Key words: census, community, household, methodology; Belize. A census is one of the ... matters of policy-making, the politics of citizenship, and the delivery of ...
Human Organization, Vol. 56, No. 1, 1997 Copyright 0 1997 by the Society for Applied
Anthropology
0018-7259/97/010064-7S1.2011
Some Methodological Issues in Counting Communities and Households RICHARD WILK AND STEPHEN MILLER In this article we discuss some of the limitations of conventional census techniques that assign all individuals to a single household in a single community. In areas with high rates of mobility and where people may belong to several households, traditional census methods can lead to very deceptive results that are poor guides for policy making and the delivery of services. The article suggests some ways census methods could be improved, so they can yield more informative and useful results. Key words: census, community, household, methodology; Belize
A
census is one of the standard tools of ethnography, a project often done as a first step in a community study or as a preliminary to more detailed surveys. The census has also historically been an indispensable tool of state administration. Benedict Anderson says that it is an essential device used by states to "regulate, constrict, count, standardize, and hierarchically subordinate" people and their institutions (1991:169). As is clear from current controversies over census categories and procedures in the United States, a census is also a more immediately political issue, because the results are often used to regulate or shape the division or distribution of resources and representation, and to justify policies which aid or hinder particular groups. Burdened with this economic and political significance, a simple census procedure should therefore bear close scrutiny, for the assumptions used in designing and enumerating may have serious consequences further along in the research process, and when the results are used to formulate policy. In this article we will confine ourselves to one particular aspect of the standard census procedure -the enumeration of households. We argue that the standard tools of household survey tend to make unwarranted assumptions about the corporate nature of
Richard Wilk is Associate Professor of anthropology at Indiana University, SB 130, Bloomington, IN 47405, and Stephen Miller is a graduate student in the same department. Earlier drafts of this article were first presented at the annual meeting of the American Anthropological Association, San Francisco, November 1992, and to the Population Institute for Research and Training at Indiana University. The research described in this article was supported by a Fulbright research fellowship and a Grant in Aid from the Wenner-Gren Foundation. The census enumeration was done by Melissa Johnson. The authors thank various anonymous reviewers for their comments and critiques of earlier drafts.
64
HUMAN ORGANIZATION
household groups, and this has serious implications both in the simple process of counting population, and in more complex matters of policy-making, the politics of citizenship, and the delivery of services. The basic problem is that the conventional census insists that each person belongs to a single domestic group, and locates that "household in a single spot on the map in an identifiable "community." As social scientists, we usually recognize that there is a clear difference between the household as an enumeration unit, chosen as a matter of convenience for grouping individuals together, and the household as a social and economic group that varies in structure and significance (Bender 1967; Wilk and Netting 1984). But in practice when we use households as enumeration units, we make it far too easy to ignore or minimize this difference. A census can then become an obscure and difficult tool for understanding both households and communities, because it presupposes the essential nature of the very units of study. The census can build particular forms of bias into the architecture of inquiry, and by its very concreteness, can make other important phenomena difficult to find, measure, and assess. It also lends a sometime spurious authority to comparative demographic analysis, producing measures with great precision and little accuracy. Our goal in this article is not, however, to suggest doing away with households a s census units. Instead we demonstrate that culturally relevant features of households can be included in census work, so that demographic figures better represent social realities, and therefore better comparisons can be made. The point is that ethnographic work can be used to add variables to a census, so that household units become more relevant for applied policy analysis. The starting assumption of the simple census - that all individuals belong to a single corporate household - is the main issue we address here. It has become an especially problematic assumption in areas where rates of migration, commuting, and other forms of residential movement are high.
The issue of where an individual "belongs," and where they should be counted, has become a matter of public debate in regions where transnational migration rates are high and new forms of householding which are not limited by national or community boundaries are emerging (Smith and Wallerstein 1992; ~ a s c et h al. 1994). As numerous studies tell us, migrants often end up belonging to two or more households in different countries (e.g., Rouse 1991). And even within a single country, people may retain active affiliations with different groups, and live in them full-time only for seasons, weekends, or holidays. Identifying a single household "head," and a roster of membership among transnational migrants, is therefore often an exercise in creative accounting. It may seem that these problems are temporally and spatially specific, and that the census might still be a useful tool for more stable and settled 'traditional' groups. We would argue, however, that mobility and flexibility are characteristics of households in many settled peasant societies, and that corporate households with fixed membership may be as often a product of state intervention as a precursor to it (see Roseberry 1991; Sider 1991;Wilk 1991). Our very perception of "traditional" societies as stable, built from normatively defined domestic units and communities, is partly an artifact of our methodology. How can we talk about social change, when our yardstick for measurement is totally insensitive to important degrees of variation?
