ff. YUAN. TIEN'
FRANK 1). BEAN
L. J. BRILL..
Theory and Method in Historical
Sociology: Household and Family
Structure London 1851
Western Michigan University, Kalamazoo, Michigan, U.S.A.
THE SOCIAL HISTORY of the family has become an increasingly
active area of research in several disciplines. Social historians such as Berkner
(1972), Blumin (1972), Demos (1968) and others have been joined by such
sociologists as Sennett (1970), following Smelser's (1959) earlier work. Ireland
and England among the European countries have received the most intensive
attention, followed by France, Sweden, Holland and Germany. Most of this
research has been focused either on small, usually rural, communities, or on the
entire society. The relatively few studies of urbanization and the family in
England have either concerned pre-industrial London (Glass, 1966; Laslett,
1969, 1964, 1970), 19th century factory towns or 20th century London and the
planned "new towns." A major gap is the social history of London exploiting the
rich possibilities from recoding the original 19th century census enumerator
schedules. Given the 100 year rule, and the inadequacy of the delineation of
households and families and occupations before 1851, it is understandable that
this source has only recently been tapped. To date only two scholars have
published work on this data; Dyos (1961) and Lees (1969), both historians.
Michael Anderson's project to computerize a 2% (400,000) sample of the entire
1851 Census for Great Britain therefore promises to open up a new era of
British social history.
One of the prime desiderata of such work would be collaboration between
historians and social scientists given the highly complementary and equally
indispensable skills each has to offer the other in the area of the social history
of the family. Such collaboration, however, has been extremely rare. The major
obstacles seem to be:
(1) Historians do not collaborate with each other-so we needn't feel
especially "left out." Cohn (1962: 15) in a delightful ethnographic participant-
observer study of the tribal mores of historians, found the collaborative climate
"chilly" and attributed this to a "recurrent phase of nativism." However,
Cohn's description of anthropology as involving inductively derived theories
and field researchers jealously guarding "their tribes" places them nearer to
historians in these respects than to sociologists. Replication (e.g., Oscar Lewis
and Robert Redfield on Tepotzlan) is still rare in anthropology, and certainly
not common enough in sociology.
This situation is changing slightly. In Rowney and Graham's reader on
Quantitative History, of the 4 (out of 25) co-authored papers, two were by pairs
of non-historians, 1 by a French (Harvard trained) historian paired with a
French Computer Science Ph.D. and one by two University of Washington
trained historians doing essentially behavioral political science analyses of U.S.
(2) There is evidently not an historical methodology which the social
scientists can learn-at least most history departments do not care to offer such
courses. The argument is that each type of history for each era and place has
a particular set of appropriate methods to be learned as an art by doing it,
or at least working as an apprentice to a master.
There is, of course, an historical methodology in a more general sense;
issues of periodization, procedures for locating and cross-checking documents,
etc., but for purposes of historian-social scientist collaboration, methodology
will be an obstacle to communication.
The sociologist might well suppose that his expertise lay in quantitative
data analysis while that of the historian consisted primarily of in-depth immer-
sion in the period and place of interest. He may, however, find historians
objecting to "servicing" sociologists in what they may perceive as a demeaning
(3) Definitions of "problems." As Cohn (1962: 17) notes, "research in
history is based on finding data; research in anthropology is based on creating
data... most historical research is done because there is a known body of
source material available." In principle, social scientists start with a model and
then look for data with which to test it. But we also generate models inductively
and most certainly we also carry out much research because the data, and
funds, are available. We are, however, supposed to be looking for the "general"
while most historians insist that each event is unique, thus rejecting the feasibil-
ity of a science of social behavior.
(4) The different market for history and historians compared to social
scientists-especially sociologists. Even in 1962, Cohn could say that anthropo-
logy faced an expanding universe while history had to confront a stable or
declining one. As a result, very able and initially highly specialized historians
end up at poor schools teacihng general lower level courses During the 60's,
sociology graduated from an earlier dubious popularity based on being one of
the "gut" departments to a booming enrollment arising from its relevance to the
social issues of this decade. While sociologists did not have the answers, they
seemed to be asking the right questions. We therefore, confront problems of
shifting power and status within our universities vis-a-vis our "own" historians.
(5) History as a discipline offering far fewer graduate fellowships than
sociology has tended to draw from a higher class background which raises
barriers of "taste" and even class accents. In fact, history may be extreme in its
THEORY AND METHOD IN HISTORICAL SOCIOLOGY
attraction to the downwardly mobile with a preference for an immersion in a
more attractive past.
(6) Status in history is highly correlated with age. It is extremely difficult
to be a "child prodigy" in history-as compared to fields such as mathematics
where genius is revealed by age 25 if ever. Social scientists on the other hand
are accustomed to respecting the rising methodological "sophistication" and
novel data produced by their most recent doctoral candidates.
