Zeitschrift für interdisziplinäre Geschichte, xlii:4 (Frühling, 2012), 503–517.

EINFÜHRUNG

George C. Ändern, Myron P. Gutmann, Susan Hautaniemi
Leonard, und Emily R. Händler
Einführung: Longitudinal Analysis of Historical-
Demographic Data The ªeld of historical demography
embraces the aspirations and challenges of both demography and
Geschichte. It analyzes the vital processes that unfold within individ-
ual life courses—particularly fertility, nuptiality, migration, Und
mortality—while tracking aggregate changes in those processes
im Laufe der Zeit. It identiªes the social, cultural, wirtschaftlich, und politisch
correlates of individual demographic outcomes and historical-

George C. Alter is Professor of History and Research Professor, Zentrum für Bevölkerungsstudien,
Universität von Michigan; Director, Inter-university Consortium for Political and Social Re-
suchen; and Director, Population Institute for Research and Training, Indiana University. Er
is the author of, with Muriel Neven and Michel Oris, “Economic Change and Differential
Fertility in Rural Eastern Belgium, 1812 to 1875,” in Noriko Tsuya et al. (Hrsg.), Prudence and
Pressure: Reproduktion und menschliche Handlungsfähigkeit in Europa und Asien, 1700–1900 (Cambridge, Masse.,
2010), 195–216; with Isabelle Devos and Alison Kvetko, “Completing Life Histories with Im-
puted Exit Dates: A Method for Historical Data from Passive Registration Systems,” Popula-
tion, 64 (2009), 293–318.

Myron P. Gutmann is Professor of History, and Information and Research Professor of
Population Studies, Universität von Michigan; Assistant Director, Nationale Wissenschaftsstiftung.
He is the author of Towards the Modern Economy: Early Industry in Europe, 1500–1800 (Neu
York, 1988); editor of, with Glenn D. Deane, Emily R. Händler, and Kenneth M. Sylvester,
Navigating Time and Space in Population Studies (New York, 2011).

Susan Hautaniemi Leonard is Research Afªliate, Zentrum für Bevölkerungsstudien, and Assistant
Research Scientist, Interuniversitäres Konsortium für Politik- und Sozialforschung, Universität
von Michigan. She is the author of, with Myron Gutmann and Glenn D. Deane, “Household
and Farm Transitions in Environmental Context,” Population and Environment, XXXII (2011),
287–317; with Myron Gutmann, “‘The Farm Should Provide Our Retirement’; Land-Use
Plans in the Aging Farm Population of the U.S. Great Plains,” Great Plains Research, XVI
(2006), 181–193.

Emily R. Merchant is a doctoral student, Dept. of History, and research area specialist,
Interuniversitäres Konsortium für Politik- und Sozialforschung, Universität von Michigan. She is
the author of, mit Myron P. Gutmann et al., "Einführung,” in Gutmann et al. (Hrsg.), Navi-
gating Time and Space in Population Studies (New York, 2011), 1–17; with Melannie D. Hart-
man et al., “Impact of Historical Land Use Changes on Greenhouse Gas Exchange in the U.S.
Great Plains, 1883–2003,” Ecological Applications, XXI (2011), 1105–1119.

The work reported in these articles was supported by grant Number R25 HD040525
from The Eunice Kennedy Shriver National Institute of Child Health and Human Develop-
ment (nichd). Myron P. Gutmann’s work was supported, in part, by the National Science
Foundation of the United States. Any opinions, ªndings, conclusions, or recommendations
expressed in this material belong to the authors exclusively and do not necessarily reºect the
views of the National Science Foundation or the nichd.

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u

/
J
ich

N
H
A
R
T
ich
C
e
–
P
D

F
/

4
2
4
5
0
3
1
6
9
9
3
2
2

/
J
ich

N
H
_
A
_
0
0
3
0
2
P
D

B
j
G
u
e
S
T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

504 | ALTER, GUT MANN, LEONARD, AND M ERCH ANT

life-history datasets covering broad spans of

demographic transitions, seeking explanations that are universal
and particular, both macro and micro. For the last ªfty years, diese
pursuits have promoted and been facilitated by a burgeoning of
individual
Zeit
and many parts of the globe. This valuable resource has allowed
researchers to examine demographic processes within lives and
across space and time. More recent developments in database soft-
ware and statistical methods have promoted increasingly sophisti-
cated analysis of these datasets, including linkage of families across
generations and identiªcation of complex causal mechanisms. Der
availability of these datasets and analytical capabilities has encour-
aged researchers to ask new questions and re-open old controver-
sies about core demographic processes and historical transitions in
ihnen, while scholars in ªelds ranging from anthropology to epide-
miology are extending the techniques of historical demography to
ever more populations and research topics.1