Defining the Household For many years anthropologists and sociologists have debated various ways of defining household units in ways that are both culturally specific and universal enough to allow comparison (Bender 1967; Hart 1992; Wilk 1989, 1994). Wilk n ~ argue that households have no single and ~ e t t i (1984) universal function or activity, but everywhere combine different sets of domestic activities from a larger corpus of possibilities, and this set must be determined in each society before households can be counted. They argue that by choosing in advance a single defining characteristic of all households, theoretical assumptions are built into the units of measurement and analysis, making empirical testing of those assumptions impossible. If we begin a study with an apriori list of "domestic functions" that define the household we are making just this error. An example would be Friedman's (1984) argument that households are defined by the possession of a common and pooled fund of income and resources. A group of people who live together, cook and eat together and work on the same farm, might not be a household by this definition (cf. Smith 1984). A number of authors have pointed out the ethnocentric, western bias that lies behind many definitions of "household" (e.g., Guyer 1984; Collins 1986; Folbre 1994; Kabeer 1994). These assumptions can be especially pernicious when the household is then assumed to be the appropriate social unit for intervention and assistance. The often separate interests, resources, and needs of men and women, young and old, owners and clients, within household units cannot be recognized if households are counted as uniform and homogenous and undifferentiated units. The problem here is not really with the concept of "household" as a domestic unit. The problem lies with the notion
of corporateness, with the idea that society is invariably composed of bounded units with restricted membership, social rules of their own, systems of leadership, common property and a social charter. The metaphor of a society as a complex structure of molecular units, each in turn built of individual human atoms, is a venerable and convincing one. Much of early sociology was devoted to the idea that in pre-modem societies people are incorporated into corporate social units that obviate individual action (Grannovetter 1985). The historical force of this argument has been blunted by a number of studies that question the corporate nature of households in pre-modern Europe (e.g., Sabean 1990) and a wide variety of non-western cultural settings. But the habits and assumptions spawned from the assumption that households are always corporate units are still with us. We still expect interactions within a household to be qualitatively distinct from those outside; we expect a clear boundary around those who belong to the group and those who do not; we expect to find possessions that are owned or held in common, and we expect clear differentiation of roles according to position in the household (i.e. "household head" or "husband"). The contradictions between the standard assumptions about households and observed reality are striking in Belize where we have conducted research on household organization and migration. Carrying out a household census in Belize, as in many Caribbean and Latin American areas with high rates of mobility, poses constant challenges to the idea that society is built of corporate household units. Some households in Belize are clearly corporate units as we have defined them above. In some social settings, households really do resemble the "little commonwealth" of the North American cultural imagination (Demos 1970; Coontz 1992). Every household has a membership roll, and every person belonged to a single household. But these are extreme and unusual cases (deserving careful study and explanation on their own). In most of the world, the corporateness of the household is a matter of degree, and various kinds of rights, duties, activities and functions are combined in great variety. One challenge is to find ways to describe and classify domestic social units according to their degree of corporateness, for this is clearly an important dimension of variability. For example, the highly bounded and corporate households that Wilk (1991) studied among the Kekchi Maya in southern Belize were income-pooling units that were responsible for the support and welfare of all member. This included support of the elderly and disabled. It also meant that the benefits of loans, training, education, and health interventions provided to one household member could reasonably be expected to improve the living standards of others in the household (though not necessarily in equal proportions). On the other hand, the noncorporate and loosely-bounded households of urban Creole Belizeans do not generally pool many important resources and income streams, and do not have the same kind of obligations to less-able members. Benefits to some members of the household do not result in improved welfare to all household members (see Catzim 1993; UNICEF 1995). The corporate welfare distributing nature of households is clearly an extremely important variable for understanding a whole host of social and economic problems. If a census counts all households as if they were identical social and economic units, what have we actually counted? VOL. 5 6 . NO. I S P R I N G 1997
65
