Patterns of gay and lesbian partnering in the larger metropolitan areas of the United States.

Since the 1990s, there has been increased interest in the trends and patterns of gay and lesbian cohabitation, a phenomenon known also as same-gender (or sex) unmarried partnering. According to data from the 1990 U.S. census, households headed by same-gender unmarried partners represented as much as one sixth of one percent of all households enumerated (Fields & Clark, 1999). Data from the 2000 Census indicate a significant increase from 1990 in the actual numbers of gay and lesbian couples who report themselves in the Census (Armas, 2001; Cohn, 2001; Gillman & Langford, 2001; Poston, Gu, & Walther, 2003; Simmons & O'Connell, 2003; Smith & Gates, 2001).

The increased interest in the distributional patterns of gay and lesbian partnering is associated with increases in the visibility of gays and lesbians. For example, the homosexual community has become a political force on the national scene and in many large cities, pushing for local and national laws dealing with inheritance and health benefits, among other issues (Castells & Murphy, 1982; Knopp, 1990; Lauria & Knopp, 1985; Weightman, 1981). Areas where gays and lesbians have settled have become, according to O'Reilly and Webster (1998), "gay spaces" with political force and activism. Those U.S. metropolitan areas with the larger gay and lesbian settlements--for instance, San Francisco, New York, Houston, and Los Angeles, among others--are often the gay spaces that receive the most national attention. But there are concentrations of gays and lesbians in virtually all metropolitan areas of the country. However, with but a few exceptions (Black, Gates, Sanders, & Taylor, 2000, 2002; Poston et al., 2003), there has been little effort among social scientists to index these concentrations and examine the extent to which the indexes are associated with the social, ecological, and political characteristics of the areas.

In this paper we used data from the 5% Public Use Microdata Samples (PUMS) of the 1990 U.S. Decennial Census. We developed for all 92 metropolitan areas of the U.S. with populations in 1990 of 500,000 or more several indexes of homosexual partnering, one set for gay men and a second for lesbians. These indexes are based on information from the 1990 census questionnaire pertaining to two-person households with two same-sex adults in which one identifies him(her)self as the "unmarried partner" of the other. At the time this paper was written (August of 2003), the development of similar indexes using 2000 data was not possible. It will be possible to develop similar kinds of indexes with data from the 2000 U.S. Census when the requisite sample data are released.

In the 1990 census questionnaire, one person in each household was designated as the householder; this person was "the member (or one of the members) of the household in whose name the home is owned, being bought or rented" (Barrett, 1994, p. 16). A subsequent census question asked every other person in the household about his or her relationship to the householder. Among the responses was unmarried partner. From the 1990 5% PUMS we selected all two-person households in which the second member identified him(her)self as the unmarried partner of the householder, in which both the householder and the partner were of the same sex, and in which both persons were age 18 or older.

We constructed four indexes for unmarried partners who are both male and four for unmarried partners who are both female. Three of the four indexes for both genders differ only in denominator; the other index is a relative measure that compares the concentration in the metropolitan area to that in the country as a whole. After constructing these indexes of homosexual partnering for each large metropolitan area, we proposed and tested hypotheses to account for variability among and between the indexes, as well as between the indexes and ecological characteristics of the metropolitan areas.

DATA

Census Data

Before the 1990 U.S. census, the development of spatial indexes for the lesbian and gay populations would have been difficult, if not impossible, and could only have been done by indirect means. In 1982, for instance, Castells and Murphy noted that "there is no statistical source that provides information on sexual preferences of residents of specific urban area ... such an obstacle appears overwhelming to the researcher trying to understand the spatial dynamics of the emerging gay culture" (1982, p. 238).

This all changed in 1990 when the federal government decided that data were needed on the patterns of heterosexual cohabitation and added the unmarried partner response to the standard "relationship to head of household" question in the 1990 census questionnaire. In the 1980s, for instance, demographers had reported increases in the prevalence of heterosexual cohabitation (Bumpass & Sweet, 1989). But there was no way then to adequately gauge the actual levels of this phenomenon with decennial census data; there was no way to identify persons in the household who were unrelated to the head of household but who had a marriage-like relationship with the householder.

For instance, in the 1980 census, one of the responses to the relationship to the head of household question was "roommate"; persons indicating this response had no blood relationship to the householder, but may or may not have had a marriage-like (i.e., an emotional) relationship with the householder. As Black et al. (2000) wrote, "before 1990, couples living outside marriage in marriage-like relationships were not identified separately from (unrelated) individuals living together as roommates" (p. 140).

Fortunately, census procedures permitted respondents to check the unmarried partner response irrespective of whether his or her sex was the same as that of the householder. Thus, researchers are able now to use these data to obtain information on gay and lesbian partnering.

Limitations of the Data

The use of these decennial census data as measures of the prevalence of gay men and lesbians in a given area has some pitfalls. First, there may be a misunderstanding of the term unmarried partner. The Census Bureau defines an unmarried partner as "a person who is not related to the householder and who shares living quarters, and who has a close personal relationship with the householder," as opposed to a housemate or roommate who "is not related to the householder and who shares living quarters primarily in order to share expenses" and as opposed to a roomer, a boarder, or a lodger of the householder (U.S. Bureau of the Census, 1992, p. B 15). However, some people may not define a close personal relationship as necessarily sexual. For example, Phua and Kaufman (1999) found in the 1990 census partnering data for New York some householders with same-gender unmarried partners who were younger than 18, and other householders with more than one same-gender unmarried partner; they found still other householders with one same-gender unmarried partner and an opposite-gender person reporting being currently married to the householder. They reported that "one cannot simply dismiss these variations as misreporting which would be superimposing one's interpretation of what 'unmarried partners' meant to these individuals" (Phua & Kaufman, 1999, p. 365).

However, Black et al. (2000) have argued that these census data may indeed be quality data for gay and lesbian unmarried partners. They have written that they

   read the evidence as suggesting that virtually none of the
   misclassification of "unmarried partner" status is due to
   a random error or to confusion about the term, but is due
   instead to mistakes concerning the relationship of an
   individual to the householder, if in fact "unmarried
   partnership" was generally being interpreted correctly,
   then misclassification is negligible when, as in the case
   for the vast majority of gay and lesbian couples, there are
   only two adults in the household. (p. 147)

Chevan (1996) nevertheless has argued that heterosexual unmarried partners may indeed mark "housemate/roommate" to avoid taboos associated with cohabitation, and this could be even more prevalent among same gender partners (Spanier, 1982). However, Casper (1992) has suggested there is probably at best only a slight taboo associated with marking unmarried partner on the census questionnaire.

Another question we need to ask is whether data on same sex unmarried partners are the same as data on self-reported gay and lesbian couples. Phua and Kaufman (1999) have written that same-gender unmarried partners may not self-identify as gay or lesbian. They write, "[B]y using the census data, one cannot assert that same-sex cohabiters are either gay men or lesbians" (pp. 376-377). While self-identification is an important consideration for gays and lesbians, Black et al. (2000) compared the number of self-reported gays and lesbians in the General Social Survey (GSS) and some other surveys with the numbers of same-gender unmarried partners reported in the 1990 U.S. census. They found that "roughly 35% of men living as partnered same-sex couples are recorded in the Census; for women the corresponding fraction is 29%," which is lower than the GSS estimated rates (Black et al., 2000, p. 147). Certainly, therefore, the number of gay and lesbian unmarried partners reported in the census data are underestimates of the numbers of self-reported gays and lesbians (Fields & Clark, 1999; Smith & Gates, 2001).

Another problem arises with regard to the editing practices of the Census Bureau. Following the stipulations of the Federal Defense of Marriage Act, the U.S. Congress defines marriage as a relationship involving opposite sex people. If a person marks his or her relationship to a householder who also is of the same gender as husband or wife, the Census Bureau is required to edit this respondent into the unmarried partner category. Fields and Clark (1999) examined the editing process of the Census 2000 dress rehearsals conducted in April 1998 in Sacramento, California, and in South Carolina. They found that same-sex couples who marked spouse were older, had children, owned property, and were less likely to have moved in the last 5 years. They wrote: "[I]t is clear from the examination of these unedited data that households which are identified as 'married couple' same gender households are a distinct group from households which are identified as unmarried partner same gender households" (p. 10). Thus, it would appear that the editing of "married" couples of the same gender into the unmarried partner category may result in inflating the numbers as well as ascribing demographic characteristics which are different for both married couples and unmarried partners of the same gender.

Of course, a major problem using the 1990 decennial census data to develop a sample of gay men and lesbians is that the resulting data are "not a random sample of people who would identify themselves as gay or lesbian, nor ... a sample of those who have engaged in same-sex sex, because the sample (by definition) contains only individuals who are involved in a cohabiting relationship" (Black et al., 2000, p. 140). The counts of gays and lesbians we developed thus are conservative estimates of the actual numbers of gays and lesbians in the population. We noted earlier that Black et al. (2000) have examined the coverage of gays and lesbians in other national data-sets and have concluded that "only about one-third of gay and lesbian couples report themselves as such in the census" (p. 153).

Although social scientists have noted the above methodological issues, we still hold that the same-sex unmarried partner data cannot be dismissed. Black et al. (2000) have noted that the number of gay and lesbian unmarried partners may not reflect the entire population of gays and lesbians, but it may be reasonable to believe that the gay and lesbian unmarried partners in the sample are more likely to be "out" because of their having marked the unmarried partner response on the census form. However, a comparison of the census data with other data sets such as the GSS allows one to conclude that the census data are somewhat less efficient in enumerating and describing the number of gay or lesbian couples. Therefore, the number of gay and lesbian unmarried partners represented in the census data is certainly an underestimate of the total population of gays and lesbians.

However, our concern is not so much with the number of gay and lesbian partners counted in the census data, but with the variation in the indexes of gay and lesbian partners within and among the U.S. metropolitan areas. Therefore, our intention in this study was to develop indexes of gay and lesbian partnering in large metropolitan areas and to relate the variation in these indexes with each other as well as with the variation in ecological characteristics of the areas. Our concern thus has to do with the validity of the variation in the gay and lesbian rates across the metropolitan areas.

How valid is this variation? There are no reliable data available to answer this question. There are no data other than census data "for calculating even the most rudimentary statistics on the (geographic) locations of the gay and lesbian populations" (Black et al., 2000, p. 149). However, one can examine the face validity of the census-developed geographical distribution data of the homosexual population by relating its variation with that of the spatial distribution of AIDS deaths. Black et al. (2000) have done this and have shown the two to be highly correlated.