This special issue of the Journal of Interdisciplinary History pre-
sents six examples of recent forays into the longitudinal analysis of
historical-demographic data, written by participants of summer
workshops held in 2006 Und 2007 at the Inter-university Consor-
tium for Political and Social Research (icpsr), the University of
Michigan. These studies focus on Utah (Vereinigte Staaten), Sart (Bel-
gium), Bologna (Italien), and Madrid (Spanien) in the nineteenth and
early twentieth centuries, attending to the particularities of these
settings and discussing the wider applicability of their ªndings.2

tracking the demographic transition Many of the articles
presented herein contribute to one of the earliest and most long-

1 Examples of the extension of historical-demographic methods to other ªelds include Alan
Bittles, Michael Murphy, and David Reher, “Inherited Dimensions of Human Populations in
the Past,” Human Nature—An Interdisciplinary Biosocial Perspective, XIX (2008), 1–6; Natalia S.
Gavrilova et al., “Does Exceptional Human Longevity Come with a High Cost of Infertility?
Testing the Evolutionary Theories of Aging,” Annals of the New York Academy of Sciences,
MXIX (2004), 513–517; Rudi G. J. Westendorp and Thomas B. L. Kirkwood, “Human Lon-
gevity at the Cost of Reproductive Success,” Nature, CCCXCVI (1998), 743–746.
2 This intensive four-week course introduces current and potential users of the growing
number of individual-level historical-longitudinal databases to the theoretical and substantive
demographic questions that can be approached with this type of data; the specialized tech-
niques used to create these datasets from surviving administrative, genealogical, and ecclesias-
tical records; and the subtle challenges that historical data present to statistical analysis.
Participants approach important issues in historical demography by developing research pro-
jects that use multivariate longitudinal analysis to exploit the rich capabilities of large individ-
ual-level life-history datasets from Europe, Asien, and North America. It began in 2006 und hat
been held biennially since 2007.

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u

/
J
ich

N
H
A
R
T
ich
C
e
–
P
D

F
/

4
2
4
5
0
3
1
6
9
9
3
2
2

/
J
ich

N
H
_
A
_
0
0
3
0
2
P
D

B
j
G
u
e
S
T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

EINFÜHRUNG | 505

standing goals of historical demography—understanding and ac-
counting for the demographic transition that occurred in Europe
and North America between the mid-eighteenth and mid-
twentieth centuries. As a description of the historical experience
of Europe and North America, demographic transition, in its
broadest terms, refers to a secular decline in mortality, accompa-
nied and followed by a decline in fertility. The result includes a
shift from a high-pressure equilibrium with a younger age struc-
ture to a low-pressure equilibrium with an older age structure,
generating dramatic but temporary population growth. Analyses
of cross-sectional aggregate historical data reveal changes in mor-
tality and fertility rates over time and identify the structural corre-
lates of varying levels of mortality and fertility. More recent
multivariate longitudinal analyses of individual-level life-history
data have added considerable nuance to our understanding of pre-
and posttransitional demographic regimes and of the transition it-
self by revealing variations in the experience of transition and as-
sessing the individual- and household-level determinants of such
demographic processes as migration, Hochzeit, Geburt, Und
death.3

Historical demographers have used longitudinal data and
methods to examine the unevenness of demographic transition
within societies, asking, Zum Beispiel, which families continued to
experience infant death as mortality declined at the societal level,
what characteristics made some couples more likely to limit their
fertility in advance of widespread fertility decline, and how mor-
tality and fertility were related at the household level. Scholars
have also utilized longitudinal fertility data to contest and compli-
cate Henry’s early description of the fertility transition as a shift
from “natural fertility,” non-parity-speciªc fertility behavior, Zu
“controlled fertility,” the cessation of childbearing when an ideal
number of offspring is achieved. Drawing on birth-interval data,
recent studies have accounted for differences in fertility levels
among natural-fertility populations, identiªed deliberate fertility

3 Aggregate cross-sectional analysis is exempliªed by the Princeton European Fertility Proj-
ect, the results of which are summarized in Ansley J. Coale und Susan Cotts Watkins (Hrsg.),
Der Rückgang der Fruchtbarkeit in Europa: The Revised Proceedings of a Conference on the Princeton European
Fertility Project (Princeton, 1986). A notable example of individual-level longitudinal analysis is
the work of the Eurasia Project: Tommy Bengtsson et al., Leben unter Druck: Mortality and Liv-
ing Standards in Europe and Asia, 1700–1900 (New York, 2004); Noriko O. Tsuya et al., Pru-
dence and Pressure: Reproduktion und menschliche Handlungsfähigkeit in Europa und Asien, 1700–1900 (New York,
2010).