Non-Corporate Households The literature on domestic organization in the Caribbean region presents a clear example of how reliance on census categories leads to intractable problems of description and classification. In their insistence on such diagnostic features as a single place of residence for each individual and the identification of a single individual as the head of household, Caribbean censuses have distorted social realities and have provided inaccurate information for policy making. In a subtler way, the use of Euroamerican census categories have supported notions that Caribbean households are "disorganized" or "broken," and that Caribbean cultures reinforce patterns of poverty. Most scholars working in the Caribbean realize that working class households there are constituted in fundamentally different ways from those of the European colonizers. But the terms through which we have tried to describe this difference are all based on the significant characteristics of an idealized European model. European corporate households can be described and categorized on the basis of the kin composition of their membership, their size and their authority structure. Describing or classifying Caribbean households with these categories leads us astray because these are simply not always the major structural features of the units. As R.T. Smith notes; "The idea of an individual living in a room,' eating at another house with the person who is 'boarding' him, visiting a lover for sexual relations and sending clothes to be washed somewhere else by a 'washer,' is a perfectly normal one for this society and always has been" (1978:353). An accurate and appropriate census of Caribbean households must therefore be sensitive to the non-corporate nature of many households, and of the fluidity of individual affiliation. It must also account for differing intensities of involvement within the household, and membership on the part of individuals not physically present, including long-term and short-term migrants. 1n some extreme cases it must even be sensitive to the social presence of the dead (Kerns 1983). Because one intention of a census is simply to count the size of population, it is tempting to just go ahead and use the local designation for a minimal domestic unit (in the Creole Belize as elsewhere in the Caribbean this is often called a "yard") as a proxy for the household. The decennial census conducted by the government of Belize follows this procedure, and defining a household as a group that "usually" resides together and is synonomous with "house" (CARICOM 1984). This is an adequate procedure for locating and counting individuals, but is not sufficient for locating them geographically and socially. This is because yards are not always corporate units, and the notion of "residence" incorporated in the census method is not appropriate for a sizable minority of the population. For many people, being counted in an individual household in a particular community or neighborhood is no more than a polite fiction. By accepting this fiction we sacrifice accuracy in the interests of parsimony and order.
from donor households to receiving households. The assumption here is that the household left behind is affected by the loss of labor and income, and the migrant joins a new household in the receiving area. The complex spatial process, seen in this way, becomes a change in household membership. But it is clear that in migration, household membership is often the actual object of contest, the very issue at stake. A person may physically leave the house, but not leave the household. Migrants may go through a long period when they belong to two or more households, which they can try to keep separate (and even secret), or find ways to merge and blend. Migrants sometimes belong to separated households that differ radically in their rules, rights, duties, obligations, structure and system of authority (Pessar 1988; Olwig 1993). Migration may be as much a renegotiation of domestic arrangements as it is a movement in space. Migrants constantly dispute and debate the meaning of household membership, the terms of obligation, and the nature of domestic rights and duties (Rouse 1991). When we try to impose the order of the corporate household on migration, when we try to determine which household the migrant belongs to and decide which households have "lost" members through migration, we assume away the very heart of the matter. For these issues of membership and corporateness are the issues that often remain crucially ambiguous, negotiable and vitally important to migrants. Another example of how an assumption of corporateness can bias the study of domestic behavior can be found in the field of gender analysis. Much work on Caribbean households has been preoccupied with differences in male and female household roles; a regular conclusion being that women form the core of the corporate household, while men move often between these "matrifocal" groups (e.g., Gonzalez 1969; 1984; Young 1990). But there is also evidence that women, especially younger women, often lack a permanent attachment to a single household unit, just as the bonds between men and women shift and change frequently over the life course (Pessar 1988; Safa 1995). If we begin our analysis with corporate household units, categorize them into structural types and then look at how people are attached to them, we mess one of the most important gender issues in domestic life. There are clearly a number of very different folk models of household organization available to Caribbean men and women, and all domestic arrangements are the product of complex processes of negotiation. We have often confused the folk models (which include extended, matrifocal and nuclear forms) with objective "types" of corporate household. In reality, many individuals do not belong to corporate households at all, and may be involved in a variety of non-corporate domestic arrangements. Sexual relationships and gender roles are important in the constitution of these arrangements, but so are economic resources, age relationships, descent, fictive kinship, and ideas about respectability and religious morality. To simply define types of corporate household and correlate them with a single variable, from this perspective, is a doomed methodology. Clearly we need a more culturally appropriateway of measuring variability in domestic arrangements.
