CHAPTER THREE: RESEARCH METHODS

Introduction

This chapter explores the methodological founda= tions of the study, describes the design, and details the actual process of conducting the study. It begi= ns with an introduction to the research methods and the theoretical framework = of the research. Conceptual chal= lenges to the study are outlined and the social location of the researcher is explored. Following this, the methods are documented in detail, including a description of the sampling p= lan; a detailed account of the outreach used for recruitment; data collection st= rategies; and the process for analyzing the data.&nb= sp; The chapter then concludes with a short discussion on plans for the study findings.

The Research Plan: A Triangulation of Qualita= tive Methods

This research pro= ject is being approached from the twin disciplines of sociology and public health. With this orientation, I begin to = not only explore the health issues related to a particular population, but as a sociologist I examine how social relations constitute that health outcome.<= span style=3D'mso-spacerun:yes'> From within these disciplines I us= e the feminist paradigm, rejecting the androcentric approach of positivist scient= ific methods, and focusing on gender, race, class, and sex interactions.<= !--[if supportFields]>= ADDIN EN.CITE <EndNote><Cite><Author>Yaiser</Author><Y= ear>2004</Year><RecNum>3</RecNum><record><rec= -number>3</rec-number><ref-type name=3D'Book'>6</ref-type><contributors><authors><a= uthor>Yaiser, Michelle L.</author><author>Hesse-Biber, Sharlene Nagy</author></authors></contributors><titles><t= itle>Feminist perspectives on social research edited by Sharlene Nagy Hesse-Biber, Michel= le L. Yaiser</title></titles><pages>xiv, 434</pages><keywords><keyword>Women Social conditions Research.</keyword><keyword>Women Social conditions.</keyword><keyword>Women's studies Research.</keyword><keyword>Feminism.</keyword></keywo= rds><dates><year>2004</year></dates><pub-loca= tion>New York</pub-location><publisher>Oxford University Press</publi= sher><isbn>0195171748 (cloth alk. paper)0195158113 (pbk. alk. paper)</isbn><call-num>HQ1180 .F455 2004HQ1180</call-num><urls></urls></record>= ;</Cite><Cite><Author>Marshall</Author><Year>= 1999</Year><RecNum>10</RecNum><record><rec-numbe= r>10</rec-number><ref-type name=3D'Book'>6</ref-type><contributors><authors><a= uthor>Marshall, Catherine</author><author>Rossman, Gretchen B.</author></authors></contributors><titles><tit= le>Designing qualitative research</title></titles><pages>xvi, 224</pages><edition>3rd</edition><keywords><keyw= ord>Social sciences Research Methodology.</keyword></keywords><dates>= ;<year>1999</year></dates><pub-location>Thousand Oaks, Calif.</pub-location><publisher>Sage Publications</publisher><isbn>0761913408 (pbk.)0761913= 3940761913408 (pbk.)</isbn><call-num>H62 .M277 1999H62</call-num><urls></urls></record>&l= t;/Cite><Cite><Author>Denzin</Author><Year>1998&= lt;/Year><RecNum>11</RecNum><record><rec-number>= 11</rec-number><ref-type name=3D'Book'>6</ref-type><contributors><authors><a= uthor>Denzin, Norman K.</author><author>Lincoln, Yvonna S.</author><= /authors></contributors><titles><title>Strategies of qualitative inquiry</title></titles><pages>xxii, 346</pages><keywords><keyword>Social sciences Methodology.</keyword><keyword>Social sciences Research Methodology.</keyword></keywords><dates><year>1998&= lt;/year></dates><pub-location>Thousand Oaks, Calif.</pub-location><publisher>Sage Publications</publisher><isbn>0761914358 (pbk. acid-free paper)</isbn><call-num>H61 .S8823 1998H61</call-num= ><urls></urls></record></Cite></EndNote>^162-164 = Despite my strong commitment to quantitative research, I believe addressing the questions that arise out of a review of existing literature on transgender health would not be possible through quantitative analysis. In contrast, qualitative methods o= ffered an opportunity to explore a substantive subject of which little is known, u= sing a naturalistic method that respects the humanity of the participants in the study. Qualitative findings e= volve out of holistic descriptions of natural phenomena. They are naturally interpretive and grounded in the lived experiences of people.^{163, 165, 166} = They explore the how and why of a phenomena, not just the what or= how much of quantitative methods.<= /span>¹⁶⁷ =

According to Denz= in, “no single method ever adequately solves the problem of rival interpretive, causal factors.” (p. 25)¹⁶⁸ = To overcome this weakness, a triangulation of methods should always be built i= nto the research design. Denzin identifies four basic types of triangulation: data, investigator, theory, a= nd methodological.¹⁶⁴ = Janesick later adds a fifth type, interdisciplinary triangulation.¹⁶⁹ Three of these methods have been incorporated into this research design.