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u

/
J
ich

N
H
A
R
T
ich
C
e
–
P
D

F
/

4
2
4
5
0
3
1
6
9
9
3
2
2

/
J
ich

N
H
_
A
_
0
0
3
0
2
P
D

B
j
G
u
e
S
T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

506 | ALTER, GUT MANN, LEONARD, AND M ERCH ANT

control prior to fertility transition, examined the role of increased
birth spacing in effecting fertility decline, and analyzed changes in
the determinants of fertility between the pre- and posttransitional
periods.4

But change in number of children born is not the whole
Geschichte: Mason has argued that the fertility transition also included a
shift from postnatal to prenatal methods of controlling family size
and composition. Historical demographers have tested this asser-
tion by using life-history data to explore the dynamics of infanti-
cide and child abandonment, identifying which children were
most at risk and examining how the presence of other household
members increased or decreased their vulnerability. Such work has
demonstrated that the demographic transition was multiple rather
than singular: Demographic regimes and changes in them varied
not only by region but also by class and even gender within re-
gions. The articles that follow this introduction use life-history
data and longitudinal methods to address many of these questions,
demonstrating the subtle concerns that these data and methods are
particularly suited to answering and suggesting the wealth of
knowledge yet to be revealed through the rigorous analysis of his-
torical sources.5

4 On the unevenness of the mortality transition, see Katherine A. Lynch and Joel B. Grün-
house, “Risk Factors for Infant Mortality in Nineteenth-Century Sweden," Bevölkerung
Studien, XLVIII (1994), 117–133. On the idea of an “early warning system” for fertility de-
cline, see Gutmann and Watkins, “Socio-Economic Differences in Fertility Control: Is There
an Early Warning System at the Village Level?„Europäisches Journal der Bevölkerung, VI (1990), 69–
101. For an example of work on the relationship between mortality and fertility within house-
holds, see John Knodel, “Child Mortality and Reproductive Behaviour in German Village
Populations in the Past: A Micro-Level Analysis of the Replacement Effect," Bevölkerung
Studien, XXXVI (1982), 177–2000. For Henry’s description of the fertility transition, see Louis
Henry, „Einige Daten zur natürlichen Fruchtbarkeit,„ Eugenics Quarterly, VIII (1961), 81–91. For analysis
of fertility differentials among pre-transition populations, see Knodel, Demographic Behavior in
the Past: A Study of Fourteen German Village Populations in the Eighteenth and Nineteenth Centuries
(New York, 1988). For analysis of pretransitional fertility control, see Bengtsson and Martin
Dribe, “Deliberate Control in a Natural Fertility Population: Südschweden, 1766–1864,”
Demographie, LXIII (2006), 727–746. On the role of spacing in the fertility transition, sehen
Knödel, “Starting, Stopping, and Spacing during the Early Stages of Fertility Transition: Der
Experience of German Village Populations in the 18th and 19th Centuries,” ebenda., XXIV
(1987), 143–162; Douglas L. Anderton und Lee L. Bohne, “Birth Spacing and Fertility Limita-
tion: A Behavioral Analysis of a Nineteenth Century Frontier Population,” ebenda., XXII
(1985), 169–183. See also the special issue, “Before the Pill: Preventing Fertility in Western
Europe and Quebec,” Zeitschrift für interdisziplinäre Geschichte, XXXIV (2003), 141–314.
5 On the shift from postnatal to prenatal fertility control, see Karen Oppenheim Mason,
“Explaining Fertility Transitions,„ Demografie, XXXIV (1997), 443–454. For analysis of
postnatal control in pre-transition populations, see Tsuya et al., Besonnenheit und Druck. Der

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u

/
J
ich

N
H
A
R
T
ich
C
e
–
P
D

F
/

4
2
4
5
0
3
1
6
9
9
3
2
2

/
J
ich

N
H
_
A
_
0
0
3
0
2
P
D

B
j
G
u
e
S
T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

EINFÜHRUNG | 507
historical data Life-history databases provide researchers
with data that literally took lifetimes and generations to collect.
Most are compiled from historical
sources originally created
for other purposes—typically ecclesiastical, administrative, Und
genealogical—often using such specialized techniques as family re-
constitution and back projection. Analyzing these sources there-
fore poses unique challenges and requires detailed knowledge
about the past societies and institutions that produced them and
the methods used to transform them into individual-level longitu-
dinal databases. The comprehensiveness and quality of historical
data depend upon both the administrative capabilities of the insti-
tutions that produced the original documents and the goals and
needs motivating them to collect information. The nature of the
Daten, im Gegenzug, determines how researchers can use them. For exam-
Bitte, databases produced through the reconstitution of families
from records of baptisms, Ehen, and burials provide precise
dates of vital events but no observation between events; those pro-
duced through the linkage of frequent censuses offer more obser-
vations of individuals but not exact dates of vital events. Continu-
ous registers of population, kept in several European and Asian
localities, typically combine the strengths of vital registers and fre-
quent censuses but are rarely available for the long swaths of time
necessary to analyze historical transitions and intergenerational
processes.6

dennoch, if used with the appropriate caution, each type
of database can yield important information about historical-
that both challenge and reªne demo-
demographic regimes
graphic-transition theory. Darüber hinaus, as more and more such data-
bases have become available, researchers have been able to supple-