Migration and Gender More Appropriate Census Methods The use of the corporate household as a universal census unit has also been an impediment to understanding some wider processes. For example, migration is generally seen as a flow 66
HUMAN ORGANIZATION
Counting households is often one of the first things an anthropologist does in a small community, and in urban areas
existing government censuses may be used. We suggest, however, that the results of simple counting of yards, houses, apartments, o r other visible (emic or etic) units is liable to produce figures that are difficult to interpret and use. They will also probably result in total counts of population size and density that are not strictly comparable. An example is easy to find in Belize. The 1990 national census insisted that a person must live in a house for five days a week in order to be counted as a member of that house and community. One rural community is a short bus ride from the city, s o workers and students commute daily, and are counted as residing in the village even though their village has really become an extended neighborhood, a bedroom suburb, of the city. Another village a bit further from the city has no daily bus service; many workers and students there spend three or four nights a week in the city lodging with relatives o r in small apartments. These commuters are mostly counted as living in the city, and the village where they have their property, families, houses, and farms therefore appears in the census as a very small place, much diminished in size from the previous census (when there was no bus service at all, and no commuting). In reality the number of people that can be counted in any community is always fluid; there are daily, weekly, and even seasonal fluctuations. If the first village is counted on a weekday afternoon, it is larger than the second village. If counted on a Saturday morning, the second village is much larger. If the goal of a census is not simply to count noses, but to group people into socially and economically significant units, we need to define those units a locally meaningful way before we can set about counting them (see Wilk and Netting 1984). Rather than being an end in itself, a census can be seen as a tool in defining, locating, and counting these units, and the relationship between individuals and groups. In other words, rather than operationalizing an a priori classification, a census can generate data, which can be combined with other sorts of information on domestic behavior, in a process of discovering relevant categories. There can thus be a feedback process, with multiple censuses evolving over the fieldwork period. In Belize we found that we could use existing studies of Afro-Caribbean domestic behavior for a preliminary census, which was then refined through ethnographic interviews, small surveys, and observation, in order to design a final census that truly counted significant domestic units and relationships. Our census was designed to measure the following kinds of variability:
Fricke 1986). In this case, the problem of placing an individual in a single location remains, if that individual regularly moves, or keeps a domestic affiliation in several places. An alternative approach is to keep the domestic group as a census unit, but to allow different kinds of membership in the group, and then to link households which have overlapping membership. We chose this method, as part of a long-term study of migration, consumption, and remittances in a rural Belizean community, which began in 1990. Crooked Tree is a predominantly Afro-Caribbean Creole community, located on an island in a system of inland freshwater lagoons in north-central Belize. T h e village was once predominately agricultural, though a significant minority have always left the village seasonally o r for extended periods of wage labor in Belize City or other rural areas. Today many people from the community have jobs o r attend schools in nearby cities, and commute daily, o r have split residences and spend weekends and holidays in the village. Migration to developed countries, and to Chicago in particular, has also become extremely common, and the population pyramid looks like an hourglass. Return migration is also common; 21% of the adults in the village in 1990 had lived outside of Belize for more than three months (15% in the United States), and 61 % had lived in other parts of Belize. In June of 1990 we enumerated 116 domestic units, socially defined as groups sharing a single plot of land (yard), which pool their income and are socially recognized and named (usually for the senior female in residence). Often there were several discrete houses and other structures in a yard, some temporarily uninhabited but the social boundaries around the units are physically marked with barbed-wire fences. The social group which participates in domestic activities within the yard is quite variable, and only some can be considered closed, corporate entities. Working closely with informants, we were able to define four categories of household membership. I . Full-time residents sleep in the unit every night, and contribute a major portion of their income to a domestic fund. However, they do not necessarily eat all of their meals there. 2. Part-time residents have another residence outside the community, and sleep in their Crooked Tree household a minimum of once or twice a month. They are expected to contribute a smaller portion of their income to the domestic fund, and have the right to become full-time residents in the future. This status is usually considered temporary, for students or seasonal workers, but in practice it may last for many years. 3. Absent members live full-time in another community, but retain the right to return and become residents in the Crooked Tree household. They are expected to visit at least every two years, and to contribute cash or goods to the household at Christmas, and in cases of emergency. Absent members can send their children to live in the household, and can have other kinds of negotiable and variable obligations and rights. They are expected to keep in touch through letters, phone calls, or sending messages with other visitors.
