What Is Demographic Analysis?

Demography is the study of population change in human societies, of the life cycle events that generate this change, and of the various factors and circumstances that influence these events.
How long do we live? How many children do we have? How do demographic conditions today compare with those of our ancestors? How does our society compare with the rest of the world? What explains differences between societies and change over time? How do life span and family size influence population growth? How fast is the population of the world growing? What are the prospects for future growth? Why are human populations growing older and what are the social and economic consequences of this aging?
Demographic analysis is concerned with the best way to frame and answer such questions. It is the "how to" of demography. Suppose we wish to answer the question "How long did people live in 19th century America?" What sort of facts must be at our disposal? What sort of answers do we expect? How do we get from the facts to the answer? If we want to know further how length of life in America today compares with life in the 19th century, how do we proceed? How then to explain the difference? And, perhaps, generate a prognosis for length of life in the 21st century? Demographic analysis does not tell us everything we need to know to answer all such questions, but it tells us many things we must know to answer them successfully.
Demographic information takes many forms, from fossilized remains to birth and death dates on tomb stones to modern censuses, surveys and vital and population registration systems. The accomplished demographer learns how to make the most of any source and how to integrate information from all available sources. The key to doing this is to learn population concepts in their pure form, rather than in the manifestations they take when confronted with particular sources of information. One may then be as comfortable dealing with the historian's documentary sources or the anthropologist's field notes as with the modern censuses and surveys that are the demographers stock in trade.
As with any discipline, the substance of demographic analysis lies in details that can be appreciated only as they are mastered, one step at a time. There are however three broad themes that pervade and inform the whole.

Populations The raw materials of demography are records of the persons who comprise particular populations, records that tell of the various demographic events these persons experienced and the various characteristics that described them at different times in their lives. When assembling this raw material, we are concerned with particular persons. Demography is not about persons as such, however, but about populations and about persons as members of a population. Populations have an existence and identity in time beyond the existence and identity of their members. Persons are born, live out their lives, and die. Populations endure, often for scores or hundreds of generations.
Demography focuses in the first instance on populations of persons comprising some human society, but we quickly learn to think more generally of populations as groups of persons or other entities whose membership changes with time. In studies of marriage and the family we consider populations of married couples, populations of single parents, and populations of families with a given number of children. The ideas we learn for the purpose of studying how long people live apply equally well to studying the ages and which men and women marry, how long marriages last, and how long couples wait before having a first or another birth. This generality is perhaps the most important single concept of modern demographic analysis.
Exposure to Risk Demography studies numbers of births, deaths and other events, but these numbers alone rarely get us very far. That there are more deaths in China than in Monaco says nothing of the comparative mortality conditions in the two countries: there are more deaths in China because there are more people in China. To compare mortality conditions we must relate numbers of deaths to the numbers of persons ‘at risk' of death. More generally, the number of occurrences of an event of any given class should be related both to the number of persons who might have experienced that event and to the length of time over which they might have experience the event. This is the principle of exposure to risk.
Period and Cohort Different demographic questions point to different ways of looking at demographic data. To study population growth we organize data by time periods. How big was the population at the beginning of a given period? How many persons were added during the period? What was the balance between birth, death and migration? Answers to such questions are period statistics. The most familiar demographic indicators, such as the population growth rate and crude birth and death rates, are period statistics.
If we are interested in how long people live, or in how many children families have, we will prefer in the first instance to work with life histories, records of individuals showing time of birth and death and other information. Given a suitable assemblage of life histories, it is a simple matter to calculate average length of life for all persons who are born during some time period or average number of children born to women who themselves were born or marry during some time period. These are examples of cohort statistics, so called because they refer to the experience of a group of persons defined by having experienced a particular event during a particular time period.
Period and cohort statistics are both natural and important ways of looking at demographic phenomena. Some questions call for period statistics, other questions for cohort statistics, and still other questions for understanding the relationship between them. The study of population growth must begin with period statistics. The question "How rapidly is the population growing?" cannot sensibly be posed in cohort terms. On the other hand, it makes no sense to ask "How long did people live in 1980?" We live out our lives over many decades. Yet there ought to be some way to say how long we live "these days" and how this compares to how long people lived in the 19th (or the 10th) century. Such questions require understanding relationships between period and cohort statistics.

Griffith Feeney
January 1998

Griffith Feeney is a demographer specializing in demographic measurement and analysis. He received a Ph.D. in demography from the University of California at Berkeley in 1972 and is currently a Senior Fellow at The East-West Center in Honolulu. He may be contacted by email at gfeeney@hawaii.edu, by mail at East-West Center, 1601 East-West Road, Honolulu, Hawaii 96848, or by telephone at 1-808-944-7456. You may download this digital document and give digital copies to other persons provided that you do not alter the file(s) in any way. Print copies may be made for personal use. All other rights reserved. © 1998 by Griffith Feeney.