multiplicity of the fertility transition in Europe is emphasized and well analyzed in John R.
Gillis, Louise A. Tilly, und David Levine (Hrsg.), Die europäische Erfahrung sinkender Fruchtbarkeit,
1850–1970 (New York, 1992).
6 For discussions of back projection and family reconstitution, sehen, jeweils, James
Oeppen, „Rückprojektion und Umkehrprojektion: Members of a Wider Class of Constrained
Projection Models,„Bevölkerungsstudien, XLVII (1993), 245–267, and E. Anthony Wrigley and
Roger S. Schoªeld, Die Bevölkerungsgeschichte Englands, 1541–1871: Eine Rekonstruktion (New York,
1981). The publications of the Eurasia group are a prominent example of this type of cross-
cultural analysis (see n. 3). Each source of life-history data carries speciªc limitations: Geneal-
ogies tend to under-represent those with no children and those of lower socioeconomic
Status; population registers do not always record all of the out-migrations; and family reconsti-
tution privileges those who do not migrate.

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u

/
J
ich

N
H
A
R
T
ich
C
e
–
P
D

F
/

4
2
4
5
0
3
1
6
9
9
3
2
2

/
J
ich

N
H
_
A
_
0
0
3
0
2
P
D

B
j
G
u
e
S
T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

508 | ALTER, GUT MANN, LEONARD, AND M ERCH ANT

ment knowledge gained from one type of source with information
from other types, and to draw cross-cultural comparisons by ana-
lyzing analogous data from different societies.

longitudinal analysis Life-history datasets include continu-
ously or periodically updated information for each individual over
Zeit, and therefore require speciªcally longitudinal analytical
Methoden. Each article in this issue employs one of two multivariate
longitudinal regression models: Cox proportional hazards and dis-
crete-time event history. These methods, which originated in en-
gineering and biomedical research in the 1960s and 1970s, war
adopted by social scientists for use with duration or time-to-event
data in the early 1980s. The dependent variable is whether or
when a particular event occurs, such as childbirth or death; results
indicate the effects of independent variables on the instantaneous
probability of experiencing the event in question (the hazard rate)
across the period of analysis. The baseline hazard reºects the haz-
ard rate for an individual possessing values of 0 for all of the inde-
pendent variables as a function of the time at risk for the event in
question (analysis time). Although values of independent variables
can change during the period of analysis, their effects are assumed
to be constant and in proportion to the baseline hazard.7

The shape of the hazard function is not speciªed in advance
when using the Cox model, allowing the baseline hazard to vary
continuously. Cox models are therefore well suited to datasets
based on continuous observations in which the exact dates of vital
events are known, such as those constructed from family reconsti-
tution or population registers. These models are less appropriate
for datasets compiled from periodic observations, which can reveal
only that an event occurred (or did not occur) between one obser-
vation and the next. For this type of data, discrete-time event-
history analysis divides the time under study into “spells” or “epi-
sodes” equal to the period between observations, with the as-
sumption that the baseline hazard remains constant within each
spell. Results of both Cox and discrete-time models indicate

Jay D. Teachman, “Analyzing Social Processes: Life Tables and Proportional Hazards
7
Models,” Social Science Research, XII (1983), 263–301; Paul Allison, “Discrete-Time Methods
for the Analysis of Event Histories,” Sociological Methodology, xiii (1982), 61–98; David Cox,
„Regressionsmodelle und Sterbetafeln (with Discussion),” Journal of the Royal Statistical Society B,
XXXIV (1972), 187–220. These methods can also accommodate nonproportional hazards.

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u

/
J
ich

N
H
A
R
T
ich
C
e
–
P
D

F
/

4
2
4
5
0
3
1
6
9
9
3
2
2

/
J
ich

N
H
_
A
_
0
0
3
0
2
P
D

B
j
G
u
e
S
T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

EINFÜHRUNG | 509

which individual attributes increase or decrease the hazard rate at
any given time, and quantify the relative risk associated with that
attribute.

Longitudinal analysis establishes causal relationships because
the state of attributes is speciªed prior to the occurrence of the
event in question. In the Cox model, the values of covariates that
change over time—such as age, Beruf, or parity—may be
updated continuously; in discrete-time models they are updated at
the beginning of each spell. Databases rich with individual attrib-
utes that are monitored continuously or frequently can produce
ªne-grained analyses of the relationships between these attributes
and vital events. Statistical software readily handles such analysis,
but complex methods of data management are often required to
translate data from the relational databases in which they are stored
to the rectangular format required by statistical packages. Der
complexity increases with the number of time-varying covariates,
especially when those covariates refer to other individuals in a
household. Another advantage and challenge of longitudinal mod-
els is their ability to accommodate incomplete records and indi-
viduals for whom the event in question does not occur during the
period of analysis. Valuable information could be lost if, for exam-
Bitte, only individuals who died under observation were included in
mortality analysis or if women who died, moved away from the
observation area, or were widowed or divorced before reaching
the end of childbearing age were excluded from fertility analysis.
In longitudinal models, individuals can be included for the length
of time during which they are under observation, and excluded,
or censored, when observation ends, regardless of whether the
event in question has occurred. Use of these observations allows
researchers to retain information, maintain sample sizes, and avoid
sample-selection bias.8

