The Future of Historical Family Demography

SO38CH18-Ruggles ARI 20 April 2012 15:7

V I E E W R S

I E

N C N A D V A

The Future of Historical Family Demography

Steven Ruggles

Minnesota Population Center and Department of History, University of Minnesota, Minneapolis, Minnesota 55455; email: [email protected]

Annu. Rev. Sociol. 2012. 38:18.1–18.19 Keywords The Annual Review of Sociology is online at family transition, family composition, historical sociology, patriarchal soc.annualreviews.org family This article’s doi: 10.1146/annurev-soc-071811-145533 Abstract Copyright c 2012 by Annual Reviews. An explosion of new data sources describing historical family compo- All rights reserved sition is opening unprecedented opportunities for discovery and anal- 0360-0572/12/0811-0001$20.00 ysis. The new data will allow comparative multilevel analysis of spatial patterns and will support studies of the transformation of living arrange- ments over the past 200 years. Using measurement methods that assess family choices at the individual level and analytic strategies that assess variations across space and time, we can dissect the decline of patriar- chal family forms in the developed world and place Northwest Europe and North America in global comparative context.

18.1 SO38CH18-Ruggles ARI 20 April 2012 15:7

INTRODUCTION moment or a brief period between the seventeenth and the nineteenth centuries. This is an extraordinary moment for historical This chronological focus misses the main family demography. We are in the midst of an action. The transformative changes in the explosion of new data resources that are open- family—such as the decline of intergen- ing unprecedented opportunities for discovery erational coresidence, the rise of marital and analysis. This review explores the impli- instability, and the increase of single- cations of the new data sources for the study parent families—occurred after the of the composition of households, families, and mid-nineteenth century. kin groups.1 I have three main arguments:  We should use demographically ap- I begin by describing the new data resources propriate measures. To make big that have recently been released or will soon comparisons and analyze household become available. I then provide an overview of and family composition across time and the origins of historical family demography and space, we must use the best available describe the ways that the new data can allow us methods and be sensitive to the effects of to break free from limiting debates and methods variation in both population composition of the past. Finally, I comment on some of the and the availability of kin. This means broad theoretical issues facing the field. that we should avoid common measures currently in widespread use.  We should study spatial variation in fam- NEW DATA ilies and households. Historical research For the past half-century, the great bulk of on the demography of households and historical studies of families and households re- families has focused on analysis of small lied on small data sets laboriously gathered by communities. One of the rationales for individual scholars to describe particular com- the community approach has been that munities in the eighteenth or nineteenth cen- local conditions have a powerful influence tury. The data sets were usually proprietary and on residence decisions. In fact, I argue, were rarely made public. the only way we can evaluate the impact Recent large-scale international collabora- of local conditions on household and tions have provided scholars with access to hun- family structure is to conduct multilevel dreds of millions of records from dozens of data analyses that assess the characteristics of sets, and new projects now under way are poised many communities simultaneously. to disseminate an avalanche of new historical  We should study long-run historical data in the coming years. These initiatives are change. Most historical studies of liv- designed to maximize comparability across time ing arrangements focus on a single and space, and they are conceived as shared re- sources for the research community. The new initiatives are listed in Table 1. 1In accordance with United Nations (2001) definitions, I use the term household to mean a person or group of people who live together and make common provision for food or other Integrated Public Use essentials for living, and family to mean a group of people residing in the same household who recognize a kin rela- Microdata Series tionship, ordinarily through descent, marriage, or adoption. The Integrated Public Use Microdata Series I use the term kin group to refer to the broader group of persons recognized as kin, regardless of living arrangements. (IPUMS) was conceived in 1991 as a method Following Cherlin (2003), I define family demography as the for making heterogeneous US Census data study of (a) the configuration of families, households, and kin sets for the period since 1850 easily interop- groups, and (b) transitions that affect those configurations. Such transitions include departure from a parental home, erable, thereby simplifying analysis of histori- marriage, marital dissolution, and cohabitation. cal change. In 1999, IPUMS began to expand

18.2 Ruggles SO38CH18-Ruggles ARI 20 April 2012 15:7

Table 1 Major new historical data infrastructure projects International IPUMS Integrated Public Use Microdata Series http://www.ipums.org NAPP North Atlantic Population Project http://www.nappdata.org IDS/EHPS-Net European Historical Population Samples http://www.iisg.nl/hsn/news/ehps-net.php Network Mosaic Recovering Historical Census Records http://www.censusmosaic.org National CMGPD China Multi-Generational Panel Dataset http://www.icpsr.umich.edu/icpsrweb/CMGPD/ HPR Norwegian Historical Population Registers http://www.rhd.uit.no/nhdc/hpr.html POPLINK Swedish Historical Population Registers http://www.ddb.umu.se/databaser/poplink/ CCRI Canadian Century Research Infrastucture http://www.canada.uottawa.ca/ccri/CCRI/index.htm I-CeM Integrated Census Microdata (I-CeM) http://www.essex.ac.uk/history/research/ICeM/default.aspx Project

beyond US borders, and today the project part- Egypt, Ireland, and Mexico. All 12 partners ners with the national statistical agencies of will contribute new data sets, so by 2016 the over 100 countries to disseminate census and project will include 64 censuses taken between survey microdata for the period since 1960. 1787 and 1930 describing the living arrange- IPUMS is the world’s largest demographic mi- ments of approximately 365 million persons crodata collection; it is currently disseminating (Ruggles et al. 2011). For 90% of these cases, over 800 million cases of integrated microdata the IPUMS project will provide comparable drawn from more than 750 censuses and surveys data pertaining to the period since 1960. To (IPUMS 2011). Over the next five years, the make long-run comparisons easy—and to help database is scheduled to double in size (Sobek ensure the long-run sustainability—the NAPP et al. 2011). database is being incorporated into the IPUMS infrastructure.

The North Atlantic Population Project The North Atlantic Population Project Intermediate Data Structure (NAPP) is a collaboration of researchers in Historical demographers have created a Britain, Canada, Iceland, Norway, Sweden, variety of large-scale longitudinal databases and the United States to make a harmonized constructed by linking information from database of nineteenth-century census enu- population registers, genealogies, or church merations (NAPP 2011). Most of the NAPP records to provide continuous information data sets are not samples, but rather include on historical life courses. Among the most entire national populations. In addition to prominent are the Demographic Data Base disseminating cross-sectional microdata, the (Umea˚ University), the Scanian Demographic partnership is exploiting data-mining technol- Database (Lund University), the Balzac Popu- ogy to develop representative linked samples lation Register (Quebec), the Utah Population with multiple observations in individuals, Database, the Enqueteˆ TRA (France), and the families, and households (Goeken et al. 2011). Historical Sample of the Netherlands. Many A new phase of the NAPP project is now being of these databases provide direct measures of launched, with an expansive interpretation household composition, and they all include of the term “North Atlantic”: New NAPP information of crucial importance for historical partners include Albania, Denmark, Germany, family demography. The creators of these data

www.annualreviews.org • The Future of Historical Family Demography 18.3 SO38CH18-Ruggles ARI 20 April 2012 15:7