Given these considerable subculturall" differences, what sort of historians
are sociologists most likely to be able to communicate if not collaborate with-
marginals or the vanguard? Where does social history stand within the situs
History in the general sense seems to be on our side. The recent popularity
of sociology has been partly at the expense of history, especially "irrelevant"
periods before the Industrial Revolution. Social and quantitative historians
(not entirely overlapping groups) are predictably younger. Quantitative
history, of course, also qualifies for high levels of research support. Many older
historians have fled to academic administration in disproportionate numbers
(followed more recently by physicists and agriculturalists). In addition, the
sociologist pursuing the social history of specific communities (should) soon
discover both a rich but dubious source of material-that of "local history"
written by amateurs. Most earlier sociological community studies now fall into
this low status category, at least in the eyes of elite historians.
The next issue is then, what does the sociologist do when dealing with
"historical" periods. Does he simply "do" history and thus try to become an
historian, at least by dealing with original documents; or should he rework
heavily researched topics such as the history of the textile industry in the
Industrial Revolution in terms of a strong interpretative model a la Smelser?-
(Even in the latter case of course, "holes" in the data appeared requiring
primary source research.) The latter approach is infra dig among traditional
historians as "derivative" "secondary source" work, at best relevant to text
books and popularizations or Toynbean philosophizing.
Both Erikson and Cohn maximized their status relationship problems with
historians by daring to invade elite preserves-American and British colonial
history respectively. They thus were especially open to the "fear and contempt"
historians could be said to have for sociologists.
No simple formula for collaboration will be suggested. Co-authored produc-
tion is extremely rare among historians or between historians and social scientists.
At the other extreme, the sociologist would have to publish in an interdiscipli-
naryjournal, or have his work reviewed by historians only then to discover his
gaffes in missing relevant data or deficiencies in his contextual depth. However,
the growing number of interdisciplinary Social History workshops and news-
letters should provide a fruitful arena for communication.
Besides historians and a few sociologists, the major stream of research
bearing on English Family History has been that of demographers and econo-
mic historians (Drake, 1969; Glass and Eversley, 1965; and Wrigley, 1966,
1968). For the most part their work has dealt with the traditional population
parameters of births, deaths, migration and the resultant changes in population
size and growth rates and age structure, largely at the aggregate level. Their
interest in the family and household structure has been limited largely to their
contribution to, or reflection of, these societal demographic issues.
Burch (1967) fairly well exhausted the available aggregate data from con-
temporary censuses on the residential family as a part of the household unit. He
clearly recognized the limitations of his findings on most of the issues of interest
to sociologists of the family e.i., intrahousehold interaction and extra household
kin and fictive kin networks. Together with Laslett (1970), he has fairly defini-
tively provided cross-cultural evidence as to the small size of the household de-
fined as the residential family unit, although clearly sociometric studies of housing
arrangements in primitive societies could permit rather large co-resident
The next steps beyond Burch's works should be: (1) disaggregated socio-
metric studies (like Adams, 1968; Aldous, 1961; and Litwak, 1966) as to the
significance of co-residence and household size. As Burch (1967: 351) implies,
the most easily testable hypotheses may have a limited range of theoretical
significance; (2) historical studies of the same areas over time and of specific
My own project is much less ambitious in temporal, geographic and docu-
mentary scope. It will extend a parish-stratified 1 % sample of Greater London
(Division I) forward from 1851 to 1861 and 1871 thus, in principle, permitting
family reconstruction. However, given the rate of change in London during
this period and the small percent of the sample, only the stable survivors living
in the same households can feasibly be followed up. Clearly this group is going
to be a biased sample of "stables" (by missing families moving only a block
away into a parish not included in the sample) and thus not typical of London
society in general. However, since the nature of this bias can be determined
mobiles such a follow-up seems well worth the effort.
Size and Structure of Households
The 1851 census for the first time reliably differentiated housefulss" (all
residents at one address) from households and families. The information availa-
ble on individuals is: name, address, relation to head of familily, marital status,
age, sex, occupation, birthplace, and whether deaf, dumb or blind. For all in-
dividuals the following additional characteristics were generated in the initial
coding: (1) household size; (2) number of other households in the same
houseful; (3) size of their family, if a relative of the head. In addition, for
a later fertility project the following data were coded on "all children under
one year of age": (1) age in months; (2) mother present or absent; (3) father
THEORY AND METHOD IN HISTORICAL SOCIOLOGY
present or absent; (4) marital status of mother; (5) occupation of father;
(6) mother's birthplace compared to child's; and whether or not child was
an orphan. My interest in domestic servants also led to generating additional
household characteristics for all servants whether living with their employers
or in their own homes.