8 The editors of this issue are currently working with representatives of life-history data-
bases throughout the world to develop a standard format for storing these types of data that
will allow for the development of a common set of programs to rectangularize data for various
forms of demographic analysis. Such standardization will also promote comparative analysis
across databases. Ändern, Kees Mandemakers, and Gutmann, “Deªning and Distributing Longi-
tudinal Historical Data in a General Way through an Intermediate Structure,” Historical Social
Forschung, XXXIV (2009), 78–114. Successful censoring requires knowledge of the date at
which observation ended. This date must be determined with care to avoid introducing new
biases. Death clearly ends observations and, in data derived from population registers, dates of
migration are often known. Determining the end of observation is much more challenging in
family-reconstitution datasets, which include only births, Ehen, and deaths that occurred

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u

/
J
ich

N
H
A
R
T
ich
C
e
–
P
D

F
/

4
2
4
5
0
3
1
6
9
9
3
2
2

/
J
ich

N
H
_
A
_
0
0
3
0
2
P
D

B
j
G
u
e
S
T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

510 | ALTER, GUT MANN, LEONARD, AND M ERCH ANT

The articles presented in this issue successfully manage the
challenges of longitudinal analysis and fully utilize the capabilities
of life-history data to answer important questions about demo-
graphic processes and historical change: “Offspring Sex Prefer-
ence” by Nora Bohnert, Hilde Leikny Jåstad, Jessica Vechbanyon-
gratana, and Evelien Walhout and “Is Sibling Rivalry Fatal?”
by Rebecca Kippen and Sarah Walters examine the inºuence of
family composition on fertility and mortality. “Intergenerational
Transmission of Reproductive Behavior” by Julia Jennings, Alli-
son Sullivan, and J. David Hacker and “Migrants and Diffusion of
Low Marital Fertility” by Mathew Creighton, Christa Matthys,
and Luciana Quaranta analyze the intergenerational transmission
and spatial diffusion of fertility behavior. “The Determinants of
Reproductive Behavior” by Rosella Rettaroli and Francesco
Scalone and “Releasing Mother’s Burdens” by Bárbara Revuelta
Eugercios explore family formation in times of social and eco-
nomic transformation.

contributions made by the articles in this issue The extent
to which couples’ desire for male (or female) children led them to
adopt fertility control has become an important issue in analysis of
the fertility transition, arising ªrst in studies of Asian societies
where son preference is well established, and more recently exam-
ining possible son preference in agricultural households. Bohnert
et al. expand the question to ask whether couples also preferred
daughters or a balance of sons and daughters, to examine whether
these preferences changed over time, and to assess whether the
prevalence of sex-preferential fertility behavior increased with the
onset of fertility transition and smaller families. Außerdem, Sie
directly address the question of agriculture and religion as indica-
tors of couples who may have been particularly motivated to pro-
duce sons. Both parity-speciªc control (anhalten) and progression
to another birth (spacing) are considered. They utilize data from

under observation. A missing date of death might imply migration, but the date of migration
is unknown. As a pioneer of this method, Henry speciªed that an event related to the process
in question cannot serve to mark the end of observation. In the case of fertility analysis, wenn die
dates of child birth or child death were used as the end of observation, women with more
children would be overrepresented in the analysis. Henry, “Some Data on Natural Fertility”;
Gutmann und Alter, „Familienrekonstitution als ereignisgeschichtliche Analyse,“ in Reher und
Schoªeld (Hrsg.), Alte und neue Methoden in der historischen Demographie (New York, 1993), 159–177.

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u

/
J
ich

N
H
A
R
T
ich
C
e
–
P
D

F
/

4
2
4
5
0
3
1
6
9
9
3
2
2

/
J
ich

N
H
_
A
_
0
0
3
0
2
P
D

B
j
G
u
e
S
T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

EINFÜHRUNG | 511

the Utah Resource for Genetic and Epidemiologic Research
(rge)—compiled by the Utah Genealogical Society in the 1970s
from the family histories of the descendants of the Utah pioneers.
To create the database, the three-generation family group sheets
used by the Church of Jesus Christ of Latter-day Saints (lds),
which closely resemble the family-reconstitution forms developed
by historical demographers, were linked into longer genealogies
and supplemented with vital records kept by the state of Utah.
The time depth of the rge makes it a valuable source for analyzing
change in demographic behavior over time.9

