ceremonies. Contemporary influences include subsistence activities, reflecting stark economic conditions on many Indian reservations. In the absence of employment opportunities on the reservation, many residents seek jobs elsewhere. Since their efforts may still be contributing to the economic well-being of the family at the reservation, “enduring ties” may drive both the worker and family members to consider the family home as the “usual residence” even if the worker is away most of the time. Bonvillain (1989), in an ethnographic study of the St. Regis Mohawk tribe for the 1990 census, found that men of the tribe are in high demand by construction companies for their worksmanship. Hence, groups of men from the Mohawk tribe regularly set up temporary group households at their employment sites, returning to St. Regis when they could.
An examination of complex ties between people and residences is incomplete if it does not consider people with ties to no fixed residence in addition to those with ties to two or more residences. As a panel, we have not weighed recommendations on the exact manner by which the street homeless population should be counted, a group that also includes those making use of shelters and relief facilities like soup kitchens. However, we do briefly comment on the nature of homelessness and the operations by which the census has tried to include them in the counts.
A basic question we have tried to answer for the various groups and living situations profiled in this report is the size of the population in question. But the basic question—“how many people are homeless in the United States?”—is very difficult to answer. Homelessness is not necessarily a permanent state; instead, people and families living in impoverished conditions experience homelessness on an episodic basis. Thus, a definition that considers homelessness strictly as a point-in-time phenomenon—portraying only a snapshot of “persons literally without a roof over their head, or forced to sleep in public or private shelters” at a single instance—is unduly restrictive and “seriously underestimates the level of homelessness in society” (Kusmer, 2002:4).16
Wright and Devine (1992b:212) summarized the measurement difficulties:
Indeed, some advocacy groups like the National Coalition for the Homeless oppose “the release of a separate ‘count’ of people enumerated in homeless situations (at selected service sites and identified outdoor locations) because such a number would be, by its very nature, both inaccurate and misleading, and therefore lead to uninformed decision-making by policymakers.” See “NCH’s Position on the Census and Homelessness” at http://www.nationalhomeless.org/publications/reports.html [8/1/06].