collections have sometimes tried to share them release in August 2011 provided 10 data sets with outside researchers, but owing to the from 7 countries with 315,000 observations, complexity of the data, access is cumbersome and collaborators have already pledged to con- and usage of each database is concentrated tribute dozens of additional files. Over the next among its producers. decade, Mosaic organizers hope to expand the The Intermediate Data Structure (IDS) is collection to cover 40 countries with millions an ambitious collaboration among the world’s of observations spanning five centuries. major producers of longitudinal historical databases. The goal is to open access to the collections and to make them interoperable. National Projects The project involves converting about 15 lon- The explosion of new data is not limited to gitudinal databases into a common structure these four international collaborations. Within with common metadata and disseminating specific countries, various additional large-scale them through a sophisticated Web-based data historical data projects have either recently access tool that will produce customized files been completed or are now under way. Space specialized for analysis of particular topics, limits preclude a full listing of these initiatives, such as fertility or household transitions (Alter but a few examples can illustrate their extraor- et al. 2009). The consortium plans to release dinary scale and scope. a preliminary version of the data access tools  The China Multi-Generational Panel with contributions from four databases by the Dataset, released in 2010, provides in- end of 2012 and to add six more databases in the dividual and household longitudinal data following year (EHPS-Net 2011). A complete for 260,000 persons in Liaoning province system with the full roster of participating data between 1749 and 1909 (Lee & Campbell collections is planned to be complete by 2016. 2011); the database will soon be expanded with records from Shuangcheng.  Separate initiatives in Norway (Norwe- The Mosaic Project gian Historical Data Centre 2011) and A new collaboration of researchers from Sweden (POPLINK 2011) are bridging most European countries aims to assemble the gap between modern population reg- thousands of census and census-like listings for isters and existing historical databases by local communities in the past (Mosaic 2011). digitizing records from the twentieth cen- In most European countries, the enumeration tury, providing continuous series of lon- forms for national censuses taken before the gitudinal data from the early nineteenth mid-twentieth century were not preserved; century to the present. accordingly, complete microdata collections  The Canadian Century Research In- and national samples like those in NAPP and frastructure (Sager & Baskerville 2010) IPUMS are not possible. In virtually every will fill the gap between nineteenth- country, however, census manuscripts or century data sets available through NAPP census-like listings known as libri status ani- and twentieth-century files disseminated marum survive in scattered archives. Historical through IPUMS by creating large sam- demographers have already digitized a great ples of the censuses of 1911, 1921, 1931, deal of these data to support community-level 1941, and 1951. studies. The Mosaic Project will gather existing  The Integrated Census Microdata machine-readable microdata and harmonize Project (I-CeM 2011) will provide it according to IPUMS and NAPP standards, complete microdata for every census identify gaps in the collection and locate of England, Wales, and Scotland be- sources to fill them, and coordinate digital tween 1841 and 1911—a total of about conversion for the new additions. The first data 210 million records.

18.4 Ruggles SO38CH18-Ruggles ARI 20 April 2012 15:7

There are still gaps. Pre–twentieth century The nuclear families Le Play identified were data for most of the developing world remain mainly located in England and the manufactur- scarce. There are no substantial public data ing districts of Western Europe. There, “the sets describing family and household compo- young adults leave their parental firesides as sition in sub-Saharan Africa or South Asia be- soon as they gain any confidence in themselves” fore 1960, and only limited data are available (Le Play 1872, p. 41). The result was disastrous: for Japan, Korea, and Latin America. Never- “the parents are isolated in their old age and die theless, for the first time researchers can freely abandoned” (Le Play 1871, p. 9). access a vast body of records describing histor- Stem families, Le Play believed, were ideal. ical family and household composition across They offered greater flexibility than joint fam- thousands of places and long time spans. In just ilies without the instability of nuclear families. a few years, the total available historical micro- Accordingly, Le Play was alarmed by what he data will exceed a billion records. To capitalize saw as a gradual shift from stem families to nu- on the new resources, we must make significant clear families. In part, he blamed Napoleonic changes in the style of research that has become inheritance law, which mandated equal division standard over the past five decades. of property among all heirs, eliminating the power of the patriarch to designate his succes- sor. At the root, however, he saw the changing LE PLAY, LASLETT, AND THE organization of labor as a fundamental threat ORIGINS OF HISTORICAL to the stem family. For the stem family to suc- FAMILY DEMOGRAPHY ceed, the patriarch must be the proprietor of The first systematic investigation of change the family farm or workshop. With the rise in the configuration of families was conducted of large commercial and manufacturing pop- by the reactionary mid-nineteenth century ulations, the tie between work and family was social scientist Fred´ eric´ Le Play (1855, 1871, severed, and the stem family was undermined. 1872). Le Play gathered case studies describing In these circumstances, the younger generation individual families across Europe and Western was vulnerable to the lure of high wages and the Asia and concluded that there were just three “attractions of city life” (Le Play 1872, p. 79). family systems found at all times and places: Subsequent observers built on Le Play’s the joint family ( famille patriarcale), the stem research. Devas (1886) wrote that joint families family ( famille souche), and the nuclear family were common in China and India and could ( famille instable). Le Play’s family types have also be found in Southeastern Europe and had extraordinary influence on sociological Central Italy but concurred with Le Play that thought, and they continue to overshadow in places such as modern France, England, debate and discussion among historical family and North America, the diminished “power of demographers. the father over his children” had led to “rapid Joint families and stem families are both dissolution” of multigenerational families multigenerational. In joint families, “parents al- (Devas 1886, pp. 44, 211). Durkheim (1888) ways retain near them all their married sons, introduced the idea that with the growth of and the children issuing from such marriages,” social differentiation, specialized functions of whereas in stem families, “the father transmits the family were lost and kin ties were weakened his fireside and place of labor to that one of (Lamanna 2002). By the mid-twentieth cen- his children which he thinks most capable,” and tury, sociologists agreed that industrialization sends the other children out into the world (Le and modernization had reduced the functions Play 1872, pp. 40–41). Le Play observed joint of the family; turning Le Play’s interpretation families mainly in Eastern Europe, and argued on its head, many concluded that flexible that stem families predominated in many parts and mobile nuclear families were ideally of Western Europe, including parts of France. adapted to the needs of modern industrial

www.annualreviews.org • The Future of Historical Family Demography 18.5 SO38CH18-Ruggles ARI 20 April 2012 15:7