We have extended the work of Black et al. (2000) regarding the validity of census-based statistics on the homosexual populations of the U.S. metropolitan areas. For each of the 92 metropolitan areas in our sample, we took the count of reported AIDS cases for 1981 to 1990 and divided this number by the number of unmarried men ages 15 and over. We then correlated these AIDS proportions with the homosexual partnering indexes for the metropolitan areas. Our correlations were in the range of .30 to .40.

These tests somewhat increase our confidence in the quality of the gay and lesbian data obtained via the 1990 decennial census, particularly the validity of the geographical distribution of these data. We turn next to the development of indexes of the gay and lesbian population for the major metropolitan areas.

INDEXES OF THE GAY AND LESBIAN PARTNERING POPULATIONS

Prior Research on Index Construction

With the exception of the above-cited analyses by Black et al. (2000, 2002) and Poston et al. (2003), we know of no studies in the social science literature that attempt to develop indexes of the prevalence of the homosexual population. The study by Black et al. (2000) is extensive and well-known. They developed gay and lesbian partnering indexes for the 20 metropolitan areas with the largest gay and lesbian populations. They first used the 5% Public Use Microdata Samples (PUMS) from the 1990 census to obtain data on gays and lesbians who were in what they termed "marriage-like" relationships. They next examined the geographic distributions of the gays and lesbians in their sample. They found that almost 60% of their gay sample resided in 20 metropolitan areas, and that over 45% of their lesbian sample resided in 20 metropolitan areas (16 of which were among the 20 inhabited by gays).

Their data indicated that Los Angeles contained 9.8% of all the gay partners in their sample, New York 8.4%, and San Francisco 7.9%. The same three cities also contained the most lesbian partners of all the cities in the U.S.: New York, 6%, Los Angeles, 5.4%, and San Francisco, 3.4%. Their data are interesting in their own right and provide further evidence that the census data on gays and lesbians do not consist "predominantly of mis-recorded opposite sex couples" (Black et al., 2000, p. 148). However, their percentage indexes are highly associated with the size of the cities. The largest cities tend to have the largest numbers of gay and lesbian partners.

Development of Numerators and Indexes

We began our data-gathering by first identifying the largest metropolitan areas in the U.S. as those with total populations in 1990 of at least 500,000 persons; there are 92 such metropolitan areas (they are listed alphabetically in the Appendix). For each of these areas we then used the 5% PUMS from the 1990 census and extracted all two-person households in which the second member identified him(her)self as the unmarried partner of the householder, in which both the householder and the partner were of the same gender, and in which both persons were of age 18 or greater. For each metropolitan area we multiplied the number of male-male households by 2 to produce an estimate of the number of gay male partners for the area; we did the same for the number of female-female households to obtain an estimate of the number of lesbian partners for the area. These data constituted the samples of gay and lesbian partners for each of the 92 metropolitan areas.

How should one index the prevalence of gay and lesbian partners in a metropolitan area? One way would be to produce incidence rates by determining the number of gays or lesbians out of those who comprise the demographic and statistical population from which gays and lesbians are drawn. In demography, a rate is defined as the number of persons experiencing an event at a given time (the numerator) divided by the population at risk of the event (the denominator). For example, a rate of first marriage for an area is the number of persons in the area in a given year marrying for the first time divided by the total number of unmarried persons in the population of the area above a certain age, usually 18. Likewise an out-migration rate for an area in a given period of time is the number of persons moving from the area in a given period divided by the number of persons living in the area at the start of the period (see Hinde, 1998, pp. 80-81,193). These rates measure the incidence or prevalence of first marriages and out-migrations occurring in the area. Geographical areas may then be ranked on the basis of the numbers of first marriages or out-migrations per 1,000 persons who are at the risk of a first marriage or an out-migration. It is important when calculating incidence or prevalence rates that the persons in the numerator correspond with those in the denominator; that is, the denominator should include the persons in the numerator. This is known as the principle of correspondence (Hinde, 1998, p. 4).

Four Indexes of Gay and Lesbian Partnering

Following the above examples, one way to develop a rate of gay (or lesbian) partnering for an area at a particular time would be to divide the number of gay (lesbian) partners by the number of unmarried males (females) in the population of people age 18 or older. The denominator would be restricted to unmarried persons, because according to the statistical and demographic definitions used here as well as Census Bureau coding procedures, married persons are by definition heterosexual and would thus not be "at risk" (in a statistical sense) of being a gay or lesbian partner.

Thus, our first index of gay (lesbian) partnering, Gay (Lesbian) Index 1, is a straightforward rate with the statistically and demographically most appropriate denominator, namely unmarried males (females) ages 18 and up. The denominator is most appropriate because only unmarried persons are statistically at risk of being in the numerator. This index is defined as follows:

Gay (Lesbian) Index 1 = [# of gay (lesbian) partners/# of unmarried males (females) age 18] 1,000

We used the constant of 1,000 so the rate would be expressed as a whole number rather than as a proportion. The first two rows of Table 1 present descriptive information on this first set of indexes for the gay and lesbian populations. The first two data columns of the Appendix list the values for this index for gays and lesbians for each of the 92 metropolitan areas.

Index 1 for gays had a mean value among the 92 metropolitan areas of 12.9 (Table 1). This means that across the largest metropolitan areas of the U.S. in 1990, there was an average of almost 13 gay cohabiters for every 1,000 unmarried men of age 18 or over. San Francisco had the highest value on Gay Index 1 with a score of almost 102; for every 1,000 unmarried men of age 18 and over in San Francisco in 1990, 102 of them (more than 10%) were in a gay partnering relationship. Birmingham had the lowest Gay Index 1 value, less than 1 per 1,000.

This first index for lesbians (Lesbian Index 1) had an average across the metropolitan areas of just under 10 per 1,000. San Francisco had the highest value of all the metropolitan areas with a score of 34.2; just over 3.4% of unmarried females age 18 and older in San Francisco were in lesbian partnering relationships. Birmingham had the lowest score, just slightly less than its lowest score on Gay Index 1.

Our second gay (lesbian) index uses the same numerator as the first index, but its denominator is comprised of heterosexual partners. The index indicates the number of gay (lesbian) partners per 1,000 heterosexual partners. Following demographic convention, this index is a ratio, not a rate, because the denominator does not contain the population that is captured in the numerator. This second index of gay (lesbian) partnering, Gay (Lesbian) Index 2, is defined as follows:

Gay (Lesbian) Index 2 = [# of gay (lesbian) partners/# of heterosexual partners] 1,000

Index 2 for gays (Gay Index 2) had a mean value among the 92 metropolitan areas of 31.6 (Table 1). Among the largest metropolitan areas of the U.S. in 1990, there were on average almost 32 gay cohabiters for every 1,000 heterosexual (unmarried) cohabiters. San Francisco had the highest value on Gay Index 2 and Bakersfield the lowest. The third and fourth data columns of the Appendix show the values of this second index for the 92 metropolitan areas.

The second index for lesbians had an average value among the 92 metropolitan areas of almost 23. San Francisco had the highest value with a score of over 82; for every 1,000 heterosexual cohabiters in San Francisco, there were almost 82 lesbian cohabiters. Birmingham had the lowest score.

Our third gay (lesbian) partnering index employs the same numerator as the first two sets of indexes, but its denominator is less restrictive than that used in the first set: It includes all males (females) age 18 or over, not just those who are unmarried. Unlike Gay (Lesbian) Index 1, this third index does not exclude married people from the denominator. Gay (Lesbian) Index 3 is a rate and is defined as follows:

Gay (Lesbian) Index 3 = [# of gay (lesbian) partners/total # of males (females) age 18+] 1,000

Since the denominator of Gay (Lesbian) Index 3 is so much larger than the denominator of Gay (Lesbian) Index 1 (because it also includes married persons), we have multiplied it by a constant of 100,000, instead of 1,000 as we did in the first set. The fifth and sixth data columns of the Appendix show the values of this third index for the 92 metropolitan areas.

Gay Index 3 had an average score across the 92 metropolitan areas of 112.5 (see Table 1). For every 100,000 males age 18 and up there was an average of almost 113 gays in partnering relationships. San Francisco had the highest score of all the metropolitan areas, 898.9 gay partners per 100,000 males in the population; almost 0.9% of all males age 18 and older in San Francisco in 1990 were in gay partnering relationships. Bakersfield had the lowest score on this index.

For lesbians, the mean value on this third index was 89. Across the 92 metropolitan areas, there was an average of 89 lesbian partners per 100,000 females age 18 and older. Boston had the highest value on Lesbian Index 3 (711) and Charleston, South Carolina, had the lowest value.

A fourth type of index is not a rate but a relative measure of the degree to which gay men and lesbians are overrepresented or underrepresented in a given metropolitan area compared to all metropolitan areas with regard to the percentage of unmarried partners who are gay (or lesbian). It is analogous to the index developed by Black et al. (2000, 2002). To calculate this set of indexes, we first determined for each metropolitan area the total number of unmarried partners, both homosexual and heterosexual. For each metropolitan area we then calculated the percentages of gay men and lesbians; the remaining partners were heterosexual unmarried partners. For example, in San Francisco, the percentage distribution of all unmarried partners in 1990 was gays, 18.6%; lesbians, 6.2%; and heterosexuals, 75.2%.

We then determined the percentage distribution of unmarried partners for the sum of all 92 metropolitan areas. The results were gays, 3.7%; lesbians, 2.4%; and heterosexuals, 93.9%.

For each metropolitan area we then subtracted the percentage of gay unmarried partners in all 92 metropolitan areas from the corresponding percentage in the given metropolitan area; this gave us a relative value of the percentage amount of the metropolitan area's overrepresentation or underrepresentation of gay partners compared to the percentage of gay partners in all the metropolitan areas together. These calculations produced index values of the percentage distribution of gay (or lesbian) partners that were adjusted on the basis of the percentage distribution of gay (or lesbian) partners in all the metropolitan areas.

We refer to this index as Gay (Lesbian) Index 4, and it is defined as follows:

Gay (Lesbian) Index 4 = % gay (lesbian) partners of all unmarried partners in the metropolitan area - % gay (lesbian) partners of all unmarried partners in all 92 metropolitan tan areas combined

The value for San Francisco on Gay Index 4 was 14.9%. Thus, San Francisco's percentage of gay partners out of all unmarried partners is 14.9% greater than the percentage of gay partners out of all unmarried partners in all 92 areas combined. San Francisco's value of 14.9% on the Gay Index 4 was calculated by subtracting the value of the percentage of gay partners in all 92 metropolitan areas, 3.7%, from San Francisco's corresponding value of 18.6%, equaling 14.9%. The last two data columns of the Appendix show the values on this index for all 92 metropolitan areas.

