Sociation Today® 
The Official Journal
The North Carolina
Sociological Association: A 
Refereed Web-Based 
ISSN 1542-6300
Editorial Board:

George H. Conklin,
 North Carolina
 Central University

Richard Dixon,

Chien Ju Huang,
 North Carolina
 Central University

Ken Land,
 Duke University

Miles Simpson,
 North Carolina
 Central University

Ron Wimberley,
 N.C. State University

Robert Wortham,
 North Carolina
 Central University




Volume 2, Number 2
Fall 2004

 The Concept of The Family: Demographic and Genealogical Perspectives


Charles B. Nam
Center for Demography and Population Health, Florida State University

     The social sciences include many concepts that are basic to the understanding of the subject matter.  Some of these concepts are defined differently and measured differently by social researchers, such as those concerning socioeconomic status.  Others are defined and measured in a fairly consistent way across studies and over time.  These latter concepts usually follow previous usages either because of research tradition or because of limitations in the collection of data, race perhaps being an example.

    A key concept in the social sciences, and especially in demography and sociology, is that of the family.  The family is generally regarded as a major social institution and a locus of much of a person's social activity. It is a social unit created by blood, marriage, or adoption, and can be described as nuclear (parents and children) or extended (encompassing other relatives). 

    It is generally assumed today that the modern family has undergone significant transformations in its structure.  We are told that societal changes have contributed to a sharp reduction in the percentage of classical  "typical" families, principally "nuclear" families.  Replacing these, we are made to understand, are childless families, one-parent families, other family configurations, and quasi-family units based on non-marital cohabitation.   This argument of the decline has been advanced for a number of decades, but little research has been conducted to test the premise.  Bane (1976) disagreed with that conclusion and pointed out that family sizes were getting smaller and mobility was splitting up some families, but the family remained as a functional social institution.

    The main contention of this paper is that analysis of changing family patterns is distorted by the definition of the family that is generally used and the way relevant data are collected.  In support of this contention, two different approaches will be used to gauge family status, and the two will be compared.  First, the standard demographic approach to defining and measuring the family concept will be reviewed. Second, the genealogical view of the family will be examined. A comparison will then be made of the two perspectives and their consequences for understanding the nature of changes in the modern family.

 Standard Demographic Approach

The family is generally recognized as an element of a broader kinship network that links ancestors and descendants of a person.  Most published statistics on the family are based on census or household survey questions and responses.  In the United States (and, for the most part, throughout the world), the "family" is defined in censuses and surveys as two or more persons  related by blood, marriage, or adoption, AND living in the same residence  (Fields and Casper 2001).  The first part of the definition excludes non-marital cohabitation but can include extended as well as nuclear family members.  However, the second part of the definition severely restricts family composition by limiting the family members to those who share living facilities under the same roof  (Glick 1957).  This standard definition is basically an accommodation to requirements of data collection in censuses and surveys in which identifying population in geographic contexts down to the separate dwelling unit is necessary.  Moreover, the questions needed to identify non-residential family members would be burdensome and the information costly to obtain.

    Persons who might be considered part of a family but do not reside at the same residential address are not included in demographic data.  They may be part of a family at another address or they may be living alone or in group quarters (housing for a substantial  number of unrelated individuals).  This is the case even if such persons live close by (maybe even next door) and/or visit or otherwise regularly communicate (by phone or mail) with their family of origin.  Additionally, because of census and survey residence rules, college students living in a college community and some long-term workers at remote places are excluded from the family group even if their intention is to return to the family's residence after school or work is completed.  In other words, the family definition is controlled by the household definition, where households describe current or potential housing markets.

    In fact, some persons who meet the standard demographic definition of the family and are included may have little association with other family members in the same residence.  For example, they may have different schedules of sleep, work, or other activities, and they may not communicate by phone or mail.  Their inclusion in the family is pro forma and based only on the given family definition. These facts raise questions about the boundaries of the standard demographic definition of the family and its consequences for interpretations of how family structure might be changing over time. 

The Genealogical Approach

    Genealogy is the study of family structural history, drawing basically on demographic data sources such as censuses, birth and death certificates, immigration records, and other administrative records.  The aim of genealogical research is to construct a family tree of ancestors and dependents of a key person (Smith and Mineau 2003).  The tree can be limited in its extension to cousins and other persons remotely related, but typically the attempt is to be inclusive of related kin.   Some genealogists prefer the term "family history" to "genealogy" because the latter term implies a genetic connection that may not be real because of questionable paternity and because it would not apply to adoptive persons.

    Many types of information can be included in family trees, but the pattern of relationships is not dependent on residential locations.  Residence can be one item of information for each individual in the tree, along with such items as dates of birth and death, place of birth, occupation, and other personal markers.  One can examine a family tree and extract a family structure using a variety of family definitions, based on how extensive one wishes to consider the family (Finnegan and Drake 1994).  Family trees typically distinguish between living and dead members of the family, so that several family definitions can be applied to only living members.  In this sense, the genealogical approach to looking at family structure provides for a broader range of family forms than is possible from the demographic approach.  Thus, one can describe a couple and their offspring, living together or not; a multi-generation family, living together or not; as well as extended family groupings. 

    Genealogies have not been incorporated into family research very much.  Smith (1987) indicates that obtaining any type of kin count or structure (and, by implication, residence-based families) "is often difficult or impossible ….  Genealogical research, even when done with the aid of computers, is labour- intensive and requires extensive archival data."  The use of genealogies in demographic research has been heavily oriented to estimating population size, as well as fertility and mortality of communities.