society (Ogburn 1933; Parsons 1944, 1955). parish records of England in search of similar Goode (1963, p. 6), reflecting this consensus, listings. They found census-like listings for 100 wrote that “wherever the economic system English villages between 1574 and 1821, and expands through industrialization ...extended an average of just 10.1% included extended kin kinship ties weaken, lineage patterns dissolve, (Laslett 1969). This figure was virtually identi- and a trend toward some form of the conjugal cal with the percentage of households contain- system generally begins to appear.” ing extended kin in the 1961 census of England Before the mid-twentieth century, theoriz- and Wales (General Register Office 1968). The ing about family change took place without the interpretation was clear: Family structure had benefit of statistical evidence. Le Play’s case been extraordinarily stable over four centuries. studies belong to a prestatistical era; he carefully There was no shift from extended to nuclear selected particular families for analysis because families with industrialization; the English fam- he felt they were typical. The British Census of ily had always been overwhelmingly nuclear. 1931 and the US Census of 1930 produced the In 1969, Laslett held a conference at the first national tabulations of household compo- newly formed Cambridge Group for the sition, but the classifications were rudimentary History of Population and Social Structure and in the US case they were not published until to discuss the new findings and place them 1940 (Nixon 1952, Glick 1941). From the 1940s in broader context. He invited scholars from onward, however, family and household statis- Europe and the United States who were tics from censuses and surveys became more working on historical family structure, and in and more abundant and detailed, and the sub- advance of the meeting he circulated his paper field of family demography emerged to track the describing families in the 100 English com- rapid transformations of the postwar era (Glick munities. Participants contributed essays on 1988). the United States, several European countries, Statistical research on the configuration of and Japan. When the conference volume was families and households in earlier times began published under the title Household and Family in the 1960s. While on sabbatical at the Li- in Past Time (HFPT), it was a bombshell. brary of Congress in 1959, English historian Laslett concluded that family organization Peter Laslett stumbled on a reproduction of was “always and invariably nuclear” (Laslett two census-like listings of the 400-odd inhab- & Wall 1972). This idea resonated powerfully itants of the village of Clayworth, in Not- among sociologists, many of whom had already tinghamshire, one taken in 1676 and the been questioning the connection between other in 1693 (Schurer¨ 2003). The population industrialization and nuclear family structure was divided into small units—presumed to be (Greenfield 1961, Litwak 1960, Sussman 1959). households—and within each unit the family The same year that Laslett’s findings about relationships were given. Laslett had expected Clayworth first became widely known, Hajnal to find that households in preindustrial England (1965) published a landmark essay arguing that were typically large and complex, filled with Northwest Europe from the seventeenth to the “grandfathers, aunts and uncles sitting round nineteenth centuries had a distinctive pattern of the cottage fire with grandchildren, cousins and late marriage and a high percentage of people second cousins all under the mastership of the who never married, compared with most of the patriarchal head” (Laslett 1970, p. 87). As he ex- rest of the world. Hajnal and Laslett combined amined the households, Laslett was shocked to their findings and concluded that early modern find that they were generally small, and just 1 in England, Northern France, the Low Countries, 10 units included kin other than the household Germany, and Scandinavia had an unusual fam- head, his wife, and their children (Laslett 1965). ily formation system of nuclear families, late Over the next few years, Laslett assembled marriage, and high domestic service (Hajnal an army of volunteers to comb through the 1982, Laslett 1982). Hajnal’s Northwest

18.6 Ruggles SO38CH18-Ruggles ARI 20 April 2012 15:7

European family system is regarded by some Papua New Guinea had the same death rate scholars as an essential condition for the in 2009. But neither of these comparisons is development of early modern capitalism and meaningful. the industrial revolution (Macfarlane 1986, Like deaths, living arrangements are highly 1987; Cain & McNicoll 1988; Hartman 2004). sensitive to age. Controlling for age is thus This is a radical departure from the structural essential whenever making comparisons of theories of the family touted by Ogburn and household composition across populations. Parsons: According to the new interpretation, But age structure is not the sole comparability the nuclear family was not a consequence of issue; demographic conditions also affect family industrialization but rather its cause. composition by determining the availability of Laslett opened HFPT with an 89-page in- kin for coresidence. One cannot, for example, troduction that was partly manifesto and partly reside with a parent if both of one’s parents a detailed set of instructions about how to con- have died. The particular configuration of kin duct research on the history of families and available for coresidence in a population is households, including an elaborate system of a direct function of the prevailing levels and household classification. This guidebook has timing of fertility and mortality. had great success: Hundreds of studies of com- As soon as Laslett first published his finding munities all over the world have been conducted that there were few extended families in prein- using Laslett’s approach and his household clas- dustrial England, Glass (1966) and Wrigley sification scheme. Not all of these have en- (1969) developed simple analytic models dorsed Laslett’s interpretation. Regardless of suggesting that preindustrial demographic whether they endorse or criticize Laslett’s ideas, conditions posed substantial demographic con- however, most studies in historical family de- straints on the potential for multigenerational mography have followed the basic template for families. Berkner (1972) argued eloquently that research he introduced in HFPT. This model Laslett’s cross-sectional data were incapable of research has three basic characteristics: of detecting Le Play’s stem family, which (a) It uses methods and measures derived from describes a process of inheritance and family Laslett’s cookbook, (b) it focuses on a sin- succession, not the configuration of families at gle community or a few nearby communi- a particular moment in time. ties, and (c) it analyzes data from a relatively We now know that Berkner was correct: brief period that occurred before the twentieth Laslett’s classification cannot detect Le Play’s century. stem families. The proof was recently supplied by Wall (2009), Laslett’s longtime collab- orator. Wall examined the case studies of MEASUREMENT individual families that Le Play compiled in In 2009, Switzerland and Papua New Guinea the mid-nineteenth century, and he classified had an identical crude death rate of 8 deaths per them according to Laslett’s system. Among 1,000 persons. Among other nations, the death Le Play’s stem families, Wall classified 81% rate per 1,000 was 3 in Syria, 5 in Nicaragua, 9 as simple families under the Laslett scheme. in Japan, and 10 in Sweden (World Bank 2011). At the moment of Le Play’s interview, those There are good reasons demographers do families consisted only of parents and their not use crude death rates to make comparisons children. Le Play nevertheless considered them across countries with differing age composi- stem families because of their system of family tion: The results are invariably misleading. succession: One child was designated as the heir Laslett’s finding that the percentage of nuclear and would remain in the home after marriage. households in England was unchanged over In nineteenth-century Northwest Europe, the centuries was technically correct, just as marriage occurred late and death occurred it is technically correct that Switzerland and early, so the period of overlap between marriage

www.annualreviews.org • The Future of Historical Family Demography 18.7 SO38CH18-Ruggles ARI 20 April 2012 15:7

and the death of parents was often brief; in many include three generations. In the twentieth cases, people died before the birth of their first century, the potential for multigenerational grandchild. Only the brief overlapping period families increased dramatically; more than 1 in can be observed as a three-generational family 5 persons in England and Wales is now age 60 in a census. Under stem family formation rules, or older, and 1 in 6 is age 65 or older. the issue is compounded by fertility: Only one Figure 1 illustrates the difference between married child remains in the parental home, and a household-level approach and one based all the others go out to form new nuclear fam- on individual-level coresidence using IPUMS ilies. Accordingly, under a stem family system census microdata for the United States between the maximum observed frequency of multigen- 1880 and 1950. The top line shows the percent- erational households is inversely proportional age of persons aged 65 or older residing with to the level of fertility (Ruggles 1987). children and grandchildren, and the bottom In preindustrial England, the average line shows the percentage of households that age at childbirth for women who survived contained three generations. Both measures use the childbearing years was almost 35, and the 1970 Census Bureau definition of house- their husbands were generally a year or two holds to maximize comparability (Ruggles & older (Ruggles 1987). Because of the long Brower 2003). Between 1880 and 1950, the per- generations, multigenerational families usually centage of elderly persons residing with three had to include an elderly person. The old, generations fell by more than half, but during however, were thinly spread through a very the same period the percentage of households young population; only about 1 in 15 persons with three generations was nearly constant, was aged 60 or older, and 1 in 20 was 65 or dropping by just one percentage point. older (Wrigley & Schofield 1981). The scarcity Demographic opportunities to form of the elderly meant that few families could three-generation households roughly doubled between 1880 and 1950, for two reasons. First, mortality declined, so more elderly parents 30 were available for coresidence. Second and Percent of persons age 65 or more residing even more important, fertility declined, so the with children and grandchildren 25 pool of available elderly parents was shared among a smaller younger generation. But the increasing opportunities for coresidence 20 were counterbalanced by a massive decline in the proportion of people who capitalized on those opportunities, creating the illusion of 15 stability. Thus, household-level measurement in the context of rapid demographic change masks massive changes in family decision 10 making, just as it did when Laslett compared Percent of households with the percentage of extended households in three generations preindustrial England with the Census of 1961. 5 Measuring intergenerational coresidence from the perspective of the oldest generation min- imizes the effects of variation in demographic 0 conditions. Even in populations where few 1880 1890 1900 1910 1920 1930 1940 1950 households have the potential to include three Figure 1 generations, most persons over the age of 65 Measures of three-generation families in the United States, 1880–1950. have the demographic potential to reside with Source: IPUMS (2011). children and grandchildren.