The bottom two rows of Table 1 display descriptive information for gays and lesbians for this fourth set of indexes. Across the 92 metropolitan areas, the average percentage of gay partners to all unmarried partners was 0.8 less than the corresponding percentage for all 92 areas. As just discussed, San Francisco had the highest Gay Index 4 value, and Bakersfield the lowest. Bakersfield's percentage of gay partners out of all unmarried partners was 3.1% less than the percentage of gay partners out of all unmarried partners in all 92 metropolitan areas.

The mean score for lesbians on the fourth index was -0.3%. San Francisco had the highest score and Birmingham the lowest. Compared to the percentage distribution of lesbian partners out of all unmarried partners in all 92 metropolitan areas, lesbians were overrepresented in San Francisco by 3.8% and underrepresented in Birmingham by 1.8%.

Relationships Among and Between the Indexes

We produced four sets of index values, each representing in different statistical and conceptual ways the degree of prevalence of gay and lesbian partners in the 92 largest metropolitan areas of the U.S. in 1990. We next examined the extent to which these indexes were associated with one another. We first considered the relationships of the four gay indexes, and then the relationships of the four lesbian indexes. Finally, we considered the degree to which the gay indexes were related to the lesbian indexes.

Table 2 is a matrix of zero-order correlations showing the relationships between the various indexes. The four gay indexes were highly correlated with one another. Of the six zero-order correlations involving the relationships between each pair of gay indexes, the highest was .995 between Gay Index 2 and Gay Index 4; the lowest was .692 between Gay Index 2 and Gay Index 3. The four lesbian indexes were also positively and highly correlated with each other, although the correlations were not quite as high as those for the four gay indexes. The correlations for the six pairs of lesbian indexes ranged from a low of .574 (Lesbian Index 1 with Lesbian Index 3) to a high of .992 (Lesbian Index 2 with Lesbian Index 4). For both the gay and lesbian index sets, the correlations between the second and fourth indexes were over .99, indicating that these two indexes were sharing the same variation.

The relationships between the gay and lesbian versions of each of the four indexes were also strong and positive. We found a correlation of .709 between Gay Index 1 and Lesbian Index 1 ; a correlation of .702 between Gay Index 2 and Lesbian Index 2; a correlation of .817 between Gay Index 3 and Lesbian Index 3; and a correlation of .623 between Gay Index 4 and Lesbian Index 4.

From a methodological perspective, the first set of indexes is better than the second and third sets because the denominator is statistically more defensible. On the other hand, the fourth set of indexes introduces an additional and very interesting consideration, that is, a relative dimension. The fourth set of indexes represents an adjustment indicating the greater or lesser prevalence of gays (or lesbians) in a given area compared to the prevalence of gays (or lesbians) in all the metropolitan areas combined. For both gays and lesbians, we obtained high correlations between the first and fourth indexes. Gay Index 1 correlated at .943 with Gay Index 4, and Lesbian Index 1 correlated at .871 with Lesbian Index 4. The variation for the first and fourth indexes for gays and lesbians was nearly the same, as was the variation for the second and fourth indexes. We will therefore use the first set of indexes as the preferred set.

Considering the data values in the appendix table, the large majority of metropolitan areas (68 of the 92) had Gay Index 1 values larger than their Lesbian Index 1 values. Only 17 metropolitan areas had lesbian index values larger than their gay index values. Eight metropolitan areas (Bakersfield, Cleveland, Lowell, Louisville, Oklahoma City, Scranton, Syracuse, and Tulsa) had Gay Index 1 values that were essentially the same as their Lesbian Index 1 values. With a few exceptions, notably San Francisco, the Gay Index 1 values were not remarkably different from the corresponding Lesbian Index 1 values. Metropolitan areas with high rates of gay partnering had high rates of lesbian partnering, and areas with low gay rates had low lesbian rates.

SOCIAL, ECOLOGICAL, AND POLITICAL CORRELATES OF HOMOSEXUAL PARTNERING

We next turned to the issue of accounting for variation in the rates of gay and lesbian partnering, asking why San Francisco had the highest gay and lesbian partnering rates and why Birmingham had the lowest rates (for another view, see Black et al., 2002). Firey's (1945, 1947) work of several decades ago assisted in addressing the first part of the question. Firey noted that certain areas in Boston, particularly Beacon Hill and the Boston Common, tended to resist economic pressures for redevelopment because they carried symbolic and sentimental value. One could extend Firey's use of "sentiment as an ecological variable" to San Francisco and its high prevalence of homosexuals by noting that over time San Francisco has become a Mecca for homosexuals. Spaces in San Francisco, for instance the Castro Valley, have come to be known as sacred and sentimental sites for homosexuals, thus attracting homosexuals to San Francisco above and beyond the numbers expected on the basis of more rational kinds of considerations. However, the use of sentiment only goes so far. It is not of much utility in explaining why all 92 metropolitan areas in this study varied in the prevalence of homosexuals. Various social, ecological, and political considerations need to be introduced. We first proposed an ecological orientation for the theoretical justification of our hypotheses.

Our investigation of reasons for the variability in rates of gay and lesbian partnering among the metropolitan areas assumed that the prevalence of gay and lesbian partnering in these areas was to a significant degree the result of migration. San Francisco's high prevalence of gay and lesbian partnering is not due to large numbers of gays and lesbians being born in San Francisco, but to large numbers moving to San Francisco. To assist in hypothesis development, we considered sociological human ecology and its specific focus on migration.

From the perspective of human ecology, migration is the major mechanism of social change and adaptability for human populations. Knowledge of migration patterns tells us about how "populations ... maintain themselves in particular areas" (Hawley, 1950, p. 149). The ecological approach asserts that human populations redistribute themselves to attain an equilibrium between their overall size and the life chances available to them. Migration is the principal mechanism for effecting this adjustment because it is a demographic response attempting to preserve or attain the best possible living standard by reestablishing a balance between population size and organization (Poston, 1981, p. 138; Poston & Frisbie, 1998, p. 30).

The theoretical foundation of human ecology is based on the interdependence of the four conceptual rubrics of population, organization, environment, and technology. The interrelationships among and between these dimensions inform our understanding of migration patterns in the following way: All populations adapt to their environments, and these adaptations vary among populations according to their social and sustenance organization, their technology, and the size, composition, and distribution of their population. The environment is comprised of both social and physical factors and sets constraints on the population and the form and characteristics of its organization. The technology at the population's disposal sets the boundaries for the form and type of environmental adaptation the population assumes. Human ecology posits that, of the three demographic processes, migration is the most efficient agent for returning the human ecosystem to a state of equilibrium or balance between its size and organization (Poston & Frisbie, 1998).

A hypothesis typically investigated in ecological studies of migration (e.g., Frisbie & Poston, 1978; London, 1986; Poston & Frisbie, 1998; Saenz & Colberg, 1988; Sly, 1972, among many others) is that that variability among human groups in their patterns of migration is a function of differences in their patterns of sustenance organization, technology, environment, and population.

A metropolitan area such as San Francisco has high rates of gay and lesbian partnering because of ecological factors that draw gay and lesbian migrants there, but also because of ecological considerations that draw migrants in general, not only homosexual migrants. An ecological factor that should tend to pull migrants in general to a metropolitan area is its level of sustenance-producing activities. Areas with low levels of unemployment should have higher migration rates than areas with high unemployment. It makes sense, therefore, to hypothesize that gays and lesbians as well as heterosexuals will be drawn to areas with an abundance of jobs, and thus low levels of unemployment (Poston & Mao, 1996). We obtained data on rates of unemployment in 1989 for each metropolitan area (U.S. Bureau of the Census, 1988). We hypothesized that the greater the levels of unemployment, the lower the concentrations of gay and lesbian partners.

In sociological human ecology, the environment is defined as "whatever is external to and potentially or actually influential on the phenomenon under investigation" (Hawley, 1968, p. 330). According to this definition, the environment includes not only the biotic or physical characteristics of an area, such as climate, but also the nonphysical "influences that emanate from other organized populations in the same and in other areas" (Hawley, 1981, p. 9). A characteristic of the physical environment, namely climate, is hypothesized to draw or repel migrants, irrespective of whether they are homosexual or heterosexual. Empirical research in the social sciences using climate as an independent variable often includes temperature as a key consideration of climate and sometimes as the only consideration (Graves, 1980; Karp & Kelly, 1971; Poston & Mao, 1996, 1998). A temperature index involves the measurement of average daily temperature during a cold month such as January or a warm month such as July. These measures are closely related (Poston & Musgrave, 1999). We obtained data on average July temperature, which measures the mean average daily July temperature for each metropolitan area, from 1951 to 1970 (U.S. Bureau of the Census, 1972). If physical climate plays a role in the settlement patterns of gay and lesbian partners, we would expect a negative association between July temperature and gay and lesbian concentrations because most individuals do not wish to live in areas which are extremely hot, preferring mild climates.

There are also environmental factors of metropolitan areas that may be hypothesized to specifically draw gays and lesbians, and not necessarily heterosexuals. For instance, Black et al. (2000, 2002) have noted the importance of the metropolitan area's social attitudes and political and religious orientation as factors that should be related to the prevalence of gays and lesbians, and O'Reilly and Webster (1998) wrote that the social and political characteristics of communities should be associated with levels of gay and lesbian concentration. Accordingly, we hypothesized that the more Republican the voting patterns in an area and the more conservative the religious attitudes (as measured by the prevalence of Southern Baptists), the lower the concentrations of gay and lesbian partners.

Why should the prevalence of Republicans and Southern Baptists be negatively associated with the prevalence of gays and lesbians? The Republican Party has long been identified, rightly or wrongly, as having an anti-homosexual orientation. Although there is a vocal homosexual group in the Republican Party (the Log Cabin Republicans), its influence on the party is thought to be minimal (Green, Guth, Kellstedt, & Smitdt, 1995; Guth, 1995; O'Reilly & Webster, 1998, p. 501). Also, the Southern Baptist conference repeatedly passes resolutions that consider homosexuality an "abomination in the eyes of God" (Steinfels, 1988, p. 6).

We gathered data for each metropolitan area on the percentage of votes cast in the 1988 election for George H. Bush (Scammon & McGillivray, 1989). We also obtained data for each area on the number of Southern Baptist members per 1,000 population in 1989 (Green, Bradley, Bradley, & McNeil, 1992).