In relation to the findings and issues Burch discusses with respect to house-
hold size and structure, I will be able to specify thirty-eight types of kin and
twenty-two types of non-kin. At this point Burch's suggestion that the major
deviation from universally small sized households would occur in urban areas in
industrializing countries would appear to be the case in London.
This raises the issue of differential fertility by class and origin. Many of the
general demographic histories discuss aggregate trends nationally or by differ-
entiating rural/urban or class. Banks' (1954) diffusion hypothesis maintains
that the middle class initiated birth control to maintain its life style and
eventually diffused its practice to the lower class. The mechanism of this
diffusion, and even the existence of the class differential in 1851 London, could
well differ from his highly general mid-Victorian pattern. Banks (1964: 83-
84) noted a 36% increase from 1851-71 in households (separate occupants)
and a 56% increase in the number of domestic servants. The resulting per
"household" ratio of servants apparently rose during this period from one for
every 4.9 households in 1851, to one for every 4.6 households in 1861, to one
for every 4.2 households in 1871.1
Moreover, the composition of the servants shifted toward the replacement
of males and a larger retinue per household, toward the single all-purpose (and
increasingly single) female. Thus the mechanism of "diffusion" of birth control
may have been "structural" social control rather than normative imitation.
Women with children were not hired as servants, since their primary function
became increasingly the care of their employer's children. Thus in London,
where the per cent of servants was far higher than for the country as a whole,
a reverse class fertility differential is possible. The fertility of married lower class
housewives may have been higher than their middle class counterparts, but
this may have been offset in the aggregate by the larger number of single female
The use of "children under 1" related to their mother's characteristics is
the census source for the 1851 London fertility structure. It has obvious defects
relative to birth registration data in terms of enumerating all live births occurr-
ing throughout the year (missing those born in 1851, either after the census was
taken or born in 1851, but dying before the enumeration), but the link to the
mother's characteristics plus other sources for the annual birth rate make this
data of great interest. Servants will unfortunately appear to have been more
childless than may have been the case since, especially if they were unmarried
1 As Michael Anderson noted (letter to author), the changed definition of households in 1861
could mean that the actual increase in servants was even greater than that suggested by
mothers, they would have had to abandon or farm out their children in order
to obtain the usual "indoor" (live-in) employment.
A non-demographic focus of this study will be the delineation of the
degree of, and trends in, the "commercialization" of households, or, in Smelser's
terms, the functional segregation of residence and employment. My expecta-
tion is that London, due to its economies of scale and its rising middle class
with an increasing proportion of discretionary income, will "permit" the
persistence or even flourishing of "archaic" luxury handicraft manufacture
in the homes of the manufacturer long after they have died out in the rest of
England (for example, in St. Botoloph's: silk weavers, hatters, jewelers, etc.).
Thus while large capital cities may also symbolize and control the prime
movers of modernization, they also will embody the last vestiges of many tra-
ditional patterns, not as "culture lags," but as an integral part of the process
of industrialization in a free enterprise economy.
Larger households, and "archaic" occupations are of course, not unrelated.
A major related variable not provided in the census but being researched by
my English collaborator, Leonore Lockwood, is the question of tenancy status.
Evidently the availability of private dwellings for purchase was delayed even
for those willing and able to buy, owing to the policy of the stewards of those
landed aristocrats who held large blocks of London real estate. With landlords
insisting on renting rather than selling, tenants were moved to deflect their
income from capital improvements in household technology like plumbing,
central heating, etc., into "portable" labor intensive-services (servants carrying
pails of hot water up three flights of stairs to fill a tin bathtub, and later bailing
it out!). The popular cultural interpretation of the notorious lag in Britain's
domestic amenities has leaned on British conservatism but it would seem to
be due at least as much to an ample supply of servants and the unwillingness of
landlords to sell property.
The above focus on degrees of "commercialization" of households will also
serve to analyze changes in the same residence over time, and thus contribute
to urban as well as family sociology. St. Botoloph's, for example, was unusual
in declining in population given the rapid growth in the city as a whole. It was
changing from a light manufacturing to a commercial central business district.
Other Methodological Problems
J. A. Banks (1968: 408-409) makes the common observation that "sociol-
ogists often mistrust historical data because they have found that many popular
impressions about social relations in our own time are incorrect." However,
he goes on to show how the generally much more critical and exhaustive use of
all relevant evidence by historians can counter such biases while sociologists
can too easily become credulous about the validity of their face-to-face contact
with their subjects.