1) Degrees of household corporateness: how closely tied to a domestic unit are different individuals? To what extent is a group recognized as bounded?
2) Different kinds of household membership: are there
categories or degrees of household membership? What bundles of rights, obligations and commitments exist? 3 ) Overlapping and multiple household membership: can physically absent people still belong to a household? Can people belong to more than one group at a time? What kinds of social ties bind people to different or distant groups (ie. descent, sexual, legal).
4. Visitors are people who do not sleep in the household,
but who have a specific long-term economic and social relationship with it. Most often they are what we called "people who eat but don't sleep." These are young men or retired people who have their own house and yard,
O n e way to deal with the problems of the conventional household census is to switch to an approach that counts only individuals, and tracks their relationships and movements (see VOL
56, NO. 1 S P R I N G 1997
67
but go to a relative's house regularly for evening meals, or have the meals brought to them. Young men or women may also become visitors in a household where they have a girlfriend or boyfriend and/orchildren. A visiting member may receive other services like clothes washing, use of tools, and storage of personal effects, and they may give contributions of labor, goods, and cash to the household. Individuals can combine these kinds of membership in different households in various ways, and they move between them during their lives. One can be a full or part-time member of one village household, and have multiple visiting relationships in others. A person can be an absent member of one village household, while they are a visitor in another. When w e used these categories in our census, we got a different view of the community from what we would have had if we had insisted on nominal categories of membership. The very category of "the village" itself emerges as a debatable, contentious, and both spatially and temporally unbounded. Using this census we have also been able to ask some important questions about the relationship between household strategies and economic fortunes. The total breakdowns for the 116 domestic units is given in Table 1. Mean household size can be calculated as 4.53persons or 6.36, depending on what kinds of members are included in the household. The strictest method of counting shows 525 village members, while the most inclusive gives a figure of 755. With the totals broken down in this way we can develop household typologies that make sense in terms of different kinds of economic strategies, and different levels of economic success. The full details of this analysis will be presented elsewhere; here we include just two tables that show some of the results. Table 2 breaks down village households by the kinds of members they have. The crosstabulation in Table 3 shows that households with part-time members are more likely to also have visitors. This suggests that some households are consistently following an 'open' strategy of incorporating non-full time members. In our census we also collected some very basic information on the kinds of appliances and goods kept in households. The product is a "commodity index" ranging from 0 to 4,which can be used as a very rough measure of household wealth. The right column on Table 2 shows that there are differences in the index in households with different kinds of membership. Households with part-time members have above average indices of wealth, and households with more part-time members have more wealth than those with few part-time members. The affiliation of absent members with a household is also strongly linked with higher wealth. The wealthiest households (commodity index=3) had an average of 3.1foreign household members, while the poorest households had an average of .9 foreign household members each. These relationships are probably reciprocal instead of causal; extending household membership may effectively increase wealth (as suggested by Selby et al. 1990), and at the same time, better-off households are more likely to attract part-time members and retain the affiliation of migrants. This analysis and other data show that household members who would not be counted by a regular census are nevertheless vitally important in the household and community economy. Their participation and membership cannot be measured by physical presence on the day or month of the actual census count. 68
HUMAN ORGANIZATION
Table 1.
Basic Household Membership Data for Crooked Tree, Belize
Households (#) Total full time residents (#) Mean size, full time residents (#) Total Part time residents (#) Mean Part time residents (#) Households wIPTRs (%) Total Foreign residents (#) Mean Foreign residents (#) Households w/FRs(%) Total Visitors (#) Mean Visitors (#) Households wlvisitors (%) Total Full time + Part-time Mean household size (#) Total Full + Part + Foreign Mean Household size (#)
Table 2.
(#) (#)
Household Counts by Types o f M e m b e r s
Full time members only Part time Visitors Foreign members Part time + Visitors Visitors + Foreign Part time + Foreign members Part time + Visitors + Foreign ALL
Table 3.