The literature we have read does not address empirically whether there is a relationship between the presence of sodomy laws and the concentrations of gays and lesbians. However, the presence of sodomy laws is another example of an ecological feature of the nonphysical environment that may be hypothesized to be associated with the migration of gays and lesbians. Prior to 1990, numerous states enforced sodomy laws specifically against gay men (e.g., in the state of Georgia, Bowers v. Hardwick, 1986). In 1990, 18 states and Puerto Rico had sodomy laws (Knopp, 1990). Thirteen states and Puerto Rico had sodomy laws applying to both heterosexuals and homosexuals (Alabama, Arizona, Florida, Idaho, Louisiana, Massachusetts, Michigan, Minnesota, Mississippi, North Carolina, Puerto Rico, South Carolina, Utah, and Virginia), while 5 states had sodomy laws directed only against homosexuals (Arkansas, Kansas, Missouri, Oklahoma, and Texas). The potential effect of sodomy laws, however, may not be as applicable in analyses of levels of homosexuality post-2003, given the 6-3 ruling by the U.S. Supreme Court on June 26, 2003, overturning the law in Texas that prohibited consenting adults in their own residences from participating in homosexual sex (Welch, 2003).

We thus constructed two dummy variables to measure the presence of sodomy laws: Sodomy-1 was scored 1 if the metropolitan area was in a state with a sodomy law directed against homosexuals and heterosexuals, and 0 if not. Sodomy-2 was scored 1 if the metropolitan area was in a state with a sodomy law directed only against homosexuals. We expected that both sodomy variables would be negatively related to the levels of concentration of gay and lesbian partners; that is, we expected the levels of gay and lesbian concentration to be lower in metropolitan areas in states with sodomy laws. Owing to the greater pervasiveness of the first sodomy measure, however, we expected that the effect of sodomy-1 would be stronger than the effect of sodomy-2.

Table 3 presents zero-order correlations for the metropolitan areas between the preferred rates of gay and lesbian partnering (Gay Index 1 and Lesbian Index 1)and the six social, ecological, and political characteristics mentioned above. Because of missing data for one or more of the ecological characteristics, the correlations were calculated for only 83 metropolitan areas.

All 12 correlations were negative, as hypothesized, although their magnitude varied; 7 of the correlations were statistically significant at p < .05. The strongest correlations with the gay and lesbian indexes were for the percentage Republican variable and the July temperature variable. The higher the percentage of Republicans in a metropolitan area, the lower the levels of gay and lesbian concentration; r = -.349 and -.409, respectively. Also, the higher the average July temperature in a metropolitan area the lower were the levels of gay and lesbian concentration; r = -.346 and -.478, respectively. In both instances, the Republican and July temperature variables were more strongly associated with the lesbian index than with the gay index.

The percentage of Southern Baptists in an area was also negatively and significantly associated with levels of gay and lesbian concentration. Of the four correlations involving the sodomy measures, only one was statistically significant. Metropolitan areas in states with sodomy laws directed against both homosexuals and heterosexuals had lower levels of lesbian concentration than metropolitan areas in states without such laws. There were no significant associations between sodomy measures and levels of gay concentration.

Finally, we performed regression analyses of the gay and lesbian partnering indexes. We initially used the six independent variables shown in Table 3 as independent variables. However, when all six variables were entered in the same regression equation, the statistical tolerances of the sodomy, July temperature, and percentage Republican variables were unacceptable, ranging in value from .41 to .55. But when we dropped from the regressions the sodomy measure pertaining to homosexuals and heterosexuals (sodomy-l), which not coincidentally is the variable with the lowest tolerance, the tolerances of the remaining five independent variables were all above .6, with a mean tolerance of .73.

Table 4 displays the results of two ordinary least squares regressions, one each for the gay and lesbian rates. The results in both equations were similar. The signs of all the independent variables, except that dealing with sodomy, were negative as expected. The metric coefficients for the Republican variable, the unemployment rate, and mean July temperature were all statistically significant, while those for the Southern Baptist and sodomy variables were not. The five independent variables accounted for more than 19% of the variance in the gay rates and for more than 29% of the variance in the lesbian rates.

Of the five independent variables used in the regression equations, one might ask which were the most influential in predicting levels of gay and lesbian concentration. In Table 4 we show for each independent variable its corresponding standardized coefficient, enclosed in parentheses and directly below its metric coefficient. In the gay equation, the variables gauging July temperature, percentage Republican, and unemployment had the greatest effects on levels of gay partner concentration. For every 1 standard deviation increase in average July temperature, for instance, there was on average about a 0.29 standard deviation decrease in the level of gay partner concentration, controlling for the effects of the other four variables. The Republican and unemployment variables had just slightly weaker standardized effects on gay partner concentration.

In the lesbian equation, the same three independent variables had the largest standardized effects, but the effect of July temperature was far stronger than those of the Republican and unemployment variables. For every 1 standard deviation increase in July temperature there was an almost .42 standard deviation decrease in the level of lesbian partner concentration. The standardized effects on the lesbian index of the unemployment rate and the percentage Republican variable were much weaker, at -.224 and -.278, respectively. We turn now to a discussion of these results.

DISCUSSION AND CONCLUSION

In this paper we addressed two issues. The first was the development of four statistical indexes to measure the levels of concentration of gay and lesbian partnering in the major metropolitan areas of the U.S. in 1990. Two are rates, one is a ratio, and one is a measure of the relative difference between the national rate and the metropolitan area. While each index has statistical, descriptive, and demographic merit, we focused on one of the rates (that measuring the number of partnered gays or lesbians per 1,000 unmarried males or females ages 18 and over). Statistically and demographically, this index has the most defensible denominator; that is, it is based on the statistically "best" population from which partnered lesbians and gays are drawn.

We then described the settlement patterns of gay and lesbian partners. Results from all four indexes indicated that San Francisco had the highest concentration of gay partners in 1990. For lesbian partners, the largest concentration in 1990 was found in Boston, while Bakersfield consistently had the lowest level. Also, the four gay and lesbian indexes were all highly correlated with each other, suggesting that any of these rates could be used almost interchangeably. Because of the high correlations between the gay indexes and the lesbian indexes, we also concluded that gay and lesbian couples tend to settle in similar metropolitan areas.

Second, we asked about the kinds of ecological characteristics that influence and are related to the geographical locations of gay and lesbian partners. Drawing on sociological human ecology and a somewhat more limited literature review of gay and lesbian settlement patterns, we identified characteristics of metropolitan areas that have been argued to be related to levels of gay and lesbian concentration, such as percentage of the area voting Republican, the unemployment rate, the presence of sodomy laws, the natural climate, and the degree of Southern Baptist membership.

Although the presence of Southern Baptist members has been used by many as an indicator of more conservative and traditional family values, in our analyses the Southern Baptist membership variable seemed to influence gay and lesbian settlement patterns only modestly. In the multivariate context, it was not a statistically significant predictor. Also, variables measuring the presence of two kinds of sodomy laws--one in which the laws apply to both heterosexual and homosexual couples and the other in which the laws apply only to homosexuals--failed in the multivariate context to be significantly related to levels of gay and lesbian concentration.

The variables that were most influential in predicting levels of gay and lesbian concentration were July temperature, percentage voting Republican, and unemployment rate. Extremes of weather were found to be negatively associated with the prevalence of gay and lesbian partnering. We also found that gay and lesbian partners tended not to settle in areas with high unemployment. Finally, our reported negative relationship of the gay and lesbian indexes with percentage voting Republican suggests that gay and lesbian partners seeking to reside in large metropolitan areas may well avoid those areas with more traditional and conservative political values. Such information about the political climate of possible settlement areas could possibly be spread through networks, the gay press, or word of mouth (Bell, 1991; Valentine, 1993).

With increases in societal awareness of lesbians and gays, we expect that the kinds of social and political characteristics examined here will continue to have an influence on gay and lesbian settlement patterns. Lesbian and gay couples can easily access the Internet and determine if an area is relatively open to gays and lesbians. Internet websites commonly list cities that are lesbian and gay friendly and that have domestic partnership benefits and gay and lesbian establishments.

Our analysis also indicates that more research is needed to determine the quality of census data on gays and lesbians and the methodology that best captures their levels of geographic concentration. For instance, data from the 2000 census indicate an increase from the 1990 census in the reporting of same-sex couples. Indeed, almost all counties in the United States reported at least one same-sex couple living within its boundaries (Smith & Gates, 2001, pp. 1-2). Texas, for example, had a percentage change in reporting from 1990 to 2000 of 445%. But even with increased reporting, Smith and Gates (2001) have argued that gays and lesbians remain underenumerated, with an estimated 30% to 40% of same-gender couples reporting their relationship on the census schedule.

One way to study this issue of data quality would be to carry out in-depth qualitative interviews of samples of gay and lesbian households, asking them how and why they answered the question on the 2000 census pertaining to relationship to the householder and used or did not use the unmarried partner response. One of the more important monographs published in the social sciences in the 1990s was Ethnic Options by Waters (1990). Waters interviewed individuals and asked them why they answered as they did the ancestry question on the 1980 census. Her research told us about the dynamics of answering the ancestry question and the pros and cons of using these data for estimates of the numbers of persons of different ancestries (Irish, French, etc.) residing in the U.S. Similar research now needs to be undertaken among gay and lesbian couples. The results of such research could tell us a great deal about the dynamics among gay and lesbian couples in using the unmarried partner response to the relationship to the householder census question and the pros and cons of using these data for estimates of the numbers of gays and lesbians in the U.S.

We expect that a qualitative analysis may provide as good an understanding as any of the fluidity of sexuality. Census based estimates of the numbers of same-sex couples can only tell us about their levels of concentration and the variation in their settlement patterns. Census-based data will tell us little about the fluidity of diverse sexualities and the networks or organizations that inform self-identified gays and lesbians. Future research related to how and why nonheterosexual couples fill out government forms should provide significant insights about gay and lesbian communities and the boundary formation within and among them.