In the case of this census material we have a variety of opportunities for
internal and external documentary verification. For the "stables" between 1851
THEORY AND METHOD IN HISTORICAL SOCIOLOGY
and 1861 the necessarily unchangeable facts of age (adding 9-11 years) and
birthplace offer reliability checks on at least this subsample. There are other
presumably unchangeable traits which can be checked such as blood kinship.
Other "relationships to head" are an interesting challenge. If landlords or
revenue agents disfavored subleasing, perhaps a number of boarders were
reported as relations, "friends," visitors or servants. In the latter case, a real
multiplicity of functions is quite possible; apprentices serving also as domestic
servants and being paid "board wages." Our best leads in this area are bio-
graphies of authors who cropped up in our sample. Carlyle's description of his
household matches that of the census perfectly but further efforts in this
direction may not prove so rewarding.
A major methodological problem is the nature of the maps available and
the problem of continuity in residential enumeration. The original census tract
maps presumably used in the 1851 census were apparently "recycled" as anti-
aircraft shell casings during WWII despite David Glass' reported efforts
to save them. The surviving most relevant map has most housing correct for
the late 1840's with the new rail lines superimposed up to 1857. One cannot
then be sure, for instance, that 51 Petticoat Lane was the same building in both
years. The family changed but the head's employment remained "Licensed
Victualler." Is this any less likely to be the same place as #1 Petticoat Square
where Henry Israel and his family ran a General Dealers shop in both years?
Hopefully we will encounter some guide to residential enumeration but since
Greater London was a loose agglomeration of independent boroughs and
counties no single set of rules may have been followed throughout.
The most dubious element in my analysis at this point is the ranking of
households in each residence. I have yet to determine definitively if the first
one listed is: (1) randomly the first one accessible to the enumerator; (2) first
floor front, etc., (3) the owner, if a resident. The data in tables 1, 5 and 6 do,
however, shed some light on the question of ranking as discussed below.
Following the "same" residences over time we find some with enough stable
families to suggest that it is the same building; but if the third household listed
has changed, were they occupying the same rooms? Certainly no adequate
verification of this problem is even going to be attempted beyond internal
crosschecks of the same residences with some stable families over three censuses.
A reasonably plausible pattern of subdividing in sections of the city known
otherwise to be declining may be ascertainable.
In Table 1, the 60 discrete "relationships to head of household" found on
the original census enumerators schedules were recorded for the 19,822 non-
institutionalized individuals in terms of the 3,972 households. This "household
structure" characteristic was then given to the head as an additional personal
attribute, which, together with the size of his or her household and family,
served as the data for this table. Originally nine alternatives were used. Three
"extended kin" combinations yielded too few cases to justify separate considera-
tion at this point. They were:
1. Nuclear unit and children of head's generation and younger generation,
Household Structure by Number of Other Households in Same Houseful in
Terms of Household and Family Size
One Person "Solitary" Heads Nuclear Unit Nuclear & Extended
Households with Only Non-Kin Only Kin Only
Number of Other Size Size Size Size
Hslds. In Houseful of Households of Households of Households of Households
Hslds. No. % Hslds. No. % Hslds. No. % Hslds. No. %
Single Household 1.0 28 11.73.8 128 66.3 5.0 635 36.2 5.3 179 48.1
1 other hsld. 1.0 66 27.6 3.3 36 18.7 3.8 523 29.8 4.1 97 26.1
2 other hslds. 1.0 67 28.0 2.6 11 5.7 3.5 361 20.6 4.2 53 14.3
3 other hslds. 1.0 46 19.3 3.0 9 4.7 3.2 106 6.0 4.2 20 5.4
4 other hslds. 1.0 13 5.4 2.3 3 1.6 3.8 63 3.6 4.0 7 1.9
5 other hslds. 1.0 9 3.8 2.2 5 2.6 3.1 33 1.9 4.0 6 1.6
6 other hslds. (1.0) (1) .4 (2.0) (1) .6 4.3 28 1.6 5.5 6 1.6
9+ other households 1.0 9 3.8 0 0 0 3.7 7 .4 (5.8) (4) 1.1
Total 1.0 239 100% 3.5 193 100% 4.1 1756 100% 4.7 372 100%
6.0% 4.9% 44.3% 9.4%
Nuclear Plus Non-Kin Nuclear and Extended Total
Family and Non-Kin
Number of Other Size Size Size Size Size Size
Hslds. in Houseful of of Households of of Households of of Households
Hslds Fam. No. % Hslds Fam. No. % Hslds Fam. No. %
Single Hsld. Houseful 6.5 4.4 720 71.3 6.9 4.5 308 79.0 5.7 4.5 2004 50.4
1 other hsld. 5.6 4.1 171 16.9 5.5 3.9 56 14.4 4.0 3.7 953 24.0
2 other hslds. 5.3 3.8 79 7.8 4.8 3.8 18 4.6 3.5 3.3 590 14.9
3 other hslds. 5.1 3.8 22 2.2 (5.7) (4.0) 3 .8 3.0 2.8 206 5.2
4 other hslds. (6.3) (4.0) 4 .4 (5.0) (4.5) 4 1.0 3.5 3.3 96 2.4
5 other hslds. (4.3) (3.3) 6 .6 (4.0) (3.0) 1 .3 2.9 2.7 60 1.5
6 other hslds. (6.0) (4.0) 6 .6 0 4.6 4.1 42 1.0
9 or more other hslds.(6.0) (4.5) 2 .2 0 3.2 3.1 22 .6
6.2 4.3 1010 100% 6.6 4.4 390 100% 4.7 4.0 3973 100%
25.5% 9.8% 100%
such as nieces, nephews, children-in-law, step-children and grandchildren.