18.8 Ruggles SO38CH18-Ruggles ARI 20 April 2012 15:7

Historical studies of family structure still measures of family composition where the rely heavily on the Laslett classification. impact of demographic factors is straightfor- For example, among 12 quantitative studies ward and predictable. We can then introduce of household composition appearing in the statistical controls for the general demographic journal History of the Family during the past five characteristics of the population—such as years, eight used the Laslett scheme. Similarly, fertility, age at marriage, and age structure— in a recent landmark collection on stem families as well as individual-level controls for age (Fauve-Chamoux & Ochiai 2009), 10 of 15 (Ruggles 2010, Gruber & Szoltysek 2011). quantitative analyses of household composition In addition to allowing us to control for the use the Laslett system, even though several of effects of demographic conditions, individual- those studies stress its limitations. level measures of family composition have other International and cross-temporal compar- big advantages over household-level measures. isons of the percentage of households that are The strategy of assessing living arrangements of extended or multiple are often meaningless. Just households as a whole in a unitary way makes as the crude death rates of Syria and Sweden it impossible to assess the impact of individ- do not reveal which country has the higher ual characteristics on residence decisions. This risk of mortality, the percentage of complex makes it difficult to study generational and gen- households tells us virtually nothing about the der dynamics. Household-level analysis makes family system. This is not a problem specific life-course analysis—as opposed to a life-cycle to Laslett’s classification. No household-level approach—impossible (Elder 1978, Kok 2007). measurement system can control for age com- The conventional measures prevent us from as- position or the availability of kin. The measure- sessing differentials in the familial experience ment schemes and classifications proposed by of men and women. If we wish to analyze the Burch (1967), Hajnal (1982), Verdon (1998), residential decision making of individuals, we and the United Nations (2001) suffer from ex- must adopt individual-level measures (Ruggles actly the same problems. & Brower 2003). Demographers have developed a variety of If existing household-level classifications methods to assess the effects of demographic are limiting and misleading, what should we conditions on the availability of kin (e.g., do instead? I do not have a Laslett-style cook- Bongaarts et al. 1987, Ruggles 1987, De book, but I do have some suggestions. If the Vos & Palloni 1989, Wolf 1994, Murphy microdata for many countries are readily avail- 2011). If sufficient demographic information able in a standardized form, that reduces the is available, microsimulation and life-table necessity of standard classifications. Individual methods can provide reasonable estimates of researchers can develop measures customized the percentage of persons with a particular to particular research problems and apply set of characteristics that had living kin of a them to a wide range of data sets, controlling particular type. In a few cases, historical sources for population composition and demographic provide sufficient information to measure the characteristics as necessary. Family measures availability of kin directly (Post et al. 1997, should follow four basic principles: Nevin 2002, Van Baelen & Matthijs 2007). But 1. We should always control for age, by re- even if demographic or empirical analysis can stricting analysis to either a specific age reveal the rough magnitude of demographic band, standardization, or regression. constraints on the availability of kin under a 2. We should anticipate the potential effects particular demographic regime, they do not of- of demographic conditions on the avail- fer practical solutions that can be incorporated ability of kin and control for them when into comparative or cross-temporal analyses feasible. of family composition. The best available 3. We should keep the measures simple. approach is to employ simple individual-level Complex classifications make it difficult

www.annualreviews.org • The Future of Historical Family Demography 18.9 SO38CH18-Ruggles ARI 20 April 2012 15:7

Table 2 Examples of individual-level family measures Name Numerator Universe/denominator Useful controls 1. Elderly with adult children With child aged 18+ Age 65+ Age, fertility level 2. Three generation With child or child-in-law and Age 65+ Age, fertility level, age at grandchild marriage 3. Stem family With ever-married child, Age 65+ Age, fertility level, age at child-in-law, or grandchild marriage 4. Joint family With two married children Age 65+ Age, fertility level, age at marriage 5. Adult sons with parents With parent Men age 30–39 Age, fertility level 6. Children with single mother With mother only Age 0–18 Age 7. Mothers without spouse Without spouse Mothers of children age Age 0–18 8. Unmarried fertility With child age 0–4 Unmarried women age Age 14–49 9. Marital instability Separated or divorced Women age 20–39 Age

to anticipate and control for the in- ages are beyond the usual ages of leaving home tervening effects of demographic condi- and yet are young enough that even in the tions. Dichotomous measures identifying nineteenth century about half still had a surviv- coresidence of specific kin types are ideal. ing parent with whom they could potentially reside (Ruggles 1994). Measure 5 is highly sen- 4. When designing measures, we should pay sitive to fertility and mortality, so unadjusted as much attention to denominators as to trends in the absolute level of coresidence numerators; in most cases, we should re- could be highly misleading, but comparison of strict the analysis to a population at risk socioeconomic differentials is less problematic. of living in a particular situation. The remaining measures are straightfor- Table 2 provides examples of cross- ward. Measure 6 focuses on the experience of sectional measures that adhere to these children; it is appropriate for analyzing the principles. The first five measures focus on in- impact of single parenthood, but it is not ideal tergenerational coresidence, which is especially for assessing residence decisions given that sensitive to demographic conditions. Most of children rarely decide where to live. Measure these intergenerational measures focus on the 7 assesses lone motherhood relative to all population aged 65 or older, an age at which mothers, and measure 8 is a simple indicator of most people in all populations have grown unmarried fertility. Marital instability, measure children with whom they could reside. Cores- 9, is an indicator of prevalence, so it reflects the idence decisions, however, depend not only combined effects of past divorce, separation, on the characteristics of the older generation, and remarriage; with a few exceptions, one can- but also on the characteristics of the younger not measure divorce rates with cross-sectional generation. Measure 5 is designed to assess the data, but the prevalence measure is an effective effects of occupational status on coresidence indicator of the level of marital instability that with parents and focuses on men aged 30–39. It maximizes comparability across time and space is limited to men because occupation is a poor (Heggeness 2010). indicator of socioeconomic status for women The demographic controls listed in the before the late twentieth century. I define the right-hand column can either be individual- younger generation as 30 to 39 because those level variables or contextual variables measured