Households
Mean Wealth
(4%)
(#)
116
84
Crosstabulation o f Household C o u n t s
Households with part-time members none 1 or more
Households with Visitors 1 or more 81 6
none
All 87
All
Conclusion Our point is that household form and structure are related in complex ways to migration, labor markets, and transnational connections, as a number of others have also argued (e.g., Gisbert etal. 1994; Smith and Wallerstein 1992). The growing interrelatedness of rural, urban, and migrant economies is concealed or obscured by methods which still assume that every person has a single place and every household has a stable roster of resident members. Census methodology often assumes a
degree of corporateness and boundedness that can no longer be justified in many regions of the world. We suggest that census methods can and should be adjusted, using ethnographic methodology, in order to describe and analyze this phenomenon. and to serve as a guide for more effective policy making. At the very local level, the miscounting of population with inappropriate household membership categories can lead to misallocation of resources. In Belize, for example, populationweighted measures are used in assessing priorities for rural water supplies, electrification, road improvement, rural health care delivery, and other services. Communities like Crooked Tree, which have large numbers of part-time and temporarilyabsent members, are consistently under-counted and under served as a result. And given the hourglass-shaped age distribution of the village population, the proportion of very young and very old in the community (those most urgently in need of social and health services), is much higher than in communities with fewer absent or part-time members. In other areas, reliance on census figures could easily lead to other forms of misallocation and poor distribution of resources. A t the national level, the counting and allocation of population to households and communities has important political dimensions. In Belize there is considerable official and public alarm over changes in the ethnic composition of the country (Vernon 1990). Yet the decline of the Afro-Caribbean proportion of the population is at least partly an artifact of a census method that does not count those who are studying or temporarily working abroad. The whole issue of ethnic balance really hangs on the definition of residence. The citizenship status of local sojourners in other countries is also much debated in Belize, as elsewhere in the Caribbean where the local rights of "born there" migrants are often challenged by expatriates and new migrants from other areas (Cohen and Macia-Lees 1993). Belizean politicians often appeal to the overseas Belizean-born community for money and technical assistance, but they are much less willing to follow through on the issue of enfranchising foreign residents and giving them local representation. Residence and citizenship are complex and often dangerous legal and political issues; only anthropologists seem willing, at this time, to address the fundamental issues involved at the level of the community and household. This is yet another case where local realities often directly contradict the fiction of official counts, measures, statistics, and categories.
REFERENCES CITED
Anderson, Benedict 1991 Imagined Communities. Second edition. London: Verso. Basch, L., N. Schiller and C. Blanc 1994 Nations Unbound. Amsterdam: Gordon and Breach. Bender, Donald 1967 A Refinement of the Concept of Household: Families, Coresidence, and Domestic Functions. American Anthropologist 69-493-504, C ARICOM 1984 1980-81 Population Census of the Commonwealth Caribbean: Belize. Kingston: Statistical Institute of Jamaica. Catzim, Adele 1993 Putting Gender on the Global Agenda: Women DefineTheir Needs. In Globalization and Development: Challenges and
