-West Center's Program

Robert W. Gardner, who lives in Maine, teaches demographic methods at

States (with Herbert

Barringer and Michael Levin) and Migration Decision Making (edited

L i s t of figures and tables

1. Rates, ratios, percentages, and p r o b a b i l i t i e s

3. F e r t i l i t y , n a t u r a l increase, and r e p r o d u c t i o n rates

2. R e l a t i o n s h i p between q and M values

A g e - s p e c i f i c death rates: G u a t e m a l a , 1985, and

H i g h e s t and l o w e s t crude death rates, by region:

H i g h e s t and l o w e s t age-specific death rates, b y sex:

A g e - s p e c i f i c and crude death rates f o r three h y p o thetical populations

A g e - s p e c i f i c death rates and p o p u l a t i o n s f o r M a i n e

A g e s t a n d a r d i z a t i o n of crude death rates: Japan

F o r m u l a s for direct age-standardization of the c r u d e

Standardized death rates for selected places and

Infant m o r t a l i t y rates for selected c o u n t r i e s , by age

List of Figures and Tables

H i g h e s t and l o w e s t i n f a n t m o r t a l i t y rates, by region:

2.10. C o m p l e t e l i f e table f o r females: U n i t e d States,

2.11. A b r i d g e d l i f e table for females: U n i t e d States,

2.12. A g e c o m p o s i t i o n of the stationary p o p u l a t i o n a n d

2.13. C r u d e death rates and life-table death rates f o r U . S .

2.14. S u r v i v o r s to exact age x(ej of 100,000 m a l e i n f a n t s

2.15. E x a m p l e s of h i g h and l o w values of l i f e e x p e c t a n c y

2.16. L i f e expectancies at selected exact ages f o r m a l e s and

2.17. L i f e expectancies for w h i t e males and females

H i g h e s t and l o w e s t crude b i r t h rates, by region:

D i s t r i b u t i o n of countries b y l e v e l of crude b i r t h rate:

Average crude rates of n a t u r a l increase, by region:

List of Figures and Tables

L o w e s t and highest age-specific f e r t i l i t y rates per

A g e - s p e c i f i c f e r t i l i t y rates per 1,000 w o m e n :

B i r t h rates by l i v e b i r t h order and percentage change

3.10. C r u d e b i r t h rates and d i r e c t l y standardized b i r t h

3.11. C r u d e b i r t h rates and d i r e c t l y standardized b i r t h

3.12. G e n e r a l f e r t i l i t y rates f o r selected countries, agestandardized by the direct m e t h o d : late 1980s

3.13. O b s e r v e d general f e r t i l i t y rates and age-standardized

3.14. C a l c u l a t i o n of total f e r t i l i t y rates f o r the U n i t e d

3.15. C a l c u l a t i o n of the gross r e p r o d u c t i o n rates f o r C o s t a

3.16. E s t i m a t e d crude b i r t h rates and gross r e p r o d u c t i o n

3.17. D i s t r i b u t i o n of countries b y w o r l d region and l e v e l of

List of Figures and Tables

3.18. C a l c u l a t i o n of the gross and net r e p r o d u c t i o n rates

3.19. Measures of reproduction: selected countries, recent

3.20. G r o s s and net r e p r o d u c t i o n rates: Europe, G r e a t

3.21. G r o s s and net r e p r o d u c t i o n rates, by color: U n i t e d

3.22. Percentage ever m a r r i e d and n u m b e r of c h i l d r e n ever

3.23. N u m b e r of c h i l d r e n ever b o r n per 1,000 w o m e n and

3.24. Average n u m b e r of c h i l d r e n ever b o r n per w o m a n

3.26. Year i n w h i c h any b i r t h i n t e r v a l had to begin, g i v e n a

R o n a l d F r e e d m a n f i r s t suggested the idea of w r i t i n g a series

Self-Teaching Guide to Elementary

Measures and Measur-

ing Fertility and Natural Increase: A Self-Teaching Guide to

In 1971 revised drafts of the m a n u a l s

were p u b l i s h e d as Papers of the East-West P o p u l a t i o n Institute, N o s . 15 and 16. Several m o r e r e v i s i o n s f o l l o w e d .

it has c o m e to be called, has gone t h r o u g h seven p r i n t i n g s