APPENDIX
FOUR INDEXES OF GAY PARTNERING AND LESBIAN PARTNERING: 92 METROPOLITAN
AREAS OF THE U.S., 1990

Obs. Name of MSA/PMSA                                      Gay
                                                          Index 1

1   Akron, OH PMSA                                  AK      5.2
2   Albany--Schenectady--Troy, NY MSA               AS      9.5
3   Allentown--Bethlehem--Easton, PA--NJ MSA        AB     10.5
4   Anaheim--Santa Ana, CA PMSA                     AN     18.4
5   Atlanta, GA MSA                                 AT     21.9
6   Austin, TX MSA                                  AU     25.2
7   Bakersfield, CA MSA                             BK      4.3
8   Baltimore, MD MSA                               BL     12.3
9   Baton Rouge, LA MSA                             BR      5.4
10  Bergen--Passaic, NJ PMSA                        BP      9.0
11  Birmingham, AL MSA                              BI      0.7
12  Boston, MA PMSA                                 BO     15.9
13  Buffalo, NY PMSA                                BU      5.1
14  Charleston, SC MSA                              CA      7.1
15  Charlotte--Gastonia--Rock Hill, NC--SC MSA      CG      9.9
16  Chicago, IL PMSA                                CH     15.5
17  Cincinnati, OH--KY--IN PMSA                     CI     10.3
18  Cleveland, OH PMSA                              CL      7.6
19  Columbus, OH MSA                                CO     13.7
20  Dallas, TX PMSA                                 DA     17.9
21  Dayton--Springfield, OH MSA                     DS      9.5
22  Denver, CO PMSA                                 DN     14.5
23  Detroit, MI PMSA                                DT      9.2
24  El Paso, TX MSA                                 EP      8.2
25  Fort Lauderdale--Hollywood
    --Pompano Beach, FL PMSA                        FH     27.2
26  Fort Worth--Arlington, TX PMSA                  FW      6.0
27  Fresno, CA MSA                                  FR     12.8
28  Gary--Hammond, IN PMSA                          GH      5.1
29  Grand Rapids, MI MSA                            GR     10.4
30  Greensboro--Winston-Salem
    --High Point, NC MSA                            GW      6.7
31  Greenville--Spartanburg, SC MSA                 GS      5.8
32  Harrisburg--Lebanon--Carlisle, PA MSA           HL     17.0
33  Hartford, CT PMSA                               HA      9.3
34  Honolulu, HI MSA                                HI     15.2
35  Houston, TX PMSA                                HO     15.5
36  Indianapolis, IN MSA                            IN     12.8
37  Jacksonville, FL MSA                            JA     10.5
38  Jersey City, NJ PMSA                            JC     28.9
39  Kansas City, MO--KS MSA                         KC     12.7
40  Knoxville, TN MSA                               KN      5.6
41  Lake County, IL PMSA                            LC      5.9
42  Las Vegas, NV MSA                               LV      6.8
43  Lowell, MA--NH PMSA                             LW      5.5
44  Little Rock--North Little Rock, AR MSA          LR      2.0
45  Los Angeles--Long Beach, CA PMSA                LA     28.7
46  Louisville, KY--IN MSA                          LU      5.9
47  Memphis, TN--AR--MS MSA                         ME      9.1
48  Miami--Hialeah, FL PMSA                         MA     12.4
49  Middlesex--Somerset--Hunterdon, NJ PMSA         MS      9.2
50  Milwaukee, WI PMSA                              MI      9.0
51  Minneapolis--St. Paul, MN--WI MSA               MN     17.0
52  Monmouth--Ocean, NJ PMSA                        MO      6.2
53  Nashville, TN MSA                               NA     10.8
54  Nassau--Suffolk, NY PMSA                        NS      6.3
55  New Haven--Meriden, CT MSA                      NM      6.3
56  New Orleans, LA MSA                             NO     11.1
57  New York, NY PMSA                               NY     21.6
58  Newark, NJ PMSA                                 NE      7.6
59  Norfolk--Virginia Beach
    --Newport News, VA MSA                          NV      8.8
60  Oakland, CA PMSA                                OA     27.7
61  Oklahoma City, OK MSA                           OK      9.6
62  Omaha, NE--IA MSA                               OM     10.7
63  Orlando, FL MSA                                 OR     17.8
64  Oxnard--Ventura, CA PMSA                        OV      8.7
65  Philadelphia, PA--NJ PMSA                       PA     10.1
66  Phoenix, AZ MSA                                 PH      2.5
67  Pittsburgh, PA PMSA                             PI      6.4
68  Portland, OR PMSA                               PO     17.8
69  Providence, RI PMSA                             PR      7.8
70  Raleigh--Durham, NC MSA                         RD     11.3
71  Richmond--Petersburg, VA MSA                    RP     11.7
72  Riverside--San Bernardino, CA PMSA              RS     13.9
73  Rochester, NY MSA                               RO     15.7
74  Sacramento, CA MSA                              SC     18.4
75  St. Louis, MO--IL MSA                           ST      7.3
76  Salt Lake City--Ogden, UT MSA                   SL      8.7
77  San Antonio, TX MSA                             SA     12.2
78  San Diego, CA MSA                               SD     26.0
79  San Francisco, CA PMSA                          SF    101.9
80  San Jose, CA PMSA                               Si     20.3
81  Santa Rosa--Petaluma, CA PMSA                   SR     23.5
82  Scranton--Wilkes--Barre, PA MSA                 SC      3.0
83  Seattle, WA PMSA                                SE     28.6
84  Springfield, MA MSA                             SP      8.1
85  Syracuse, NY MSA                                SY      8.8
86  Tacoma, WA PMSA                                 TA      8.2
87  Tampa--St. Petersburg--Clearwater, FL MSA       TS     12.3
88  Toledo, OH MSA                                  TO      4.0
89  Tucson, AZ MSA                                  TU      2.8
90  Tulsa, OK MSA                                   TL     11.6
91  Washington, DC--MD--VA MSA                      WA     30.9
92  West Palm Beach--Boca Rato
    --Delray Beach, FL MSA                          PB     22.0

Obs. Name of MSA/PMSA                              Lesbian     Gay
                                                   Index 1   Index 2

1   Akron, OH PMSA                                  10.4      12.4
2   Albany--Schenectady--Troy, NY MSA               14.6      19.6
3   Allentown--Bethlehem--Easton, PA--NJ MSA         6.6      20.9
4   Anaheim--Santa Ana, CA PMSA                      7.3      41.4
5   Atlanta, GA MSA                                 11.6      59.4
6   Austin, TX MSA                                  18.7      56.0
7   Bakersfield, CA MSA                              4.3       6.3
8   Baltimore, MD MSA                               13.1      24.1
9   Baton Rouge, LA MSA                              4.0      16.5
10  Bergen--Passaic, NJ PMSA                         7.6      29.2
11  Birmingham, AL MSA                               0.6      19.7
12  Boston, MA PMSA                                 14.6      51.5
13  Buffalo, NY PMSA                                 3.6      12.5
14  Charleston, SC MSA                               2.8      19.2
15  Charlotte--Gastonia--Rock Hill, NC--SC MSA       7.1      28.1
16  Chicago, IL PMSA                                 8.6      41.1
17  Cincinnati, OH--KY--IN PMSA                      4.9      26.1
18  Cleveland, OH PMSA                               7.6      22.9
19  Columbus, OH MSA                                 9.6      31.8
20  Dallas, TX PMSA                                  7.0      44.5
21  Dayton--Springfield, OH MSA                      8.1      25.9
22  Denver, CO PMSA                                 13.0      26.7
23  Detroit, MI PMSA                                 6.9      22.9
24  El Paso, TX MSA                                  4.7      27.8
25  Fort Lauderdale--Hollywood
    --Pompano Beach, FL PMSA                         8.3      44.5
26  Fort Worth--Arlington, TX PMSA                   5.5      15.1
27  Fresno, CA MSA                                   8.9      23.8
28  Gary--Hammond, IN PMSA                           3.8      13.6
29  Grand Rapids, MI MSA                             7.8      22.6
30  Greensboro--Winston-Salem
    --High Point, NC MSA                             2.7      20.3
31  Greenville--Spartanburg, SC MSA                  4.6      15.3
32  Harrisburg--Lebanon--Carlisle, PA MSA            6.8      31.4
33  Hartford, CT PMSA                                8.6      27.9
34  Honolulu, HI MSA                                 5.8      36.2
35  Houston, TX PMSA                                 8.6      37.7
36  Indianapolis, IN MSA                            12.2      23.3
37  Jacksonville, FL MSA                             7.6      23.6
38  Jersey City, NJ PMSA                            11.2      74.7
39  Kansas City, MO--KS MSA                          9.0      28.3
40  Knoxville, TN MSA                                6.7      17.8
41  Lake County, IL PMSA                            10.3      15.3
42  Las Vegas, NV MSA                                9.3       9.9
43  Lowell, MA--NH PMSA                              5.5      15.7
44  Little Rock--North Little Rock, AR MSA           0.7      38.4
45  Los Angeles--Long Beach, CA PMSA                11.4      63.7
46  Louisville, KY--IN MSA                           5.9      15.7
47  Memphis, TN--AR--MS MSA                          3.0      28.1
48  Miami--Hialeah, FL PMSA                          9.6      29.0
49  Middlesex--Somerset--Hunterdon, NJ PMSA          9.7      24.8
50  Milwaukee, WI PMSA                               7.1      19.0
51  Minneapolis--St. Paul, MN--WI MSA               21.3      32.7
52  Monmouth--Ocean, NJ PMSA                         7.5      15.6
53  Nashville, TN MSA                               14.4      27.4
54  Nassau--Suffolk, NY PMSA                         5.3      24.6
55  New Haven--Meriden, CT MSA                       9.4      26.0
56  New Orleans, LA MSA                             10.1      38.9
57  New York, NY PMSA                               11.9      65.6
58  Newark, NJ PMSA                                  6.4      23.1
59  Norfolk--Virginia Beach
    --Newport News, VA MSA                           6.4      23.4
60  Oakland, CA PMSA                                25.8      53.5
61  Oklahoma City, OK MSA                            9.6      27.9
62  Omaha, NE--IA MSA                                8.9      22.6
63  Orlando, FL MSA                                 10.0      31.8
64  Oxnard--Ventura, CA PMSA                        15.4      17.7
65  Philadelphia, PA--NJ PMSA                        9.0      25.6
66  Phoenix, AZ MSA                                  2.1      26.4
67  Pittsburgh, PA PMSA                              3.5      24.7
68  Portland, OR PMSA                               23.7      27.8
69  Providence, RI PMSA                              3.1      31.7
70  Raleigh--Durham, NC MSA                         10.5      30.7
71  Richmond--Petersburg, VA MSA                     5.9      34.6
72  Riverside--San Bernardino, CA PMSA              11.2      23.2
73  Rochester, NY MSA                                9.9      29.2
74  Sacramento, CA MSA                              23.3      31.9
75  St. Louis, MO--IL MSA                            7.0      20.2
76  Salt Lake City--Ogden, UT MSA                    4.3      28.9
77  San Antonio, TX MSA                              8.6      38.3
78  San Diego, CA MSA                               12.6      50.5
79  San Francisco, CA PMSA                          34.2     246.9
80  San Jose, CA PMSA                               17.3      39.7
81  Santa Rosa--Petaluma, CA PMSA                   28.4      38.0
82  Scranton--Wilkes--Barre, PA MSA                  3.0      13.5
83  Seattle, WA PMSA                                25.5      45.8
84  Springfield, MA MSA                              5.1      23.7
85  Syracuse, NY MSA                                 8.8      18.4
86  Tacoma, WA PMSA                                 11.7      14.2
87  Tampa--St. Petersburg--Clearwater, FL MSA       14.9      22.8
88  Toledo, OH MSA                                   5.0       9.0
89  Tucson, AZ MSA                                   2.6      29.8
90  Tulsa, OK MSA                                   11.6      32.5
91  Washington, DC--MD--VA MSA                      16.4      71.9
92  West Palm Beach--Boca Rato
    --Delray Beach, FL MSA                          10.3      35.9