The authors draw on fertility histories of women born be-
zwischen 1850 Und 1900 in order to span the pre- and posttransitional
periods, dividing their analysis by mothers’ birth cohort. Compar-
isons of the parity-speciªc sex ratios of living children between
couples who stopped at a given parity and couples who continued
to bear children address the extent to which the sex of offspring
inºuenced further childbearing; Cox regressions with independ-
ent variables indicating the sex mix of previous children assess the
effect of this mix on birth spacing. Results for the ªrst analysis in-
dicate that women who gave birth to a daughter at higher parities
were more likely to have another child; results for the second
demonstrate that, also at higher parities, the likelihood of a subse-
quent birth was elevated for women whose previous children
were mostly girls and reduced for women whose previous chil-
dren were mostly boys, both relative to those with a balanced sex
mix. Whereas the spacing pattern is evident even in the earliest
cohorts, it emerges at lower parities in later cohorts, among whom
the stopping pattern is also more pronounced. These ªndings sug-
gest that the pretransitional preference for sons persisted across the
Übergang, though the means of realizing it changed. Agricultural-
ists demonstrated a preference for larger families, regardless of sex;
lds members demonstrated a preference for sons similar to that of
non-members. Bohnert et al.’s ªnding of son preference in a
North American population breaks new ground.

9 For more information about the rge, see Lee L. Bohne, Dean L. Mai, and Mark Skolnick,
“The Mormon Historical Demography Project,„Historische Methoden, XI (1978), 45–53. Exam-
ples of intergenerational analysis using the rge include Anderton et al., “Intergenerational
Transmission of Relative Fertility and Life Course Patterns,„ Demografie, XXIV (1987), 467–
480; Ken R. Smith et al., “Effects of Childhood and Middle-Adulthood Family Conditions
on Later-Life Mortality: Evidence from the Utah Population Database, 1850–2002,” Social
Science and Medicine, LXVIII (2009), 1649–1658.

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u

/
J
ich

N
H
A
R
T
ich
C
e
–
P
D

F
/

4
2
4
5
0
3
1
6
9
9
3
2
2

/
J
ich

N
H
_
A
_
0
0
3
0
2
P
D

B
j
G
u
e
S
T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

512 | ALTER, GUT MANN, LEONARD, AND M ERCH ANT

Kippen and Walters also consider the inºuence of siblings on
demographic outcomes, in this case the survival prospects of chil-
dren under the age of ªve. They question whether having siblings
near to one’s own age increased the risk of death, whether having
much older siblings had a protective effect, and whether some
families faced poorer survival prospects for all of their children.
The theoretical underpinnings of these questions lie in the compe-
tition for resources, maternal depletion, and unobservable biologi-
cal traits that elevate the risk to related children. Since differences
in mortality for neonates, infants, and children are well docu-
erwähnt, the authors analyze these age groups separately. Kippen
and Walters’ questions are particularly suited to longitudinal analy-
sis of life-history data because the characteristics of the offspring set
vary over the course of parents’ marriages and children’s lives. Der
authors use data from the village of Sart in the Historical Database
of the Liège Region—a database created by Alter and Michel Oris
from population registers kept by eleven Belgian communes in the
second half of the nineteenth century, supplemented with infor-
mation from civil registers of births, Ehen, und Todesfälle. Begin-
ning in 1846, Belgian law required municipalities to record the
population by domicile and to update the information continu-
ously in response to births, Todesfälle, Ehen, and migration. Sart
began its population registration earlier, In 1812, providing an
even deeper record of continuous individual-level data covering
most of the nineteenth century. Because population registers list
individuals as members of households, they allow scholars to ana-
lyze the effects of household composition on demographic out-
comes.10

Using this rich data set, the authors include the number and
ages of siblings present in the household as time-varying covariates
in a series of Cox proportional-hazards models. Their models also
include the proportion of siblings dying in infancy as a way to as-
sess whether infant mortality clustered in households even after
controlling for household composition and other potentially con-
founding factors. Results conªrm that for neonates, co-resident
siblings under ªve years of age increased the index child’s mortal-
ity risk and that infant mortality clustered in families, even when

10 The Historical Database of the Liège Region is described more fully in Alter et al., “The
Family and Mortality: A Case Study from Rural Belgium,” Annales de démographie historique,
101 (2001), 11–31.

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u

/
J
ich

N
H
A
R
T
ich
C
e
–
P
D

F
/

4
2
4
5
0
3
1
6
9
9
3
2
2

/
J
ich

N
H
_
A
_
0
0
3
0
2
P
D

B
j
G
u
e
S
T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

EINFÜHRUNG | 513

controlling for the length of time elapsed since the previous sibling
was born and the presence of siblings old enough to supplement a
mother’s care. Wichtig, they also ªnd that children under the
age of one did not beneªt from Sart’s nineteenth-century mortal-
ity transition. This study reveals important determinants of neona-
tal mortality, Aber, methodologically, it also points to the impor-
tance of segmenting analysis of infant mortality by age. Taken
together, this article and that of Bohnert et al. make a compelling
argument for the importance of family composition to a proper
understanding of fertility and child mortality, while also attending
to critical methodological issues.

intergenerational

Another important dimension of family context is the rela-
tionship between generations. Entsprechend, intergenerational pro-
cesses have received a great deal of attention recently in historical
demography. Jennings et al. extend the “nature versus nurture”
debate on the basis of
fertility transmission,
making three important contributions: Erste, they examine the
relationship not only between women’s fertility and that of
their mothers but also between women’s fertility and that of their
mothers-in-law; zweite, they test whether the strength of this re-
lationship increased during the fertility transition; and third, Sie
add the vital status of a woman’s mother and mother-in-law as
time-varying covariates to assess the role of the extended family in
fertility decisions.