Prospects for Belize. Pp. 73-89. Belize City: Society for the Promotion of Education and Research. Cohen, Colleen and Frances Mascia-Lees 1993 The British Virgin Islands as Nation and Desti-nation: Representing and Siting Identity in a Post-Colonial Caribbean. Social Analysis 33:130-151. Collins, Jane 1986 The Household and Relations of Production in Southern Peru. Comparative Studies in Society and History 28: 65 1-671. Coontz, Stephanie 1992 The Way We Never Were: American Families and the Nostalgia Trap. New York: Basic Books. Demos, John 1970 A Little Commonwealth: Family Life in Plymouth Colony. New York: Oxford University Press. Folbre, Nancy 1994 Who Pays for the Kids? Gender and the Structure of Constraint. London: Routledge. Fricke, Thomas 1986 Himalayan Households: Tamang Demography and Domestic Processes. Ann Arbor: UMI Publications. Friedman, Kathie 1984 Households as Income-Pooling Units. In Households and the World Economy. J. Smith, I. Wallerstein and H. Evers, eds. Pp. 37-55. Beverley Hills: Sage Publications. Gonzalez, Nancie 1969 Black Carib Household Structure. Seattle: University of Washington Press. 1984 Rethinking the Consanguineal Household and Matrifocality. Ethnology 23: 1-12. Gisbert, Maria, Michael Painter, and Mery Quiton 1994 Gender Issues Associated with Labor Migration and Dependence on Off-Farm Income in Rural Bolivia. Human Organization 53(2): 110- 122. Grannovetter, Mark 1985 Economic Action and Social Structure: The Problem of Embeddedness.American Journal of Sociology 91 :48 1 -510. Guyer, Jane 1984 Naturalism in Models of African Production. Man 19:371388. Hart, Gillian 1992 Imagined Unities: Constructions of "The Household" in Economic Theory. In Understanding Economic Process. Sutti Ortiz and Susan Lees, eds. Pp. 111-130. Lanham, MD: University Press of America. Kabeer, Naila 1994 Reversed Realities: Gender Hierarchies in Development Thought. London: Verso. Kerns, Virginia 1983 Women and the Ancestors. Urbana: University of Illinois Press. Olwig, Karen Fog 1993 Global Culture, Island Identity. Chur, Switzerland: Harwood. Pessar, Patricia 1988 The Constraints on the Release of Female Labor Power: Dominican Migration to the United States. In A Home Divided: Women and Income in theThird World. Daisy Dwyer and Judith Bruce, eds. Pp. 195-215. Stanford: Stanford University Press. . Roseberry, William 1991 Potatoes, Sacks, and Enclosures in Early Modern England. In Golden Ages, Dark Ages. Jay O'Brien and William Roseberry, eds. Pp. 19-47. Berkeley: University of California Press. Rouse, Roger 1991 Mexican Migration and the Social Space of Postmodernism. Diaspora 1(1):8-23. VOL. 56, NO. 1 SPRING 1997
69
Safa, Helen 1995 The Myth of the Male Breadwinner. Boulder CO: Westview Press. Sabean, David 1990 Property, Production and Family in Neckarhausen, 17001870. Cambridge: Cambridge University Press. Selby, Henry, Arthur Murphy, and Robert Stephen Lorenzen 1990 The Mexican Urban Household: Organizing for Self Defense. Austin: University of Texas Press. Sider, Gerald 1991 House and History at the Margins of Life: Domination, Domesticity, Ethnicity, and the Construction of Ethnohistories in "The Land God Gave to Cain." In Golden Ages, Dark Ages. Jay O'Brien and William Roseberry, eds. Pp. 208-233. Berkeley: University of California Press. Smith, Joan 1984 Nonwage Labor and Subsistence. In Households and the World Economy. J. Smith, I. Wallerstein and H. Evers, eds. Pp. 64-89. Beverly Hills: Sage Publications. Smith, Joan and Immanuel Wallerstein eds. 1992 Creating and Transforming Households: The Constraints of the World Economy. Cambridge: Cambridge Univ. Press. Smith, R.T 1978 The Family and the Modern World System: Some Observations from the Caribbean. Journal of Family History 3:337-360.
HUMAN ORGANIZATION
UNICEF 1995 The Right to a Future: A Situation Analysis of Children in Belize. Belize City: National Committee for Families and Children. Vernon, Dylan 1990 Belizean Exodus to the United States: For Better or Worse. in SPEAReports #4. Belize City: Society for the Promotion of Education and Research. Wilk, Richard and Robert M. Netting 1984 Households: Changing Form and Function. In Households: Comparative and Historical Studies of the Domestic Group. R. Netting, R. Wilk and E. Arnould, eds. Pp. 1-28. Berkeley: University of California Press. Wilk, Richard 1989 Decision Making and Resource Flows Within the Household: Beyond the Black Box. In The Household Economy: Reconsidering the Domestic Mode of Production. R. Wilk, ed. Pp. 23-54. Boulder: Westview Press. 1991 Household Ecology. Tucson: University of Arizona Press. 1994 Inside the Economic Institution: Modeling Household Budget Structures. In Anthropology and Institutional Economics. James Acheson, ed. Pp. 365-390. Lanham, MD: University Press of America. Young, Virginia H. 1990 Household Structure in a West Indian Society. Social and Economic Studies 39(3): 147-179.