18.10 Ruggles SO38CH18-Ruggles ARI 20 April 2012 15:7

at the local, regional, or national levels. Age SPACE AND TIME should always be controlled at the individual Ever since Le Play, geography has been a level, but it is also useful to add contextual central preoccupation of historical family de- controls for age structure of the population mography. A significant literature has focused because for many measures age distribution is on the spatial distribution of joint families closely associated with the availability of kin. (Wheaton 1975), which have been identified For example, when comparing measure 5 across in Russia (Czap 1982), Southeastern Europe populations, one can include an independent (Kaser 1994), and parts of Italy (Kertzer variable controlling for the percentage of the 1989). Historical family demographers have population aged 65 or older. That indicator also devoted substantial effort to refining the is closely correlated with the availability of boundaries of the Northwest European nuclear parents for men in their thirties and provides a family region identified by Laslett and Hajnal simple and effective way to account for the dif- (e.g., Plakans & Wetherell 2005, Szoltysek ferential availability of kin. We often lack good 2009), and many investigators have questioned mortality data, especially at the local level, so the premise that stem families were absent controlling for population age structure as well from Northwest Europe and North America. as individual-level age is a useful substitute. They point to evidence of stem families in With cross-sectional data, fertility and age nineteenth-century England (Ruggles 1987), at marriage can occasionally be controlled at Ireland (Gibbon & Curtin 1978), Finland the individual-level but almost always can be (Moring 2009), France (Fauve-Chamoux controlled at the population level. 1984), Norway (Sogner 2009, Ja˚stad 2011), None of these measures depends upon and the United States (Ruggles 1994). headship; they are strictly configurations of Despite decades of debate about the geog- kin sharing a household. Headship is critical raphy of historical family and household com- to analysis of living arrangements, but we position, however, there has been little spatial should not conflate family composition and analysis. The vast majority of quantitative evi- headship. Many studies use the percentage dence on family and household composition has of female-headed households as a measure of consisted of studies of a single community or a lone motherhood, but as Heggeness (2010) small group of communities. When different demonstrated, that measure misses up to half investigators analyze different communities, of lone mothers in some populations. Analysis comparability problems arise. The authors each of headship is critical but should be distin- use their own computer programs for manipu- guished from analysis of composition. Thus, lating and classifying their data and the results for example, it makes much more sense to are not always comparable from one author assess the percentage of lone mothers heading to another. Moreover, when data from differ- households than to assess the percentage of ent places are analyzed separately, no statistical households headed by lone mothers. analysis of spatial variation can be undertaken. Crude household-level measurement cre- Since the publication of Laslett’s original ated the illusion that the Northwest Euro- conference volume, comparison has been ac- pean family was unchanged across multiple cen- complished by publishing volumes of collected turies, and this misunderstanding dominated essays, each pertaining to a different place and the literature for decades. It is not appropri- analyzed by a different author. Fauve-Chamoux ate to use any household-type classification for & Ochiai (2009) includes studies of stem fami- analysis of trends or differentials in living ar- lies in more than a dozen historical populations rangements. Instead, we should adopt well- of Europe and Asia. Among the 20 essays in the designed individual-level measures, control for massive volume, most analyze either a single age structure, and be sensitive to the potential locality or two or three communities within a impact of kin availability.

www.annualreviews.org • The Future of Historical Family Demography 18.11 SO38CH18-Ruggles ARI 20 April 2012 15:7

local area. There are three exceptions that offer for example, the percent of elderly residing international comparisons, but even these offer with adult children dropped 80% from 1850 to only basic descriptive statistics that compare a 1990, and marital instability increased 500% small number of communities. (Ruggles 1997, 2007). By comparison, during One of the main rationales for community the same period fertility declined by 65% studies is to uncover the impact of local con- and life expectancy doubled (Carter et al. ditions on family composition. It is impossible, 2006). But although the fertility and mortality however, to empirically establish a relationship transitions have been intensively studied, there between conditions and family composition if has been comparatively little research on the the N is 1; one must have data from many family transition. communities to assess variation in conditions Time is curiously absent from much histor- and behavior. Not infrequently, investigators ical family demography. Most studies focus on compare communities with different character- specific communities at a particular moment or istics. For example, in a recent article Breschi over a brief period. They sometimes study the et al. (2009) analyze the impact of agricultural effects of short-run economic, demographic, or regimes on family complexity and remarriage of political events on families but less often ex- widows. Although the study is fascinating and amine secular trends. Among a dozen quantita- suggestive, it is ultimately inconclusive because tive analyses of household structure published it is based on just three observations. in the journal History of the Family during the To statistically exploit the comparative past five years, only one-third looked at long- strategy through spatial analysis, we need run change. information about many places. The new data The greatest period of change in the family sources give us detailed information about was the twentieth century. The United States tens of thousands of communities. They also is presently the only country with a continuous provide information at the level of individuals, series of microdata spanning the twentieth cen- households, neighborhoods, provinces, and tury; as a consequence, there has been an explo- countries, allowing true multilevel analysis. A sion of studies of long-run family change in the systematic approach to spatial analysis incorpo- United States. Dozens of investigators have ex- rating multiple indicators of the characteristics plored a wide range of topics, including marital of places can provide powerful new tools with disruption (Cvrcek 2011), transitions to adult- the potential to transform the field (Gutmann hood (Fussell & Furstenberg 2005), the racial et al. 2011). crossover in family complexity (Goldscheider It is equally important that we exploit the & Bures 2003), the living arrangements of chil- new potential to assess change over broad dren and the decline in coresidential support spans of time. The families of the developed for young mothers (Tolnay 2004, Short et al. world have been transformed over the past 2006), living arrangements of the elderly (Costa two centuries. In the mid-nineteenth century, 1999, Gratton & Gutmann 2010), the rise of elderly people ordinarily lived with their cohabitation and interracial marriage (Rosen- children, divorce was exceedingly rare, and feld 2006), and the impact of home ownership births to unmarried women were infrequent. on family structure (Collins & Margo 2004). Today, the great majority of the elderly in Within the next few years, we will have com- most developed countries reside alone or with parable microdata for both the pre-1920 period a spouse, divorce has gone up as much as and the post-1960 period in eight or nine de- 50-fold, and in a few countries most babies are veloped countries. The period between 1920 born to unmarried mothers. and 1960—a period of especially rapid family The magnitude of family change has been change in Europe and North America—poses just as dramatic as the more famous transitions greater challenges. In many countries, the mi- of fertility and mortality. In the United States, crolevel source materials from this period have

18.12 Ruggles SO38CH18-Ruggles ARI 20 April 2012 15:7

been lost, and confidentiality restrictions gen- theorists, however, the source of ideational erally limit access to the sources. As described change is unclear. The new norms and values above, however, a Canadian census data se- seem have a life of their own, and they spread ries is nearing completion, and population reg- and develop autonomously and on their own ister projects under way in Norway, Sweden, schedule; they are disconnected from the mas- and the Netherlands promise to make continu- sive structural changes that have transformed ous family data available for the entire century. and are continuing to transform the world. In the developing world, there are few publi- The transformation of family composition cally accessible microdata describing families coincided with declining authority of family before the 1960s, but there will be several sig- patriarchs. Before the mid-nineteenth century, nificant additions in the next few years. More- European and North American families were over, in many developing countries rapid family firmly controlled by the household head, who change first began in the late twentieth century, had the right to command the obedience of a time for which data resources are becoming his wife and children and to use corporal abundant. punishment to correct insubordination (Siegel 1996, pp. 2122–23). In the second half of the nineteenth century, legal restrictions on wife INVESTIGATING THE beating began to appear in Sweden, the United FAMILY TRANSITION States, and Britain, and in several countries Intergenerational coresidence in North Amer- married women acquired limited property ica and Northwest Europe began to drop in the rights (Calvert 1974, Emerton et al. 2005). In nineteenth century, and the change accelerated the twentieth century, patriarchal authority in the twentieth century (Ruggles 2009, 2010). diminished to the point that the very concept of This shift is just one facet of a broader trans- “Household Head” became obsolete; in 1980, formation of family behavior. In each country, the US Census Bureau abandoned the concept marital stability began to erode around 1900 to avoid offending respondents, and over the (Goode 1963). Over the course of the twen- following two decades most European census tieth century, Northwest Europe and North authorities followed suit (Ruggles & Brower America saw remarkable increases in lone par- 2003, Minnesota Population Center 2011). enthood, solitary residence, nonmarital fertil- I am persuaded that both the changes in ity, and cohabitation (Cherlin 1992, Sobotka & family structure and the decline of patriarchal Toulemon 2008). Taken together, this set of ideology are related to the massive structural changes represents a transition in family com- changes of the Industrial Revolution, which position with consequences as profound and transformed production and led to a tenfold far-reaching as the demographic transition of increase of per capita real income (Lindert mortality and fertility. 2004). In the early nineteenth century, house- Ideational theorists argue that family hold heads controlled the means of subsistence, changes reflect the erosion of traditional and residence decisions reflected the unequal norms about appropriate family behavior distribution of power. It was typically worth it (Lesthaeghe 1983). The rise of individualism for adult sons to obey their fathers; they of- and secularization, they argue, led to shifts in ten had no attractive alternatives. There were the moral code that allowed the major changes few employment opportunities available other in family behavior. At one level, the ideational than farm service; the prospect of inheriting interpretation is obviously correct: Current the farm or workshop was typically their best family behavior would have been prohibited by option. The lack of alternatives also reinforced nineteenth-century social mores, and the family patriarchal gender relationships. Wives submit- transition could never have happened without ted to the authority of their husbands or fathers an ideational shift. In the writing of most partly because they had no other way to subsist.