2. Nuclear unit and collateral relatives of the head-cousins, siblings and
siblings in law.
3. Nuclear unit and kin of the parent or grandparent of head generations.
4. Nuclear units with combinations of relatives in the first three cases.
In the first stage of this recording the observed frequencies were: 1-20;
2-33; 3-4; 4-315. It was then decided to combine 1 to 3 with 4 as "nuclear and
extended kin" in Table 1. In the next stage of our analysis other combinations
THEORY AND METHOD IN HISTORICAL SOCIOLOGY
of this critical type of household structure will be attempted, such as "married
son of head with wife" where they were not originally defined as a separate
Before discussing this table any further, the latter recode possibility raises
a basic methodological issue which must be examined. The explicit policy
followed in coding this "raw" data was to adhere as closely as possible to the
original in order (1) to preserve as much detail as possible so that other scholars
could recode it for their own needs, and (2) to inductively arrive at the coding
procedures actually used whether or not explicit rules were given. One of the
basic problems of this type was to determine when a household was said to
exist in the marginal cases of boarders, servants, etc., and kin of heads with one
or more members of their own nuclear units. For instance, the given directions
instructed the enumerators to code solitary boarders as a separate household,
but this rule was frequently disregarded. Rather than alter the actual data to
fit this rule, boarders were coded as given. At this first stage of aggregation,
we can see many types of households which may be affected by such ambiguities.
The "problem" households will then be examined to see if plausible patterns
can be determined. Since each sub-parish was enumerated by a single person,
variance in the "reliability" of this classification may be adequately explained
at this level. On the other hand, we may ultimately accept such classifications
as possibly reflecting the quality of personal relationships involved, or as in-
The other axis of Table 1 measures the extent of multiple household-
housefuls in London. "Housefuls" are defined solely as all households at the
same address. This classification, fortunately, is supported by explicit long (or
double) lines drawn between housefuls by the original enumerator (with short
or double lines between households) and so probably does not raise issues of
"forcing" dubious cases.
Another basic problem with such household structure recoding is that no
established model has been agreed on to facilitate comparative research.
Blumin (1972: 14) for instance, excludes boarders, and apparently, any single
person households (hence his larger family household and family sizes). Those
more interested in fertility and population change naturally are less interested
in such isolates, but the present concern for the full range of household life
styles called for their inclusion in my study.
Overall, the sizes of families and households in London were certainly
unexceptional relative to current or historically comparable patterns. India's
1951 urban household size was also 4.7 (Table 2), virtually the same size as the
national average of 4.8 for England and Wales. Compared to the other urban
data in this table, London has relatively more non-kin. (Its average number
of non-nuclear kin must await a further type of analysis.)
London's 6% of one-person households (Table 1) falls in the upper range of
Burch's (1967: 354-355) U.N. data, typical of northern, largely urban, Euro-
pean countries (no urban breakdown was given). In subsequent analysis an
attempt may be made to separate the "boarders" from the single person-single
Urban Household and Family Size
Size of Size of No. of No. of
Hsld. Nuc. Family Other Rel. Non-Kin
India, 1951* 4.7 3.5 1.1 .1
U.S., 1960* 3.2 3.0 .2 .6
Venezuela, 1961* 5.2 4.0 .8 .4
London, 1851 4.7 4.1 + .7
*Page 361, Burch, American Sociological Review, June, 1967.
household housefuls, but the small numbers probably involved may not make
this effort worthwhile. In addition, the above mentioned problem of lodgers'
frequent inclusion in family households may make such a distinction difficult.
It is interesting to note that 61% of one-person households are in housefuls
with at least 2 other households which suggests that even if all of these people
were lodgers, there were few "hotel-type" boarding houses. Poorhouses, to be
sure, are excluded from this tabulation. There were many in London at this
time with populations of 200-600.