The genealogical basis and long historical span of the rge (von-
scribed above) make it ideal for Jennings et al.’s intergenerational
Analyse. Results from bivariate correlations and Cox proportional-
hazards models demonstrate that correlations between a woman’s
fertility and that of both her mother and her mother-in-law
emerged at the end of the nineteenth century, providing evidence
that the vital status of the mothers of both spouses affected fertility,
controlling for age, cohort, and other demographic factors. Der
authors suggest that, although the behavior of a woman’s own
mother had a stronger inºuence on her fertility, her mother-in-
law’s ideals of the number and timing of births may have been
transmitted directly or through her son, indicating the importance
of husbands in conception and contraception. They emphasize
that explanations of intergenerational fertility patterns must go be-
yond simple biological models of
fecundity to include social
norms, support mechanisms, and intrafamilial power dynamics.

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u

/
J
ich

N
H
A
R
T
ich
C
e
–
P
D

F
/

4
2
4
5
0
3
1
6
9
9
3
2
2

/
J
ich

N
H
_
A
_
0
0
3
0
2
P
D

B
j
G
u
e
S
T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

514 | ALTER, GUT MANN, LEONARD, AND M ERCH ANT

The conclusions of this article argue that the transmission of new
social norms from one generation to the next played an important
role in the fertility transition, and particularly in its irreversibility.

Mechanisms for the spatial diffusion of low fertility in the
nineteenth century have been more asserted than tested. In ADDI-
tion to the familial pathway explored in Jennings et al., migration
is a logical theoretical mechanism for the movement of ideas
through space. Although migration within Europe increased in
both volume and distance during the same period when fertility
declined, the question of whether movers and stayers experienced
different fertility trajectories has received little attention in the his-
torical-demographic literature. This lacuna may reºect the dif-
ªculty of tracking migration in family reconstitution and genea-
logical data. Creighton et al. begin to ªll this gap through analysis
of the Historical Database of the Liège Region (described above),
based on records noting migration into and out of the village of
Sart between 1812 Und 1900.

The authors predict at the outset that because Sart was a rural
commune that experienced a relatively late fertility transition, mi-
grants would be found to have had lower fertility than natives.
They further suggest that, if this fertility gap were produced by
conscious family limitation, it would have opened only in the sec-
ond half of the nineteenth century, when the fertility transition
was underway in the surrounding areas. The authors test these
ideas with Cox proportional-hazards models, controlling for a
number of other factors that might have affected the likelihood of
conception, including husband’s occupation, wife’s age, Parität,
breast-feeding, and the sex composition of children in the house-
hold. Acknowledging the role of both partners in making fertility
decisions, the authors assess the effects of both wives’ and hus-
bands’ migration status. Analysis conªrms their theory only for
dual-migrant couples. Although migrant status had no effect on
fertility earlier in the century, nach 1850, couples in which both
husband and wife had migrated to Sart were less likely to conceive
than were couples in which at least one partner was born in Sart.
These ªndings suggest an important relationship between the
movement of people and the movement of ideas, and a potentially
important role for migration in the European fertility transition.
The study by Creighton et al. points to the geographical un-
evenness of the fertility transition; that by Rettaroli and Scalone

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u

/
J
ich

N
H
A
R
T
ich
C
e
–
P
D

F
/

4
2
4
5
0
3
1
6
9
9
3
2
2

/
J
ich

N
H
_
A
_
0
0
3
0
2
P
D

B
j
G
u
e
S
T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

EINFÜHRUNG | 515

documents the socioeconomic unevenness of both pre-transition
fertility and fertility decline in the rural hinterland of Bologna.
Having reconstituted both individual life histories and household
contexts in two parishes of Bologna between 1818 Und 1900 aus
institutional records of the Catholic Church, the authors use dis-
crete-time event-history analysis to assess the relative effects of
biodemographic factors, household characteristics, and economic
ºuctuations on fertility prior to, and in the early years of, Bo-
logna’s fertility transition.