www.annualreviews.org • The Future of Historical Family Demography 18.13 SO38CH18-Ruggles ARI 20 April 2012 15:7

Across Northwest Europe and North To investigate the reasons for the breakup of America, wage labor opportunities emerged in the traditional patriarchal household, it is use- the nineteenth century. In the United States ful to think about incentives and constraints at in 1800, only 17% of men worked outside of the individual level, not the household level. farming; in 1900, 65% did so (Ruggles 2007). In Coale’s (1973) formulation of the precondi- Norway, the percentage of adult men in com- tions for fertility decline is instructive for the mercial and industrial occupations rose from family transition as well. Coale identified three 13% to 42% over the course of the century essential prerequisites for fertility limitation: (NAPP 2011). As more and better-paid wage (a) It must be within the “calculus of con- labor opportunities emerged, the incentive for scious choice,” (b) it must be advantageous, and sons to obey their fathers diminished. (c) effective techniques must be available. New opportunities also arose for women to We need to investigate all three of Coale’s work outside of the family economy and outside conditions for change. Using vast new data re- the control of husbands and fathers. In both sources, measurement methods that assess fam- Norway and the United States, for example, ily choices at the individual level, and analytic the percentage of women employed in wage la- strategies that assess variations across space and bor jobs other than domestic service doubled time, we can begin to dissect the transformation during the last three decades of the nineteenth of the family. We will soon finally have the data century (NAPP 2011). In places where job op- needed to draw a reliable map of family patterns portunities for women appeared, they created a in Europe before industrialization and to trace path to live independently and to escape from the chronology of associations between family abusive marriages. In the United States, there change and measures of secularization. We can was a remarkably consistent geographic associ- directly compare the family transition to the ation between the opening of female wage labor fertility transition at the levels of households, opportunities and the rise of marital instability communities, and regions. New longitudinal from 1880 to 1990 (Ruggles 1997). data will help us evaluate the material incen- The structural theories of family change tives for residence transitions at the individual that prevailed from Durkheim to Parsons level in a comparative framework. Complete- minimized the importance of incentives and count microdata will allow multilevel analysis constraints faced by individuals. According to of the impact of economic opportunity on the these interpretations, increasing social differen- family options available to individuals. tiation and new social institutions led families The new historical data will also allow to spontaneously shed their functions and sim- us to place Northwest Europe and North plify their structures to fit the needs of modern America in broad comparative context. There industrial society (Lamanna 2002, Parsons are increasing signs that the family transition 1955). Today, such theories seem teleological. may be global in scope. In recent decades, there Even if we reject functionalist explanations has been dramatic family change in Eastern and of change, however, the impact of structural Southern Europe (Lesthaeghe 2009). The de- factors on the family should not be understated. veloped countries of East Asia have also seen a The changing configuration of the family was sharp decline in intergenerational coresidence, made possible by a shift in the balance of a substantial rise of divorce, increased cohabi- power within the family created by the opening tation, and declining marriage (Chattopadhyay of economic opportunity. Individuals made & Marsh 1999, Lee 2006, Raymo et al. 2009, decisions about whether to stay with families or United Nations 2005). Family patterns in de- to leave; when wage labor opportunities arose, veloping countries are less clear. Marital insta- children and wives had a path to escape patri- bility is increasing in much of the developing archal discipline. But to take advantage of that world (Heggeness 2010). There is some evi- path, they had to be willing to defy authority. dence of a drop in intergenerational families,

18.14 Ruggles SO38CH18-Ruggles ARI 20 April 2012 15:7

but there are also conflicting data, and patriar- indelibility of family norms and values (e.g., chal authority remains strong in most countries Therborn 2004) are beginning to lose ground. (Palloni 2001, Ruggles & Heggeness 2008). The question of convergence, however, is in- It is too early to declare that convergence herently a historical one. By empirically de- theorists were correct all along and that the scribing sequences of change across time and entire world will eventually come to resemble space, we can evaluate theory and provide his- the West. As evidence of family change around torical context for understanding the ongoing the world accumulates, however, theories of the transformation of family demography.

DISCLOSURE STATEMENT The author is not aware of any affiliations, memberships, funding, or financial holdings that might be perceived as affecting the objectivity of this review.

LITERATURE CITED Alter G, Mandemakers K, Gutmann MP. 2009. Defining and distributing longitudinal historical data in a general way through an intermediate structure. Hist.Soc.Res.34:78–114 Berkner L. 1972. The stem family and the developmental cycle of the peasant household: an eighteenth century Austrian example. Am. Hist. Rev. 77:398–418 Bongaarts J, Burch TK, Wachter KW, eds. 1987. Family Demography: Methods and Their Application. Oxford: Clarendon Breschi M, Forasin A, Manfredini M, Zaccchigna M. 2009. Family composition and remarriage in pre- transitional Italy: a comparative study. Eur. J. Popul. 25:277–96 Burch TK. 1967. The size and structure of families: a comparative analysis of census data. Am. Sociol. Rev. 32:347–63 Cain M, McNicoll G. 1988. Population growth and agrarian outcomes. In Population, Food and Rural Develop- ment, ed. RD Lee, WB Arthur, AC Kelley, G Rogers, pp. 101–11. Oxford: Clarendon Calvert R. 1974. Criminal and civil liability in husband-wife assaults. In Violence in the Family, ed. SK Steinmetz, MA Straus, pp. 88–91. New York: Harper & Row Carter S, Gartner SS, Haines M, Olmsted A, Sutch R, Wright G, eds. 2006. Historical Statistics of the United States: Millennial Edition. Cambridge: Cambridge Univ. Press Chattopadhyay A, Marsh R. 1999. Changes in living arrangement and familial support for the elderly in Taiwan: 1963–1991. J. Comp. Fam. Stud. 30:523–37 Cherlin AJ. 1992. Marriage, cohabitation, sexual behavior, and childbearing in North America. In The Peopling of the Americas, Proceedings of the IUSSP Conference at Veracruz, Mexico, May 17–23, 1992, 3:223–43. Liege,` Belg.: Int. Union Sci. Study Popul. Cherlin AJ. 2003. Family demography. In Encyclopedia of Population, ed. P Demeny, G McNicoll, pp. 355–61. New York: Macmillan Ref. Coale AJ. 1973. The demographic transition reconsidered. In Proceedings of the International Population Conference, Liege 1973, 1:55–72. Liege,` Belg.: Int. Union Sci. Study Popul. Collins WJ, Margo RA. 2004. Race, home ownership, and family structure in twentieth-century America. In What Has Happened to the Quality of Life in the Advanced Industrialized Nations?, ed. EN Wolff, pp. 187–213. Northampton, MA: Edward Elgar Costa DL. 1999. A house of her own: old age assistance and the living arrangements of older nonmarried women. J. Public Econ. 72:39–59 Cvrcek T. 2011. U.S. marital disruptions and their economic and social correlates, 1860–1948. J. Fam. Hist. 36:142–58 Czap P. 1982. The perennial multiple family household: Mishino, Russia, 1782–1858. J. Fam. Hist. 7:5–26 Devas CS. 1886. Studies of Family Life: A Contribution to Social Science. London: Burns & Oates

www.annualreviews.org • The Future of Historical Family Demography 18.15 SO38CH18-Ruggles ARI 20 April 2012 15:7