Obs. Name of MSA/PMSA                              Lesbian     Gay
                                                   Index 2   Index 2

1   Akron, OH PMSA                                   24.7      37.1
2   Albany--Schenectady--Troy, NY MSA                30.0      77.2
3   Allentown--Bethlehem--Easton, PA--NJ MSA         13.1      73.1
4   Anaheim--Santa Ana, CA PMSA                      16.5     136.1
5   Atlanta, GA MSA                                  31.3     181.9
6   Austin, TX MSA                                   41.6     220.6
7   Bakersfield, CA MSA                               6.3      27
8   Baltimore, MD MSA                                25.6      90.6
9   Baton Rouge, LA MSA                              12.4      40.9
10  Bergen--Passaic, NJ PMSA                         24.6      63.6
11  Birmingham, AL MSA                               15.7      30.2
12  Boston, MA PMSA                                  47.4     773
13  Buffalo, NY PMSA                                  8.8      37.4
14  Charleston, SC MSA                                7.7      50.5
15  Charlotte--Gastonia--Rock Hill, NC--SC MSA       20.3      75.2
16  Chicago, IL PMSA                                 22.9     117.8
17  Cincinnati, OH--KY--IN PMSA                      12.5      71.8
18  Cleveland, OH PMSA                               22.9      58.1
19  Columbus, OH MSA                                 22.2     103.3
20  Dallas, TX PMSA                                  17.4     131.3
21  Dayton--Springfield, OH MSA                      21.9      70.3
22  Denver, CO PMSA                                  24.0     108.7
23  Detroit, MI PMSA                                 17.1      72.2
24  El Paso, TX MSA                                  15.9      59.5
25  Fort Lauderdale--Hollywood
    --Pompano Beach, FL PMSA                         13.6     198.9
26  Fort Worth--Arlington, TX PMSA                   13.7      46.6
27  Fresno, CA MSA                                   16.4      92.5
28  Gary--Hammond, IN PMSA                           10.2      36.2
29  Grand Rapids, MI MSA                             17.0      67.9
30  Greensboro--Winston-Salem
    --High Point, NC MSA                              8.1      50.7
31  Greenville--Spartanburg, SC MSA                  12.2      40.2
32  Harrisburg--Lebanon--Carlisle, PA MSA            12.5     109.9
33  Hartford, CT PMSA                                25.6      98
34  Honolulu, HI MSA                                 13.8     107.7
35  Houston, TX PMSA                                 21.0     114.1
36  Indianapolis, IN MSA                             22.2      84.8
37  Jacksonville, FL MSA                             17.2      74.8
38  Jersey City, NJ PMSA                             28.9     251.3
39  Kansas City, MO--KS MSA                          20.1      84.9
40  Knoxville, TN MSA                                21.3      35.9
41  Lake County, IL PMSA                             26.8      35.4
42  Las Vegas, NV MSA                                13.6      50.5
43  Lowell, MA--NH PMSA                              15.7     238.4
44  Little Rock--North Little Rock, AR MSA           12.8      82.2
45  Los Angeles--Long Beach, CA PMSA                 25.4     234.3
46  Louisville, KY--IN MSA                           15.7      46.4
47  Memphis, TN--AR--MS MSA                           9.4      75.4
48  Miami--Hialeah, FL PMSA                          22.4     100.3
49  Middlesex--Somerset--Hunterdon, NJ PMSA          26.3      62.3
50  Milwaukee, WI PMSA                               14.9      66.4
51  Minneapolis--St. Paul, MN--WI MSA                40.8     122.9
52  Monmouth--Ocean, NJ PMSA                         18.7      39.7
53  Nashville, TN MSA                                36.5      73
54  Nassau--Suffolk, NY PMSA                         20.8      40.7
55  New Haven--Meriden, CT MSA                       39.1      84.2
56  New Orleans, LA MSA                              35.3     119.2
57  New York, NY PMSA                                36.2     189.4
58  Newark, NJ PMSA                                  19.4      61.6
59  Norfolk--Virginia Beach
    --Newport News, VA MSA                           16.9      72.5
60  Oakland, CA PMSA                                 49.7     212.2
61  Oklahoma City, OK MSA                            27.9      71.7
62  Omaha, NE--IA MSA                                18.9      71.9
63  Orlando, FL MSA                                  17.8     131.4
64  Oxnard--Ventura, CA PMSA                         31.4      60
65  Philadelphia, PA--NJ PMSA                        22.9      76
66  Phoenix, AZ MSA                                  21.9     103.1
67  Pittsburgh, PA PMSA                              13.5      51.9
68  Portland, OR PMSA                                37.1     115.9
69  Providence, RI PMSA                              12.7     102.7
70  Raleigh--Durham, NC MSA                          28.5     105.1
71  Richmond--Petersburg, VA MSA                     17.3      97.6
72  Riverside--San Bernardino, CA PMSA               18.7      92.4
73  Rochester, NY MSA                                18.5     118
74  Sacramento, CA MSA                               40.4     140.2
75  St. Louis, MO--IL MSA                            19.4     53.2
76  Salt Lake City--Ogden, UT MSA                    14.5     75.4
77  San Antonio, TX MSA                              27.1     97
78  San Diego, CA MSA                                24.6     201.3
79  San Francisco, CA PMSA                           82.8     898.9
80  San Jose, CA PMSA                                33.8     152.5
81  Santa Rosa--Petaluma, CA PMSA                    45.9     159.1
82  Scranton--Wilkes-Barre, PA MSA                   13.5      27.6
83  Seattle, WA PMSA                                 41.0     205.4
84  Springfield, MA MSA                              14.8     113.4
85  Syracuse, NY MSA                                 18.4      71.9
86  Tacoma, WA PMSA                                  20.2      54.5
87  Tampa--St. Petersburg--Clearwater, FL MSA        27.4      85.2
88  Toledo, OH MSA                                   11.2      32.2
89  Tucson, AZ MSA                                   28.2     122.2
90  Tulsa, OK MSA                                    32.5      76.9
91  Washington, DC--MD--VA MSA                       38.1     234.5
92  West Palm Beach--Boca Rato
    --Delray Beach, FL MSA                           16.8     142.3

Obs. Name of MSA/PMSA                               Lesbian     Gay
                                                    Index 3   Index 4

1   Akron, OH PMSA                                   74.2      -2.50
2   Albany--Schenectady--Troy, NY MSA               118.3      -1.83
3   Allentown--Bethlehem--Easton, PA--NJ MSA         45.7      -1.67
4   Anaheim--Santa Ana, CA PMSA                      54.1       0.22
5   Atlanta, GA MSA                                  96.0       1.75
6   Austin, TX MSA                                  163.9       1.41
7   Bakersfield, CA MSA                              27.0      -3.07
8   Baltimore, MD MSA                                96.5      -1.40
9   Baton Rouge, LA MSA                              30.7      -2.09
10  Bergen--Passaic, NJ PMSA                         53.5      -0.92
11  Birmingham, AL MSA                               24.1      -1.79
12  Boston, MA PMSA                                 711.2       1.00
13  Buffalo, NY PMSA                                 26.2      -2.47
14  Charleston, SC MSA                               20.2      -1.82
15  Charlotte--Gastonia--Rock Hill, NC--SC MSA       54.3      -1.01
16  Chicago, IL PMSA                                 65.6       0.17
17  Cincinnati, OH--KY--IN PMSA                      34.4      -1.18
18  Cleveland, OH PMSA                               58.1      -1.50
19  Columbus, OH MSA                                 72.3      -0.68
20  Dallas, TX PMSA                                  51.4       0.49
21  Dayton--Springfield, OH MSA                      59.5      -1.23
22  Denver, CO PMSA                                  97.8      -1.15
23  Detroit, MI PMSA                                 53.9      -1.49
24  El Paso, TX MSA                                  34        -1.03
25  Fort Lauderdale--Hollywood
    --Pompano Beach, FL PMSA                         60.7       0.52
26  Fort Worth--Arlington, TX PMSA                   42.3      -2.23
27  Fresno, CA MSA                                   64        -1.41
28  Gary--Hammond, IN PMSA                           27.1      -2.36
29  Grand Rapids, MI MSA                             51        -1.52
30  Greensboro--Winston-Salem
    --High Point, NC MSA                             20.3      -1.72
31  Greenville--Spartanburg, SC MSA                  32.1      -2.21
32  Harrisburg--Lebanon--Carlisle, PA MSA            43.9      -0.69
33  Hartford, CT PMSA                                89.8      -1.04
34  Honolulu, HI MSA                                 41        -0.24
35  Houston, TX PMSA                                 63.4      -0.13
36  Indianapolis, IN MSA                             80.9      -1.47
37  Jacksonville, FL MSA                             54.4      -1.42
38  Jersey City, NJ PMSA                             97.3       3.08
39  Kansas City, MO--KS MSA                          60.1      -0.99
40  Knoxville, TN MSA                                43.1      -1.98
41  Lake County, IL PMSA                             61.9      -2.22
42  Las Vegas, NV MSA                                69.4      -2.73
43  Lowell, MA--NH PMSA                             238.4      -2.17
44  Little Rock--North Little Rock, AR MSA           27.4      -0.03
45  Los Angeles--Long Beach, CA PMSA                 93.4       2.16
46  Louisville, KY--IN MSA                           46.4      -2.17
47  Memphis, TN--AR--MS MSA                          25.1      -0.98
48  Miami--Hialeah, FL PMSA                          77.5      -0.93
49  Middlesex--Somerset--Hunterdon, NJ PMSA          66        -1.33
50  Milwaukee, WI PMSA                               52.1      -1.86
51  Minneapolis--St. Paul, MN--WI MSA               153.6      -0.65
52  Monmouth--Ocean, NJ PMSA                         47.7      -2.19
53  Nashville, TN MSA                                97.3      -1.12
54  Nassau--Suffolk, NY PMSA                         34.4      -1.34
55  New Haven--Meriden, CT MSA                      126.3      -1.25
56  New Orleans, LA MSA                             108.3      -0.07
57  New York, NY PMSA                               104.6       2.26
58  Newark, NJ PMSA                                  51.7      -1.48
59  Norfolk--Virginia Beach
    --Newport News, VA MSA                           52.3      -1.44
60  Oakland, CA PMSA                                197.3       1.15
61  Oklahoma City, OK MSA                            71.7      -1.05
62  Omaha, NE--IA MSA                                59.9      -1.52
63  Orlando, FL MSA                                  73.4      -0.66
64  Oxnard--Ventura, CA PMSA                        106.6      -2.01
65  Philadelphia, PA--NJ PMSA                        68        -1.25
66  Phoenix, AZ MSA                                  85.6      -1.17
67  Pittsburgh, PA PMSA                              28.3      -1.31
68  Portland, OR PMSA                               154.6      -1.08
69  Providence, RI PMSA                              41.1      -0.65
70  Raleigh--Durham, NC MSA                          97.6      -0.79
71  Richmond--Petersburg, VA MSA                     48.8      -0.40
72  Riverside--San Bernardino, CA PMSA               74.7      -1.47
73  Rochester, NY MSA                                74.7      -0.90
74  Sacramento, CA MSA                              177.5      -0.72
75  St. Louis, MO--IL MSA                            51.1      -1.75
76  Salt Lake City--Ogden, UT MSA                    37.7      -0.92
77  San Antonio, TX MSA                              68.7      -0.10
78  San Diego, CA MSA                                97.9       1.01
79  San Francisco, CA PMSA                          301.5      14.88
80  San Jose, CA PMSA                               130         0.00
81  Santa Rosa--Petaluma, CA PMSA                   192.1      -0.18
82  Scranton--Wilkes--Barre, PA MSA                  27.6      -2.38
83  Seattle, WA PMSA                                183.7       0.52
84  Springfield, MA MSA                              70.9      -1.41
85  Syracuse, NY MSA                                 71.9      -1.92
86  Tacoma, WA PMSA                                  77.8      -2.32
87  Tampa--St. Petersburg--Clearwater, FL MSA       102.6      -1.52
88  Toledo, OH MSA                                   40.3      -2.81
89  Tucson, AZ MSA                                  115.5      -0.87
90  Tulsa, OK MSA                                    76.9      -0.64
91  Washington, DC--MD--VA MSA                      124.2       2.78
92  West Palm Beach--Boca Rato
    --Delray Beach, FL MSA                           66.7      -0.28