The next category in Table 1, one rarely distinguished, were individual
heads living only with non-kin. (The number of non-kin can be determined by
subtracting 1 from the size of household.) In terms of multiple-unit residence,
those households, at 66%, fall between the nuclear and nuclear-extended, and
the larger-sized non-kin households in terms of the percent of single household
The largest single category were nuclear units, but they still amounted to
less than 50% of all households. Of these, only 36% were in "single family
dwellings." Overall, only 50% of all households were in single household
housefuls with the predominant number being nuclear and non-kin. (This
group of 720 households probably includes many of the upper class households,
which will later be identified by occupational data.) The largest households
with the largest number of non-kin (except for the first two "solitary" head
categories) were the most complex type, nuclear units and extended kin and
non-kin. Nearly 80% of these (the highest such percent) were in single family
Many of the households are of the "archaic" type better preserved in
London than in rural areas or industrial cities, namely multi-purpose dwellings
doubling as artisan handicraft shop-stores with servants, manufacturing work-
ers, clerks and boarders.
The general results in Table 3 are as expected. The minority of female heads
have, on the average, a few more non-kin in their households (.9/.8) than do
male heads. It will be interesting to see if the non-kin in the former case are
largely boarders, e.g., helping widows pay the rent. In this case the households
are presumably "poor" whereas if the non-kinwere servants, we have probably
the most reliable index of middle or upper class status. The single-female
Size of Household and Family by Marital Status and Sex of Head of Household
Single Married Widowed Total
Size Size Size Size Size Size Size Size
of of Households of of Households of of Households of of Households
Hslds. Fam. No. % Hslds. Fam. No. % Hslds. Fam. No. % Hslds. Fam. No. %
Males 2.5 2.1 207 60.4 5.2 4.4 2816 95.3 4.7 3.4 178 26.7 5.0 4.2 3202 90.6
Females 3.3 1.9 136 39.6 4.0 3.1 140 4.7 3.8 3.0 490 73.3 3.7 2.8 768 19.4
Total 2.8 2.0 343 100% 5.1 4.4 2956 100% 4.1 3.1 668 100% 4.7 4.0 3971 100%
8.6% 74.4% 16.8% 100% 0
Size of Household and Family by Age and Sex of Head of Household
Male Female Total
Size of Size of Households Size of Size of Households Size of Size of Households
House- Families No. % House- Families No. % House- Families No. %
holds holds holds
15-19 1.6 1.3 9 .3 (2.3) (2.3) 3 .4 1.8 1.6 12 .3
20-24 3.0 2.5 158 4.9 2.4 1.6 22 2.9 2.9 2.4 180 4.5
25-29 3.8 3.3 383 12.0 3.1 2.2 46 6.0 3.7 3.2 429 10.1
30-34 4.9 4.2 511 16.0 3.8 3.3 58 7.6 4.8 4.2 569 14.3 n
35-39 5.4 4.9 462 14.4 4.3 3.5 56 7.3 5.3 4.7 518 13.1
40-44 5.6 4.8 412 12.9 4.3 3.2 82 10.7 5.4 4.6 494 12.5
45-49 5.7 4.8 391 12.2 4.6 3.6 77 10.0 5.5 4.6 468 11.8
50-54 5.7 5.0 286 8.9 3.8 2.9 96 12.5 5.2 4.4 382 9.6
55-59 4.7 3.9 210 6.6 3.4 2.8 71 9.3 4.4 3.6 281 7.1
60-64 4.9 3.8 189 5.9 3.3 2.3 104 13.6 4.3 3.3 293 7.4
65-69 4.9 3.6 80 2.5 3.1 2.0 67 8.7 4.1 2.9 147 3.7
70+ 4.2 3.0 109 3.4 3.7 2.4 85 11.1 4.0 2.7 194 4.9
5.0 4.2 3200 100% 3.7 2.8 767 100% 4.7 4.0 3967 100%
80.7% 19.3% 100%
THEORY AND METHOD IN HISTORICAL SOCIOLOGY
heads have relatively more non-kin than do widows (1.4/.8) or married female
heads (.9). This may arise from the greater frequency of spinsterhood among
daughters of the well to do.
Table 4 delineates the life cycle structure of households and families. The
number of non-kin in male-headed households rises steadily from an average
of.5 under age 30 to 1.2 over age 60. Female heads reveal no consistent pattern
in this respect. For both, inclusion of marital status and employment with the
type of non-kin would clarify the picture but will require considerable loss of
age detail in order to maintain adequate cell size.
In Table 5 marital status is related to the "multiplicity" of housefuls in
terms of rank order instead of "number of other households" as in Table 1.
