Geoscience Reference
In-Depth Information
neighborhoods reach a critical mass of gay representation, the representation
accelerates. In line with the pioneer viewpoint, Collins contends that it is not
amenity value, but rather cheap land values and access to night life and services,
occurring in areas with extensive physical decay and vacancies, that attracts the first
gay “pioneers”. This initial concentration of gays subsequently attracts more gays,
generating the renaissance of the neighborhood. (He also argues that in many cases
the initial settlement of gays in a neighborhood is an “accident” that subsequently
attracts more gays, more services for gays, and then yet more gays.) Other
neighborhood characteristics that have been deemed attractive to gays include
high vacancy rates (Ruting
2008
) and diversity (Gates and Ost
2004
).
Using 2000, 2005-2009, and 2010 US Census/American Community Survey
data on census tracts, we provide the first statistical analysis and evidence on the
causes and effects of the concentration of gays by gender in the neighborhoods of
the largest U.S. cities. The next section discusses the data and the specification of
the dependent variable, the census tract shares of the city's gay partnerships. The
third section presents residential segregation indices and clustering measures for
gay men and lesbians, followed by a section discussing the characteristics of census
tracts that correlate with having more gay residents. The next section examines how
growth in population and income (as an index of housing prices) over a decade is
associated with gay representation in the census tract at the start of the decade.
Conclusions are presented in the final section.
19.1.1 Data and Methods
We study 38 large central cities that are located in 35 metropolitan areas (See
Tables
19.1
and
19.2
for a list of cities by metropolitan area). We examine
segregation of gays by gender, how characteristics of a neighborhood in 2000 are
related to shifts of gay households and of all households between census tracts
within the city over the subsequent decade, and how gay presence in 2000 within
census tracts is related to population and income of residents in 2009.
For our studies of segregation and spatial concentration of gay partnerships, we
use the 2010 and the 2000 U.S. Census short form data, rather than the sample data
provided by the Census long form or American Community Survey (ACS). The
measures of segregation or concentration of a population are basically analyses of
variation, and variation increases, especially for groups that are a small share of the
tract population, when the sample sizes are small. Census tract data from 2000 and
2010 are based on the total census population
2
and, therefore, yield more accurate
measures of segregation and concentration. Because analyses of segregation or
concentration at two different points in time do not require that the spatial units
(census tracts) have the same boundaries for both time periods, we compare the
2
We use census tract data that is from the short form (SF1) answered by all households in the
census. We cannot use these data to measure differences within tracts, however, because the
boundaries of the census tracts changed between 2000 and 2010.