Obs. Name of MSA/PMSA                              Lesbian
                                                   Index 4

1   Akron, OH PMSA                                  -0.04
2   Albany--Schenectady--Troy, NY MSA                0.43
3   Allentown--Bethlehem--Easton, PA--NJ MSA        -1.17
4   Anaheim--Santa Ana, CA PMSA                     -0.87
5   Atlanta, GA MSA                                  0.44
6   Austin, TX MSA                                   1.36
7   Bakersfield, CA MSA                             -1.81
8   Baltimore, MD MSA                                0.01
9   Baton Rouge, LA MSA                             -1.23
10  Bergen--Passaic, NJ PMSA                        -0.10
11  Birmingham, AL MSA                              -0.91
12  Boston, MA PMSA                                  1.88
13  Buffalo, NY PMSA                                -1.57
14  Charleston, SC MSA                              -1.68
15  Charlotte--Gastonia--Rock Hill, NC--SC MSA      -0.49
16  Chicago, IL PMSA                                -0.28
17  Cincinnati, OH--KY--IN PMSA                     -1.23
18  Cleveland, OH PMSA                              -0.24
19  Columbus, OH MSA                                -0.32
20  Dallas, TX PMSA                                 -0.79
21  Dayton--Springfield, OH MSA                     -0.34
22  Denver, CO PMSA                                 -0.14
23  Detroit, MI PMSA                                -0.79
24  El Paso, TX MSA                                 -0.91
25  Fort Lauderdale--Hollywood
    --Pompano Beach, FL PMSA                        -1.15
26  Fort Worth--Arlington, TX PMSA                  -1.10
27  Fresno, CA MSA                                  -0.85
28  Gary--Hammond, IN PMSA                          -1.43
29  Grand Rapids, MI MSA                            -0.80
30  Greensboro--Winston-Salem
    --High Point, NC MSA                            -1.64
31  Greenville--Spartanburg, SC MSA                 -1.24
32  Harrisburg--Lebanon--Carlisle, PA MSA           -1.23
33  Hartford, CT PMSA                                0.00
34  Honolulu, HI MSA                                -1.11
35  Houston, TX PMSA                                -0.45
36  Indianapolis, IN MSA                            -0.31
37  Jacksonville, FL MSA                            -0.78
38  Jersey City, NJ PMSA                             0.19
39  Kansas City, MO--KS MSA                         -0.51
40  Knoxville, TN MSA                               -0.37
41  Lake County, IL PMSA                             0.14
42  Las Vegas, NV MSA                               -1.10
43  Lowell, MA--NH PMSA                             -0.91
44  Little Rock--North Little Rock, AR MSA          -1.21
45  Los Angeles--Long Beach, CA PMSA                -0.10
46  Louisville, KY--IN MSA                          -0.91
47  Memphis, TN--AR--MS MSA                         -1.53
48  Miami--Hialeah, FL PMSA                         -0.30
49  Middlesex--Somerset--Hunterdon, NJ PMSA          0.07
50  Milwaukee, WI PMSA                              -0.99
51  Minneapolis--St. Paul, MN--WI MSA                1.37
52  Monmouth--Ocean, NJ PMSA                        -0.62
53  Nashville, TN MSA                                1.00
54  Nassau--Suffolk, NY PMSA                        -0.44
55  New Haven--Meriden, CT MSA                       1.24
56  New Orleans, LA MSA                              0.86
57  New York, NY PMSA                                0.86
58  Newark, NJ PMSA                                 -0.57
59  Norfolk--Virginia Beach
    --Newport News, VA MSA                          -0.80
60  Oakland, CA PMSA                                 2.08
61  Oklahoma City, OK MSA                            0.21
62  Omaha, NE--IA MSA                               -0.62
63  Orlando, FL MSA                                 -0.74
64  Oxnard--Ventura, CA PMSA                         0.56
65  Philadelphia, PA--NJ PMSA                       -0.24
66  Phoenix, AZ MSA                                 -0.34
67  Pittsburgh, PA PMSA                             -1.13
68  Portland, OR PMSA                                1.06
69  Providence, RI PMSA                             -1.21
70  Raleigh--Durham, NC MSA                          0.26
71  Richmond--Petersburg, VA MSA                    -0.78
72  Riverside--San Bernardino, CA PMSA              -0.63
73  Rochester, NY MSA                               -0.66
74  Sacramento, CA MSA                               1.34
75  St. Louis, MO--IL MSA                           -0.56
76  Salt Lake City--Ogden, UT MSA                   -1.04
77  San Antonio, TX MSA                              0.12
78  San Diego, CA MSA                               -0.14
79  San Francisco, CA PMSA                           3.80
80  San Jose, CA PMSA                                0.72
81  Santa Rosa--Petaluma, CA PMSA                    1.81
82  Scranton--Wilkes-Barre, PA MSA                  -1.12
83  Seattle, WA PMSA                                 1.34
84  Springfield, MA MSA                             -1.00
85  Syracuse, NY MSA                                -0.66
86  Tacoma, WA PMSA                                 -0.47
87  Tampa--St. Petersburg--Clearwater, FL MSA        0.18
88  Toledo, OH MSA                                  -1.33
89  Tucson, AZ MSA                                   0.24
90  Tulsa, OK MSA                                    0.62
91  Washington, DC--MD--VA MSA                       1.00
92  West Palm Beach--Boca Rato
    --Delray Beach, FL MSA                          -0.83

Note. Gay (Lesbian) Index 1 is measured per 1,000 unmarried males
(females) age 18+. Gay (Lesbian) Index 2 is measured per
1,000 heterosexual cohabiters. Gay (Lesbian) Index 3 is measured
per 100,000 males (females) age 18+. Gay (Lesbian) Index 4 is the
percentage of gay (lesbian) cohabiters of all cohabiters in the area
that is greater (or less) than the percentage of gay (lesbian)
cohabiters of all cohabiters in all 92 areas combined.

Table 1. Means, Standard Deviations, and Minimum and Maximum Values
for Gay and Lesbian Partnering Indexes: 92 Metropolitan Areas of
the U.S., 1990

                                Minimum Value         Maximum Value
Index          Mean    SD            MSA                   MSA

Gay
Index 1        12.9    11.6    0.7                101.9
                               Birmingham, AL     San Francisco, CA
Lesbian
Index 1         9.6     6.1    0.6                34.2
                               Birmingham, AL     San Francisco, CA
Gay
Index2         31.6    26.5    6.4                247.0
                               Bakersfield, CA    San Francisco, CA
Lesbian
Index2         22.6    11.5    6.0                82.4
                               Bakersfield, CA    San Francisco, CA
Gay
Index3        112.5   121.4    27.0               898.9
                               Bakersfield, CA    San Francisco, CA
Lesbian
Index3         89.0    82.4    20.2               711.2
                               Charleston, SC     Boston, MA
Gay
Index4         -0.8    2.1     -3.1               14.9
                               Bakersfield, CA    San Francisco, CA
Lesbian
Index4         -0.3    1.0     -1.8               3.8
                               Birmingham, AL     San Francisco, CA

Note. Gay (Lesbian) Index 1 is measured per 1,000 unmarried males
(females) age 18+. Gay (Lesbian) Index 2 is measured per
1,000 heterosexual cohabiters. Gay (Lesbian) Index 3 is measured
per 100,000 males (females) age 18+. Gay (Lesbian) Index 4 is the
percentage of gay (lesbian) cohabiters of all cohabiters in the
area that is greater (or less) than the percentage of gay (lesbian)
cohabiters of all cohabiters in all 92 areas combined. MSA
= metropolitan statistical area.