Both are aimed at the degree of multiple/single family housing. At this point,
as explained above, the rationale, if any, of the order in which households
appear in a houseful is open to question. The best way to unravel the rationale
would seem to be inductive analysis. We will, therefore, relate marital status
of the head to both measures of "multiplicity."
In Table 5 the household sizes of single heads decrease regularly the higher
(or perhaps, socially, the lower) their household rank. This is also the case in
Table 6, if we exclude the 4th and 6th order ranks which have very few cases.
Married household heads in Table 5 have household sizes which regularly
decrease to the third order cases and then increase to their second order level
by the time they are the "last" household in a sevenplus household houseful.
The percentage of married households heads declines steeply and continually
by rank, while single heads reveal a second "mode" in that there are twice as
many who are the second household than single heads who also "head" a
houseful. Again, delineation of the role of lodgers should clarify this pattern
With respect to the number of non-kin (subtracting family from the house-
hold), both measures demonstrate a strong and similar tendency for non-kin
to be limited to single or "duplex" households. However, both also show that
once past the duplex, higher orders do not further depress the frequency of
Table 6, however, shows a less regular decline in non-kin within each
marital status category. Overall then, the simpler measure of multiplicity
used in Table 6 does not appear to reflect its effect as reliably as does the rank
order used in Table 5. Therefore, a rank ordering of households seems to
indicate some sort of rationale. Hopefully a search of the literature of this
period will yield descriptions of some of the multiple family dwellings included
in our sample.
While a more finished set of findings would have been preferable, I welcome
the opportunity for collegial suggestions at this stage in my analysis. Aside from
the immediate further steps mentioned in the text, much more will be done with
Size of Households and of Families by the Marital Status of Household Heads and Rank Order Degree of "Multiplicity" of Houseful
Single Married Widowed Total
Size Size Size Size Size Size Size Size
Rank in Houseful of of Households of of Households of of Households of of Households
Hslds. Fam. No. % Hslds. Fam. No. % Hslds. Fam. No. % Hslds. Fam. No. %
Only household in
houseful 4.1 2.0 143 41.5 6.1 4.9 508 51.0 5.0 3.6 351 52.2 5.7 4.5 2004 50.4
First 2.6 1.7 44 12.8 4.7 4.2 611 20.7 3.3 2.9 126 18.9 4.4 3.9 781 19.7
Second 1.9 1.5 88 25.5 3.8 3.5 541 18.2 2.8 2.5 118 17.7 3.4 3.2 748 18.8
Third 1.7 1.5 44 12.9 3.3 3.1 186 6.3 2.9 2.5 49 7.3 3.0 2.8 280 7.1
Fourth 1.7 1.5 17 4.9 3.8 3.6 61 2.1 1.8 1.6 12 1.8 3.2 2.9 90 2.3
Fifth 3 .9 4.1 3.8 26 .9 2.3 2.1 7 1.1 3.6 3.3 36 .9
Sixth 3 .9 4.2 4.0 13 .4 2 .3 3.7 3.5 18 .5
Seventh 1 .3 4.7 4.6 7 .2 0 0 4.3 4.1 8 .2
Eighth 2 .6 3 .1 3 .5 3.0 2.7 8 .2
Total 2.8 2.0 345 100% 5.1 4.4 2956 100% 4.1 3.1 668 100% 4.7 4.0 3973 100%
8.7% 74.4% 16.8% 100%
Size of Households and of Families by Marital Status of Household Heads and Number of Other Households in Same Houseful
Size Size Size Size Size Size Size Size
No. of Households of of Households of of Households of of Households of of Households
in Houseful Hslds. Fam. No. % Hslds. Fam. No. % Hslds. Fam. No. % Hslds. Fam. No. %
Nine + others
4.1 2.0 143 41.5 6.1 4.9 1508 51.0 5.0 3.6 351 52.5 5.7 4.5 2004 50.4
2.4 1.8 83 24.1 4.4 4.0 716 24.2 3.2 2.8 153 22.9 4.0 3.7 953 24.0
1.8 1.5 63 18.3 3.9 3.6 440 14.9 2.9 2.7 86 12.9 3.5 3.3 590 14.9
1.5 1.3 34 9.9 3.6 3.4 137 4.6 2.3 2.0 35 5.2 3.0 2.8 206 5.2
2.2 2.2 5 1.5 3.9 3.7 69 2.3 2.3 2.2 22 3.3 3.5 3.3 96 2.4
1.5 1.0 8 2.3 3.4 3.2 42 1.4 2.3 1.9 10 1.5 2.9 2.7 60 1.5
(2.0) (2.0) 1 .3 4.9 4.5 35 1.2 3.3 2.3 6 .9 4.6 4.1 42 1.1
1.0 1.0 8 2.3 4.8 4.7 9 .3 3.8 3.4 5 .8 3.2 3.0 22 .6
2.8 2.0 345 100% 5.1 4.4 2956 100% 4.1 3.1 668 100% 4.7 4.0 3973 100%
the 1851 sample, and then repeated for 1861 and 1871 in conjunction with
Michael Anderson's 2% national sample for 1851. When the raw occupation
and birthplace information is recorded into a manageable set of categories,
migration, class and probable occupational life styles can be included as
variables. In addition, parishes will be used as units of analysis with the inclusion
of parish-level data from non-census sources.