De Vos S, Palloni A. 1989. Formal models and methods for analyzing kinship and household organization. Popul. Index 55:174–98 Durkheim E.´ 1888. Introduction a` la sociologie de la famille. Ann. Fac. Lett. Bordx. 10:257–82 EHPS-Net. 2011. European Historical Population Samples Network (EHPS-Net). Amsterdam: Int. Inst. Soc. Hist. http://www.iisg.nl/hsn/news/ehps-net.php Elder GH Jr. 1978. Family history and the life course. J. Fam. Hist. 2:279–304 Emerton R, Adams K, Byrnes A, Connors J. 2005. International Women’s Rights Cases. London: Cavendish Fauve-Chamoux A. 1984. Les structures familiales au royaume des familles-souches: Esparros Ann. Econ. Soc. Civiliz. 3:513–28 Fauve-Chamoux A, Ochiai E. 2009. The Stem Family in Eurasian Perspective: Revisiting House Societies, 17th-20th Centuries. Bern: Peter Lang Fussell E, Furstenberg FF. 2005. The transition to adulthood during the 20th century: differences by race, nativity and gender. In On The Frontier of Adulthood: Theory, Research, and Public Policy, ed. RA Settersten, FF Furstenberg, RG Rumbaut, pp. 29–75. Chicago: Univ. Chicago Press General Register Office. 1968. Sample Census 1966, England and Wales: Household Composition Tables. London: H.M.S.O. Gibbon P, Curtin C. 1978. The stem family in Ireland. Comp. Stud. Soc. Hist. 20:429–53 Glass DV. 1966. London Inhabitants Within the Walls 1695. Leicester: London Rec. Soc. Glick PC. 1941. Types of families: an analysis of census data. Am. Sociol. Rev. 6:830–38 Glick PC. 1988. Fifty years of family demography: a record of social change. J. Marriage Fam. 50:861–73 Goeken R, Huynh L, Lenius T, Vick R. 2011. New methods of census record linking. Hist. Methods 44:7– 14 Goldscheider FK, Bures RM. 2003. The racial crossover in family complexity in the United States. Demography 40:569–87 Goode WJ. 1963. World Revolution and Family Patterns. Glencoe, IL: Free Press Gratton B, Gutmann MP. 2010. Emptying the nest: older men in the United States, 1880–2000. Popul. Dev. Rev. 36:331–56 Greenfield SM. 1961. Industrialization and the family in sociological theory. Am. J. Sociol. 67:312–22 Gruber S, Szołtysek M. 2011. Stem families, joint families, and the European pattern: How much of a reconsideration do we need? MPIDR Work. Pap. WP 2011–001, Max Planck Inst. Demogr. Res., Rostock, Ger. Gutmann MP, Deane GD, Merchant ER, Sylvester K. 2011. Navigating time and space in population studies. Int. Stud. Popul. 9:1–17 Hajnal J. 1965. European marriage patterns in perspective. In Population in History: Essays in Historical Demog- raphy, ed. DV Glass, DEC Eversley, pp. 101–43. London: Arnold Hajnal J. 1982. Two kinds of preindustrial household formation system. Popul. Dev. Rev. 8:449–94 Hartman MS. 2004. The Household and the Making of History: A Subversive View of the Western Past. New York: Cambridge Univ. Press Heggeness M. 2010. Essays on marital instability, household behavior, and social policy in developing countries.PhD diss., Univ. Minn. Appl. Econ. I-CeM. 2011. Integrated Census Microdata (I-CeM) Project. Essex, UK: Dep. Hist., Univ. Essex. http://www.essex.ac.uk/history/research/ICeM/default.aspx IPUMS. 2011. Integrated Public Use Microdata Series. Minneapolis: Minn. Popul. Cent., Univ. Minn. http://www.ipums.org Ja˚stad H. 2011. Northern co-residence across generations in Northernmost Norway during the last part of the nineteenth century. PhD Art thesis. Univ. Tromsø. 77 pp. Kaser K. 1994. The Balkan joint family: redefining a problem. Soc. Sci. Hist. 18:243–69 Kertzer DI. 1989. The joint family household revisited: demographic constraints and household complexity in the European past. J. Fam. Hist. 14:1–15 Kok J. 2007. Principles and prospects of the life course paradigm. Ann. D´emogr. Hist. 2007:203–30 Lamanna MA. 2002. Emile´ Durkheim on the Family. Thousand Oakes, CA: Sage Laslett P. 1965. The World We Have Lost. New York: Scribner Laslett P. 1969. Size and structure of the household in England over three centuries. Popul. Stud. 23:199– 223

