APPENDIX 2
RESEARCH METHODOLOGY
Here I provide details about the composition of the participants and about the research
methodology to give readers a full picture of how I conducted the study. Some of this
material is presented in the introduction, but I repeat details, such as the demographic
information and the process of collecting data, to give the clearest account of my meth-
odological choices.
THE PARTICIPANTS
Data were collected in face-to-face interviews I conducted with fifty-three people who
were engaged in twenty-six close intersectional friendship dyads and one triad at the
time of the interviews, which occurred between October 2002 and August 2003. The
participants were primarily residents of the San Francisco Bay Area and surrounding
counties, although six of the interviewees lived in Southern California.∞ The interviewees
self-identified their gender as male or female and their sexual orientation as straight, gay,
or lesbian (although one identified as ‘‘queer’’).≤ The total sample included twenty-eight
women (thirteen lesbian, fourteen straight, one queer) and twenty-five men (thirteen
gay, twelve straight). There are more women than men in the study because I was unable
to interview the male halves of the dyads in two cases; also, the triad included in the study
was composed of two women and one man. The participants ranged in age from twenty-
one to sixty-four, with a median age of thirty-two; the racial composition was 59 percent
white, 17 percent Latino, 19 percent Asian, and 4 percent black. Appendix 1 contains a list
of participants with corresponding demographic information and identifies her or his
intersectional friend.
The participants were recruited using a convenience and purposive snowball sam-
pling method, in which one participant refers the researcher to another, beginning with
my contacts in the San Francisco Bay Area lesbian, gay, bisexual, and transgender (lgbt)
communities and expanding through the participants’ social networks. These methods
were ideal for sampling intersectional friends because the targeted population is not
easily located using other methods of data collection. For example, building a random
and representative population sample would have been exceptionally di≈cult, because
the actual population size of intersectional friends is unknown. Some participants were
Previous Page Next Page