The primary purpose of this chapter has been to present some of the metho-
dological problems (and substantive potential) of social history in general and
of census recoding in particular. Comparative sociology requires both inter-
societal research and intrasocietal studies over time. With the relative decline
in foreign area sociological field research in the 1970's due to problems of
funding and access (compared to the 1960's) and the coming of accessible age
of usable 19th century censuses, a new era of social history is developing.
ADAMS, BERT N. Kinship in an Urban Setting. Chicago: Markham Publishing Company, 1968.
ALDOUS, JOAN. "Urbanization, the Extended Family and Kinship Ties in West Africa," Social
Forces, October, 1961.
ANDERSON, MICHAEL. Family Structure in Nineteenth Century Lancashire, Cambridge University
Press. Cambridge, 1971.
BANKS, J. A. "Historical Sociology and the Study of Population," Daedalus, Spring, 1968.
BANKS, J. A. Prosperity and Parenthood. Routledge and Kegan Paul, 1954.
BERKNER, Lrrz K. "The Stem Family and the Developmental Cycle of the Peasant Household:
An 18th Century Austrian Example," American Historical Review 77 (1972): 398-418.
BLUMIN, STUART M. "Rip Van Winkle's Grandchildren: Families and Households in the
Hudson Valley 1800-1860." Unpublished manuscript, 1972.
BURCH, THOMAS. "Some Demographic Determinants of Average Household Size," Demography,
COHN, BERNARD S. "An Anthropologist Among the Historians," The South Atlantic Quarterly,
DEMOS, JOHN. "Families in Colonial Bristol, Rhode Island: An Exercise in Historical Demo-
graphy," William and Mary Quarterly, XXV, January, 1968: 40-57.
DRAKE, MICHAEL. "The Mid Victorian Voter," Journal ofInterdisciplinary History, Vol. 1 No. 1.
DRAKE, MICHAEL. Population in Industrialization. Methuen, London, 1969.
DYos, H. J. Victorian Suburb. Camberwell: Leicester, 1961.
GLASS, D. V., and D. E. C. EVERSLEY. Population in History. London, 1965.
GLASS, D. V. London Inhabitants Within the Walls 1695. London Record Society, 1966.
HAMMEL, E. A., AND LASLETT, PETER. "Comparing Household Structure Over Time and
Between Cultures," Comparative Studies in Society and History, Vol. 16, No. 1 January, 1974.
LASLETT, PETER. "Size and Structure of the Household in England over Three Centuries,"
Population Studies (Cambridge), July, 1969, Vol. XXIII, No. 2, p. 199.
LASLETT, PETER. The World We Have Lost. New York, 1964.
LASLETT, PETER. "The Comparative History of Household and Family," Journal of Social
History, Fall, 1970, Vol. 4, No. 1.
LASLETT, PETER. Household and Family in Past Time, Cambridge University Press, Cambridge,
LEES, LYNN. "Social Change and Social Stability Among the London Irish, 1830-1870." Ph.D.
thesis in History, Harvard University, 1969.
LEES, LYNN. "Patterns of Lower Class Life: Irish Slum Communities in Nineteenth Century
THEORY AND METHOD IN HISTORICAL SOCIOLOGY 115
London." Pp. 357-385 in Stephan Thernstrom and Richard Sennett Nineteenth Century
Cities, Essays in the New Urban History. New Haven: Yale University Press, 1969.
LrrWAK, EUGENE. "Occupational Mobility and Extended Family Cohesion," American
Sociological Review, February, 1966.
ROWNEY, D. K. and J. Q. GRAHAM. Quantitative History. Dorsey, 1969.
SENNETT, R. Families Against the City. Harvard University Press, 1970.
SMELSER, NEIL. Social Change in the Industrial Revolution. University of Chicago Press, 1959.
WRIGLEY, E. A. (Ed.) An Introduction to English Historical Demography from the 16th to the 19th
Century. Basic Books, 1966.
WRIGLEY, E. A. "Historical Population Studies," Daedalus, Vol. 97, Nol 2, Spring, 1968.
WRIGLEY, E. A. (Ed.) Nineteenth Century Society Essays in the Use of Quantitative Methods for the
Study of Social Data, Cambridge, Cambridge University Press, 1972.