18.16 Ruggles SO38CH18-Ruggles ARI 20 April 2012 15:7

Laslett P. 1970. The comparative history of household and family. J. Soc. Hist. 4:75–87 Laslett P. 1982. Family and household as work group and kin group: areas of traditional Europe compared. In Family Forms in Historic Europe, ed. R Wall, J Robin, P Laslett, pp. 513–64. Cambridge, UK: Cambridge Univ. Press Laslett P, Wall R. 1972. Household and Family in Past Time: Comparative Studies in the Size and Structure of the Domestic Group over the Last Three Centuries in England, France, Serbia, Japan, and Colonial North America. Cambridge, UK: Cambridge Univ. Press Lee JZ, Campbell CD. 2011. China Multi-Generational Panel Dataset, Liaoning (CMGPD-LN ), 1749–1909. Ann Arbor, MI: Inter-Univ. Consort. Polit. Soc. Res. Lee YJ. 2006. Risk factors in the rapidly rising incidence of divorce in Korea. Asian Popul. Stud. 2:113–31 Le Play F. 1855. Les Ouvriers Europ´eens: Etudes´ sur les Travaux, la Vie Domestique, et la Condition Morale des Populations Ouvi`eres de l’Europe. Paris, Fr.: Impr. Imp. Le Play F. 1871. L’organisation de la Famille Selon le Vrai Mod`ele Signal´e par L’histoire de Toutes les Races et de Tous les Temps. Tours, Fr.: Mame Le Play F. 1872. The Organization of Labor: In Accordance with Custom and the Law of the Decalogue.Transl. Gouverneur Emerson, MD. Philadelphia: Claxton, Remsen & Haffelfinger Lesthaeghe R. 1983. A century of demographic and cultural change in Western Europe: an exploration of underlying dimensions. Popul. Dev. Rev. 9:411–35 Lesthaeghe R. 2009. The unfolding story of the second demographic transition. Popul. Dev. Rev. 36:211– 51 Lindert P. 2004. Growing Public: Social Spending and Economic Growth Since the Eighteenth Century. Cambridge, UK: Cambridge Univ. Press Litwak E. 1960. Geographic mobility and extended family cohesion. Am. Sociol. Rev. 25:385–94 Macfarlane A. 1986. Marriage and Love in England: Modes of Reproduction 1300–1840. Oxford: Blackwell Macfarlane A. 1987. The Culture of Capitalism. Oxford: Blackwell Minnesota Population Center 2011. Integrated Public Use Microdata Series, International: Version 6.1. Minneapolis: Univ. Minn. Moring B. 2009. Land, inheritance, and the Finnish stem family. See Fauve-Chamoux & Ochiai 2009, pp. 173– 203 Mosaic. 2011. Recovering Surviving Census Records to Reconstruct Population, Economic, and Cultural History. Rostock, Ger: Max Planck Inst. Demogr. Res. http://www.censusmosaic.org Murphy M. 2011. Long-term effects of the demographic transition on family and kinship networks in Britain. Popul. Dev. Rev. 37:55–80 NAPP. 2011. North Atlantic Population Project. Minneapolis: Minn. Popul. Cent., Univ. Minn. http://www. nappdata.org Nevin M. 2002. The influence of the wider kin group on individual life-course transitions: results from the Pays de Herve (Belgium), 1846–1900. Contin. Change 17:405–35 Nixon JW. 1952. On the size, constitution and housing standards of households in England and Wales, 1931 and 1951. Rev. Inst. Int. Stat. 20:135–47 Norwegian Historical Data Centre. 2011. Historical Population Registers. Tromsø, Nor.: Univ. Tromsø. http://www.rhd.uit.no/nhdc/hpr.html Ogburn WE. 1933. The family and its functions. In Recent Social Trends in the United States: Report of the President’s Research Committee on Social Trends, pp. 661–708. New York: McGraw-Hill Palloni A. 2001. Living arrangements of older persons. In Population Bulletin of the United Nations: Population Ageing and Living Arrangements of Older Persons: Critical Issues and Policy Responses, Special Issue 43/43, pp. 54–110. New York: United Nations Parsons T. 1944. The social structure of the family. In The Family: Its Function and Destiny, ed. RN Anshen, pp. 173–201. New York: Harper Parsons T. 1955. The American family. In Family Socialization and Interaction Process, ed. T Parsons, RF Bales, pp. 3–33. Glencoe, IL: Free Press Plakans A, Wetherell C. 2005. The Hajnal line and Eastern Europe. In Marriage and the Family in Eurasia: Perspectives on the Hajnal Hypothesis, ed. T Engelen, AP Wolff, pp. 105–28. Amsterdam: Aksant Acad.

www.annualreviews.org • The Future of Historical Family Demography 18.17 SO38CH18-Ruggles ARI 20 April 2012 15:7

POPLINK. 2011. Information om Hantering av personuppgifter i databasen POPLINK. Umea˚, Sweden: Demogr. Database, Univ. Umea˚. http://www.ddb.umu.se/databaser/poplink Post W, Van Poppel F, Van Imhoff E, Kruse E. 1997. Reconstructing the extended kin-network in the Netherlands with genealogical data: methods, problems, and results. Popul. Stud. 51:263–78 Rosenfeld MJ. 2006. Young adulthood as a factor in social change in the United States. Popul. Dev. Rev. 32:27–51 Raymo J, Iwasawa M, Bumpass L. 2009. Cohabitation and family formation in Japan. Demography 46:785– 803 Ruggles S. 1987. Prolonged Connections: The Rise of the Extended Family in Nineteenth Century England and America. Madison: Univ. Wis. Press Ruggles S. 1994. The transformation of American family structure. Am. Hist. Rev. 99:103–28 Ruggles S. 1997. The rise of divorce and separation in the United States, 1880–1990. Demography 34:455–66 Ruggles S. 2007. The decline of intergenerational coresidence in the United States, 1850–2000. Am. Sociol. Rev. 72:962–89 Ruggles S. 2009. Reconsidering the Northwest European family system: living arrangements of the aged in comparative historical perspective. Popul. Dev. Rev. 35:249–73 Ruggles S. 2010. Stem families and joint families in comparative historical perspective. Popul. Dev. Rev. 35:249– 73 Ruggles S, Brower S. 2003. The measurement of family and household composition in the United States, 1850–1999. Popul. Dev. Rev. 29:73–101 Ruggles S, Heggeness M. 2008. Intergenerational families in developing countries. Popul. Dev. Rev. 34:253– 81 Ruggles S, Roberts E, Sarkar S, Sobek M. 2011. The North Atlantic Population Project: progress and prospects. Hist. Methods 44:1–6 Sager EW, Baskerville PA. 2010. Canadian historical research and pedagogy: a view from the perspective of the Canadian Century Research Infrastructure. Can. Hist. Rev. 91:533–51 Schurer¨ K. 2003. Introduction: household and family in past time further explored. Contin. Change 18:9–21 Short SE, Goldscheider FK, Torr BM. 2006. Less help for mother: the decline in coresidential female support for the mothers of young children, 1880–2000. Demography 43:617–29 Siegel RB. 1996. ‘The rule of love’: wife beating as prerogative and privacy. Yale Law J. 105:2117–207 Sobek M, Cleveland L, Flood S, Hall PK, King ML, et al. 2011. Big data: large-scale historical infrastructure from the Minnesota Population Center. Hist. Methods 44:61–68 Sobotka T, Toulemon L. 2008. Changing family and partnership behaviour: common trends and persistent diversity across Europe. Demogr. Res. 19:85–138 Sogner S. 2009. The Norwegian stem family: myth or reality? See Fauve-Chamoux & Ochiai 2009, pp. 151– 72 Sussman MB. 1959. The isolated nuclear family: fact or fiction. Soc. Probl. 6:333–40 Szołtysek M. 2009. Life-cycle service and family systems in the rural countryside: a lesson from historical East-Central Europe. Ann. D´emogr. Hist. 117:53–94 Therborn G. 2004. Between Sex and Power: Family in the World 1900–2000. London: Routledge Tolnay SE. 2004. The living arrangements of African American and immigrant children, 1880–2000. J. Fam. Hist. 29:421–45 United Nations. 2001. Principles and Recommendations for a Vital Statistics System, Revision 2.NewYork:United Nations United Nations. 2005. Living Arrangements of Older Persons Around the World. New York: United Nations Van Baelen H, Matthijs K. 2007. Comparing two methods to reconstruct the kin network outside the household with nineteenth-century Antwerp data. J. Fam. Hist. 32:149–60 Verdon M. 1998. Rethinking Households: An Atomistic Perspective on European Living Arrangements. London/New York: Routledge Wall R. 2009. Ideology and reality of the stem family in the writings of Fred´ eric´ Le Play. See Fauve-Chamoux & Ochiai 2009, pp. 53–80 Wheaton R. 1975. Family and kinship in Western Europe: the problem of the joint family household. J. Interdis. Hist. 5:601–28

18.18 Ruggles SO38CH18-Ruggles ARI 20 April 2012 15:7

Wolf D. 1994. The elderly and their kin: patterns of availability and access. In Demography of Aging, ed. LG Martin, SH Preston, pp. 146–94. Washington, DC: Natl. Acad. Press World Bank. 2011. World Development Indicators 2011. Online. Washington, DC: World Bank. http://data. worldbank.org/indicator/SP.DYN.CDRT.IN Wrigley EA. 1969. Population and History. London: Weidenfeld & Nicolson Wrigley EA, Schofield RS. 1981. The Population History of England 1541–1871: A Reconstruction. London: Edward Arnold

www.annualreviews.org • The Future of Historical Family Demography 18.19