Table 2. Zero-Order Correlations for Gay and Lesbian Partnering
Indexes: 92 Metropolitan Areas of the U.S., 1990

              Gay       Gay       Gay       Gay     Lesbian
            Index 1   Index 2   Index 3   Index 4   Index 1

Gay
Index 1       --       .947      .796      .943      .709

Gay
Index 2                 --       .692      .995      .558

Gay
Index 3                           --       .828      .552

Gay
Index 4                                     --       .559

Lesbian
Index 1                                                --

Lesbian
Index 2

Lesbian
Index 3

Lesbian
Index

               Lesbian   Lesbian   Lesbian
               Index 2   Index 3   Index 4

Gay
Index 1         .706      .417      .630

Gay
Index 2         .702      .416      .617

Gay
Index 3         .692      .817      .631

Gay
Index 4         .702      .422      .623

Lesbian
Index 1         .871      .574      .871

Lesbian
Index 2          --       .680      .992

Lesbian
Index 3                    --       .683

Lesbian
Index                                --

Note. All correlations significant at p < .001.

Table 3. Zero-Order Correlations for Two Gay and Lesbian
Partnering Indexes with Social, Ecological, and
Political Correlates: 83 Metropolitan Areas of the
U.S., 1990

Correlates                  Gay Index 1          Lesbian Index 1

% Republican                -.349 (.001)          -.409 (.000)
% Southern Baptist          -.212 (.054)          -.269 (.014)
Unemployment rate           -.154 (.164)          -.084 (.450)
July temperature            -.346 (.001)          -.478 (.000)
Sodomy (homo and hetero)    -.198 (.073)          -.290 (.008)
Sodomy (homo)               -.068 (.542)          -.120 (.281)

Note. p values in parentheses.

Table 4. Metric and Standardized Regression Coefficients
for Gay and Lesbian Partnering Indexes on Social,
Ecological, and Political Correlates: 83
Metropolitan Areas of the U.S., 1990

Correlates             Gay Index I      Lesbian Index 1

% Republican          -0.339 (.018)       -.177 (.12)
                      (-.279)            (-.278)
Southern Baptist       -.005 (.811)       -.002 (.819)
                      (-.028)            (-.026)
Unemployment rate     -1.523 (.010)       -.663 (.023)
                      (-.270)            (-.224)
July temperature       -.582 (.025)       -.440 (.001)
                      (-.290)            (-.418)
Sodomy (homo)          1.104 (.786)        .747 (.708)
                       (.032)             (.042)

Constant              87.248 (.000)      57.736 (.000)
[R.sup.2] (adj.)        .194               .292

Note. p values reported in parentheses to the right of the metric
coefficients. Standardized coefficients reported in parentheses
below the metric coefficients.

REFERENCES

Armas, G. C. (2001, August 22). Census shows 600,000 same-sex homes. Retrieved June 24, 2002, from http://www.lexisnexis.com

Barrett, R. E. (1994). Using the 1990 U.S. Census for research. Thousand Oaks, CA: Sage.

Bell, D. J. (1991). Insignificant others: Lesbian and gay geographies. Area, 23, 323-329.

Black, D., Gates, G., Sanders, S., & Taylor, L. (2000). Demographics of the gay and lesbian population in the United States: Evidence from available systematic data sources. Demography, 37, 139-154.

Black, D., Gates, G., Sanders, S., & Taylor, L. (2002). Why do gay men live in San Francisco? Journal of Urban Economics, 51, 54-76.

Bowers v. Hardwick et al., 478 U.S. 1039 (1986).

Bumpass, L., & Sweet, J. (1989). National estimates of cohabitation. Demography, 26, 615-625.

Casper, L. M. (1992). Community norms and cohabitation: Effects of level and degree of consensus (Working Paper No. 92-9). Princeton, NJ: Office of Population Research.

Castells, M., & Murphy, K. (1982). Cultural identity and urban structure: The spatial organisation of San Francisco's gay community. In N. I. Fainstein & S. S. Fainstein (Eds.), Urban policy under capitalism (pp. 237-259). London: Sage.

Chevan, A. (1996). As cheaply as one: Cohabitation in the older population. Journal of Marriage and the Family, 58, 656-667.

Cohn, D. (2001, August 22). Count of gay couples up 300%: 2000 census ranks D.C., Arlington, Alexandria among top locales. Washington Post. Retrieved June 24, 2002, from http://www.lexisnexis.com

Fields, J. M., & Clark, C. L. (1999). Unbinding the ties: Edit effects of marital status on same gender couples (Population Division Working Paper No. 34). Washington, DC: U.S. Bureau of the Census.

Firey, W. (1945). Sentiment and symbolism as ecological variables. American Sociological Review, 10, 140-148.

Firey, W. (1947). Land use in central Boston. Cambridge, MA: Harvard University Press.

Frisbie, W. P., & Poston, D. L., Jr. (1978). Sustenance differentiation and population redistribution. Social Forces, 57, 42-56.

Gillman, T. J., & Langford, T. (2001, August 22). Gays stand out in Census: Latest population figures put Dallas in the top ranks nationally. Dallas Morning News. Retrieved June 24, 2002, from http://www.lexisnexis.com

Graves, P. E. (1980). Migration and climate. Journal of Regional Science, 20, 227-237.

Green, J. C., Guth, J. L., Kellstedt, L. A., & Smitdt, C. E. (1995, July 5). Evangelical realignment: The political power of the Christian right wins one. The Christian Century, pp. 676-677.

Green, N. M., Jr., Bradley, L., Bradley, M. B., & McNeil, L. (Eds.). (1992). Churches and church membership in the United States 1990 with map: An enumeration by region, state and county based on data reported by 133 church groupings. Atlanta, GA: Glenmary Research Center.

Guth, J. L. (1995). South Carolina: The Christian right wins one. PS: Political Science and Politics, 28, 8-11.

Hawley, A. H. (1950). Human ecology: A theory of community structure. New York: Ronald Press.

Hawley, A. H. (1968). Human ecology. In D. L. Sills (Ed.), International encyclopedia of the social sciences (pp. 323-332). New York: Crowell, Collier, and Macmillan.

Hawley, A. H. (1981). Urban society (Rev. ed.). New York: Ronald Press.

Hinde, A. (1998). Demographic methods. New York: Edward Arnold.

Karp, H. H., & Kelly, K. D. (1971). Toward an ecological analysis of intermetropolitan migration. Chicago: Markham Publishing Company.

Knopp, L. (1990, May). Social consequences of homosexuality. Geographical Magazine, pp. 20-25.

Lauria, M., & Knopp, L. (1985). Toward an analysis of the role of gay communities in the urban renaissance. Urban Geography, 6, 152-169.

London, B. (1986). Ecological and political-economic analyses of migration to a primate city: Bangkok, Thailand, ca. 1970. Urban Affairs Quarterly, 21, 501-526.

O'Reilly, K., & Webster, G. R. (1998). A sociodemographic and partisan analysis of voting in three anti-gay rights referenda in Oregon. Professional Geographer, 50, 498-515.

Phua, V. C., & Kaufman, G. (1999). Using the census to profile same sex cohabitation: A research note. Population Research and Policy Review, 18, 373-386.

Poston, D. L., Jr. (1981). An ecological examination of southern population redistribution, 1970-1975. In D. L. Poston, Jr., & R. H. Weller (Eds.), The population of the south: Structure and change in social demographic context (pp. 137-154). Austin: The University of Texas Press.

Poston, D. L., Jr., & Frisbie, W. P. (1998). Human ecology, sociology, and demography. In M. Micklin & D. L. Poston, Jr. (Eds.), Continuities in sociological human ecology (pp. 27-50). New York: Plenum Press.

Poston, D. L., Jr., Gu, Y., & Walther, C. S. (2003, October). The ecology of homosexuality. Paper presented at the annual meeting of the Southern Demographic Association, Washington, DC.

Poston, D. L., Jr., & Mao, M. X. (1996). An ecological investigation of interstate migration in the United States, 1985-1990. Advances in Human Ecology, 5, 303-342.

Poston, D. L., Jr., & Mao, M. X. (1998). Interprovincial migration in China, 1985-1990. Research in Rural Sociology and Development, 7, 227-250.

Poston, D. L., Jr., & Musgrave, M. (1999, June). The effect of climate on internal migration in the United States and China. Paper presented at the Open Meeting of the Human Dimensions of Global Environmental Change Research Community, Shonan Village Center, Kanagawa Prefecture, Japan.

Saenz, R., & Colberg, E. (1988). Sustenance organization and net migration in small Texas nonmetropolitan communities, 1960-1980. Rural Sociology, 53, 334-345.

Scammon, R. M., & McGillivray, A. V. (1989). American votes 18: A handbook of contemporary American election statistics. Washington, DC: Congressional Quarterly.

Simmons, T., & O'Connell, M. (2003). Married couple and unmarried-partner households: 2000. Census 2000 special reports (CENSR-5). Washington, DC: U.S. Bureau of the Census.

Sly, D. F. (1972). Migration and the ecological complex. American Sociological Review, 37, 615-628.

Smith, D. M., & Gates, G. J. (2001). Gay and lesbian families in the United States: Same sex unmarried partner households: A preliminary analysis of 2000 United States Census data. A Human Rights Campaign Report. Retrieved August 22, 2001, from http://www.hrc.org

Spanier, G. B. (1982, November). Living together in the eighties. American Demographics, pp. 17-20.

Steinfels, P. (1988, June 17). Southern Baptists condemn homosexuality as "depraved." New York Times, p. B6. Retrieved June 24, 2001, from http://www.lexisnexus.org

U.S. Bureau of the Census. (1972). 1972 county and city data book. Washington, DC: U.S. Government Printing Office.

U.S. Bureau of the Census. (1988). 1988 county and city data book. Washington, DC: U.S. Government Printing Office.

U.S. Bureau of the Census. (1992). Census of population and housing, 1990: Public use microdata sample US technical documentation. Washington, DC: U.S. Government Printing Office.

Valentine, G. (1993). (Hetero)sexing space: Lesbian perceptions and experiences of everyday spaces. Environment and Planning D: Society and Space, 11, 395-413.

Waters, M. C. (1990). Ethnic options. Berkeley: University of California Press.

Weightman, B. A. (1981). Commentary: Towards a geography of the gay community. Journal of Cultural Geography, 1, 106-112.

Welch, G. (2003). Supreme Court overturns laws banning homosexual conduct. The Touchstone, 13(3), 13.

Manuscript accepted October 6, 2003

Address correspondence to Dudley L. Poston, Jr., Department of Sociology, Texas A&M University, 4351 TAMU, College Station, TX 77843-4351; e-mail: [e-mail].

Carol S. Walther and Dudley L. Poston, Jr.

Texas A&M University