The results of their analysis indicate that such biodemo-
graphic variables as mother’s age, time since last birth, and vital
status of most recently born child were still the main determinants
of fertility throughout this period. dennoch, controlling for
these variables reveals independent effects of both socioeconomic
status and ºuctuations in grain prices, even in the earliest years of
Analyse. The authors also identify a monotonic decrease in fertility
across the nineteenth century. Jedoch, analysis of interaction ef-
fects indicates that this fertility decline was not shared equally
across social strata. The fertility of farmers and sharecroppers, WHO
typically lived in multiple-family households and had the highest
fertility overall, neither changed over time nor responded to
short-term price oscillations. This article hints at a link between
fertility decline in this region and the post-uniªcation shift in agri-
cultural production away from sharecropping, which privileged
larger families, and toward wage labor, which privileged smaller
ones. Rettaroli and Scalone’s breakthrough is the establishment of
socioeconomic status as a determinant of fertility prior to the de-
mographic transition.

Even as fertility declined throughout Europe, parents contin-
ued to employ postnatal strategies for controlling family size and
Komposition. Revuelta Eugercios explores the complexities of
abandonment, the most common European postnatal strategy.
Relinquishing children to foundling hospitals had long been part
of the family-building repertoire of poor parents, but the author
demonstrates that in the early twentieth century, parents in Ma-
drid and the surrounding areas began to utilize foundling hospitals
as a means of temporary rather than permanent relief from the
burden of caring for their children. Individual-level data from the
Admission and Exits and Entries books of the Foundling Hospital
of Madrid between 1890 Und 1935 offer a unique opportunity to

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u

/
J
ich

N
H
A
R
T
ich
C
e
–
P
D

F
/

4
2
4
5
0
3
1
6
9
9
3
2
2

/
J
ich

N
H
_
A
_
0
0
3
0
2
P
D

B
j
G
u
e
S
T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

516 | ALTER, GUT MANN, LEONARD, AND M ERCH ANT

examine how poor families adapted available maternal-welfare
provisions to meet their needs in a moment when ideas of moth-
erhood were changing and new social-reform legislation was
emerging.

Revuelta Eugercios uses discrete-time event-history analysis
to examine the correlates of this new use of the Foundling Hospi-
tal, marked by successful child retrieval. Results reveal two pat-
Seeschwalben: Among those born at the adjoining maternity hospital, legit-
imate children and the children of younger mothers were more
likely to be retrieved, regardless of sex; among those who were
brought to the foundling hospital after birth, male children and
children who were older at the time of abandonment were more
likely to be retrieved, as were children whose parents provided
some kind of information to the institution at the time of aban-
donment. These patterns suggest that some of the women who
availed themselves of the maternity hospital also utilized the
foundling hospital for temporary child care as part of this welfare
entitlement, and that some parents who brought their children to
the foundling hospital viewed abandonment as a short-term sur-
vival strategy, leaving notes to indicate their retrieval intentions.
Revuelta Eugercios’ analysis highlights the changing roles of insti-
tutions in demographic processes and parental initiatives in the use
of institutions, as well as some evidence for son preference in early
twentieth-century Madrid. Because of the high mortality rates in
the foundling hospital, particularly among infants, parents likely
recognized the risk involved in using the Foundling Hospital as a
temporary refuge.

The six articles presented in this special issue of the Journal of Inter-
disciplinary History bring new substantive and methodological
insights to the ªeld of historical demography—revealing the re-
sponsiveness of pre-transition fertility to changing household and
economic contexts, tracking the transmission of new fertility
practices between generations and across space, exploring the un-
evenness of mortality and fertility decline within societies, Und
documenting the changing role of social institutions in family for-
mation. They also suggest fruitful avenues for further research, In-
dicating the vast and still largely untapped potential for longitudi-
nal data and methods to provide additional knowledge about
demographic processes, demographic change, and the connec-

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u

/
J
ich

N
H
A
R
T
ich
C
e
–
P
D

F
/

4
2
4
5
0
3
1
6
9
9
3
2
2

/
J
ich

N
H
_
A
_
0
0
3
0
2
P
D

B
j
G
u
e
S
T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

EINFÜHRUNG | 517

tions between demographic change and social, wirtschaftlich, and po-
litical transformations. This collection demonstrates the immense
value of training new generations of scholars in the longitudinal
analysis of historical-demographic data, equipping them to extend
this work in time and space as the number and size of life-history
databases grows, Und, as analytical methods continue to advance,
inspiring them to seek new insights from the surviving traces of
die Vergangenheit.

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u

/
J
ich

N
H
A
R
T
ich
C
e
–
P
D

F
/

4
2
4
5
0
3
1
6
9
9
3
2
2

/
J
ich

N
H
_
A
_
0
0
3
0
2
P
D

B
j
G
u
e
S
T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3

D
Ö
w
N
Ö
A
D
e
D

F
R
Ö
M
H

T
T

:
/
/

D
ich
R
e
C
T
.

ich
T
.

e
D
u

/
J
ich

N
H
A
R
T
ich
C
e
–
P
D

F
/

4
2
4
5
0
3
1
6
9
9
3
2
2

/
J
ich

N
H
_
A
_
0
0
3
0
2
P
D

B
j
G
u
e
S
T

Ö
N
0
8
S
e
P
e
M
B
e
R
2
0
2
3
PDF Herunterladen