    Because sets of family trees are often hard to come by, the broader kinship network that the family tree describes can be obtained by having survey respondents reconstruct the history of a family’s changing structure by tracing  the family’s evolution from the marriage date of a couple to  the point where only one member of the family group is still living.

Comparison of the Two Approaches

    The distinction between the demographic and genealogical approaches can be illustrated by looking at the time trend in family patterns using various family definitions.  In the real, illustrative example I have chosen here family structures are compared based on the demographic concept and three alternative genealogical concepts of the family. 

The time span in this example is from 1956 to 2002 and covers the years of five decennial censuses from 1960 to 2000.  Initials are used to designate different family members.  The first person listed is the key person.  The first column describes the nuclear family composition according to census definition.  The second column adds in living members of the nuclear family who are non-residents.  The third column adds in living parents, parents-in-law, children's spouses, and grandchildren, whether or not living in the same residence  The fourth column adds in brothers and sisters of the key person, also not necessarily in the same resideence.  Of course, many other variations and extensions of the family are possible.  Over time, new members are added as they meet the family definition and members are deleted when they die or are divorced from a nuclear family member.

    In this particular example, the family begins with a 1956 marriage.   Columns one and two show that both census and genealogical approaches report the nuclear family the same way.  Columns three and  four indicate two sets of parents and two pairs of siblings.  By the time of the 1960 Census, these family patterns have remained the same.  In 1961, a child is born to the couple and is the only addition to the family.  In 1962, a second child joins the family and all family definitions expand similarly.  This pattern continues through the time of the 1970 Census. 

    In 1971, one of the couple’s parents dies and the extended family definitions are reduced.  In 1975, another parent dies, further reducing the more extended family patterns.  No change takes place through the 1980 Census period.  By 1981, one of the children leaves for college away from home.  Now, for the first time, census and genealogical family approaches define the nuclear family's composition differently.  The following year, 1982, the other child leaves home for college and further disparity in nuclear family composition ensues.  In the same year, a third parent dies, and the extended family groups become smaller.

     In 1983, the first child transfers to a college near home and resumes residence with his parents.  In 1984, the second child completes college and returns home while awaiting marriage to someone she's engaged to.  The next year, 1985, sees her departure to her own household with the newest extended family member.  By 1986 the other child completes college and sets up his own residence in the community.  In 1988, financial considerations bring that child back home for a considerable part of the year.  The following year, 1989, he departs for professional school.  In the same year, the other child has a son and thereby expands the extended family patterns.  The family patterns remain the same through the 1990 Census period.  In 1991, a daughter is born to complement the son of the one child; and in that same year one of the sisters of the initial couple passes away.  When 1995 rolls around, the son has gotten married, adding another extended family member.

    In 1999, the son divorces and rejoins the couple.  By the time of the 2000 Census, the son has found a new residence in town; and later that year he remarries, thus establishing the earlier extended family sizes.  In 2001, the nuclear family as defined by the demographic approach comes to an end as one of the original couple dies.  Then, in 2002 the remarried son has a child.

Family Definitional Implications

    A scan of the chart shows important differences in family composition according to the various definitions. 

  •     First, the nuclear family independent of residence, as revealed in the second column, did not change from 1962 to 2001, a period of 39 years.  In contrast, the nuclear family demographically defined remained intact from only 1962 to 1980, a period of 18 years. 
  •     Second, according to the genealogical approach, there is still an existent family by 2002, a period of 46 years and counting.  In contrast, based on the demographic approach, the family ceases to exist after 2000.
  •     Third, the family as defined by census or survey has been quite variable over time, and some of the changes are due to very short-term stays in the residence.  In fact, in one year, 1984, the family structure that was established in 1962 and continued through 1980 is reconstructed. 
  •     Fourth, the versions of the extended family shown here are at their smallest size when the family demographically-defined is at its largest size.

     One could reasonably argue for any of the family definitions shown here, or for other variants.  A point to be made, however, is that the family as defined demographically may not be the definition of the family that most of us think of when we contemplate our own family pattern. 

     More importantly, from the perspective of the dynamics of family changes, analysis of changing family structure using the demographic approach may overstate the fluidity and demise of the nuclear family form, whereas greater family stability is indicated by the genealogical approach.  Family researchers should consider whether residential separation of family members reduces family structure even when family functions are basically maintained, especially in light of increases in the availability of travel and communication channels.  Bane (1976) argued further that smaller family sizes probably led to stronger attachments of residence-based family members with their non-residence-based family members.  Those who design questions related to the definition of families in censuses and surveys should consider if additional information needs to be collected to add meaningfulness to the analysis of family composition and changes in it.  In any event, our analysis of changes in the family as a social unit should not be held hostage to a definition and measurement approach that may not adequately reflect its true character.


Bane, Mary Jo  (1976).  Here To Stay: American Families in the Twentieth Century.  New York: Basic Books.

Fields, Jason, and Lynne Casper  (2001).  "America's Families and Living Arrangements."  Current Population Series P-20-537.

Finnegan Ruth, and Michael Drake (1994).  From Family Tree to Family History.  Cambridge University Press.

Glick, Paul C.  (1957).  American Families.  New York: John Wiley.

Smith, James E.  1987.  "The Computer Simulation of Kin Sets and Kin Counts," in Bongaarts, John, Thomas K. Burch, and Kenneth W. Wachter, eds., Family Demography: Methods and Their Applications.  Oxford: Clarendon Press.

Smith, Ken R., and Geraldine P. Mineau  (2003).  "Genealogical Records," in Demeny, Paul, and Geoffrey McNicholl, eds., Encyclopedia of Population.  New York: Macmillan.

Return to Sociation Today, Fall 2004

©2004, The North Carolina Sociological Association.