Interdisciplinary triangulation resulted from the dual approaches of public health and sociol= ogy mentioned previously. I estab= lished data triangulation by deliberately soliciting participants with different perspectives on transgender health. Participants were divided into two major classes, transgender people= and key-influencers. Transgender = people were asked for information related to their own personal experiences of soc= ial determinants. Then, to frame = this knowledge, I interviewed a second category of people, a group I termed R= 20;key influencers”. A key influencer is someone who has information relevant to social determinants of transgender health either through their professional position, their role a= s a community ally, or their role as a partner of a transgender person. Lastly, I built triangulation of methodology into the study by choosing to use three different data collecti= on formats: life history interviews, focus groups, and short interviews.

Life history inte= rviews were chosen as the foundation of the data collection. With a long history of use in the anthropological and sociological fields, a life history interview provides = the greatest depth of information about a person.¹⁷⁰ = It emphasizes individual meaning, or “how a person copes with society ra= ther than how society copes with the stream of individuals.” (p.121)¹⁶³ = Of particular value to this study, a life history helps “capture cultural patterns and how the patterns are linked to the life of an individual.̶= 1; (p. 121)¹⁶³ I theorized that a series of life history interviews with a diversity of transgender people would provide a basis of information on social determina= nts for these individuals. But wh= ile life histories contribute valuable information about individual experiences= , I also wished to explore community-level knowledge. To accomplish this, the life histo= ry interviews were supplemented by focus groups. As Morgan states, focus groups are= best suited for creating data and insights that are dependent on group interacti= on.¹⁷¹ = This is done as people take different experiences and attempt to make “collec= tive sense” of them. (p.259)¹⁷² Wilkinson argues that focus groups = also shift the balance of power away from the researcher, reducing exploitation = and making them well-suited for soliciting information from underrepresented so= cial groups.¹⁷³ = Both of these traits made the limited use of focus groups a valuable addition to the data collection for this study. The two aforementioned methods were appropriate for the transgender participants. In order to best accommodate the key influencer participants, a third method was used, semi-structured in-depth interviews. Basically a “conversation with a purpose” (p. 149),¹⁷⁴ these interviews allowed me to explore topics related to health in the transgender communities, while allowing the respon= dent to determine the structure of their responses. These interviews were planne= d to last approximately thirty to forty-five minutes. In practice, they ended up ranging= from thirty minutes to three hours.

The three data co= llection methods above formed the methodological basis for my research plan. After considering the amount of da= ta that might emerge from each category, and the need for variation within interviewees, I decided to balance the different data collection methods as outlined in the table below. = In practice I ended up conducting one additional key informant and life history interview, and had a total of 521 single spaced transcript pages of data.

Research Plan=
Method	How many?	n per event	total n
Life histories=	12	1=	12
Focus groups	2=	8=	16
Key informant inter= views	8=	1=	8=
Planned number of study participants	36

Research Outcome
Method	How many?	n per event	total n
Life histories=	13	1=	13
Focus Groups	2=	6 FTM, 10 MTF<= /o:p>	16
Key informant inter= views	9=	1=	9=
Actual number of study participants	38

Theoretical Foundation

In 1983, feminist theorist Nancy Hartsock argued that current theoretical models were insufficient for understanding a women’s experience in the world. She proposed an alternative, stand= point theory. This theory argued th= at it was essential to create a new frame of reference for power relations. Standpoint theory attempts to understand power relations from the eyes of a dominated or marginalized population rather than from the point of view of = the dominant population group. Hartsock lo= oked at power relations from the eyes of women rather than men, showing that “one could begin to see the outline of a very different type of commu= nity if one took the mother/infant relation rather than market exchange as the prototypic human interaction.”^{=
175<=
/span>} = Later work on the theory emphasized the production of “concrete multiplicit= y,” or a call for many different marginalized standpoints to be explored, emphasizing within them the creation of communal knowledge.^{=
176<=
/span>} = In standpoint theory, “All knowledge attempts are socially situated and = some of the objective social locations are better than others as starting points…” (p.44)¹⁶² Beginning to create knowledge from within the dominant paradigm is inherently limiting, but doing so from a marginalized position can provide = more opportunity for critical perspective on social organization.

Standpoint theory= holds that gender, race, class and sexuality are not discrete influences, but rat= her intermix to create mutually constitutive identities.^{=
177,=
178} People create knowledge from their experiential standpoint, sometime= s in direct contradiction to dominant social messages. For example, “For Black wome= n, constructed knowledge of self emerges from the struggle to reject controlli= ng images and integrate knowledge deemed personally important. Usually knowledge essential to Bla= ck women’s survival.” (p.95)^{=
178<=
/span>} = Shannon Sullivan further develops this theory by exploring the notion of one’s body as not interactive with the world, but rather as “transactional.” = In this interpretation, there is “an active and dynamic relationship bet= ween things such that those things are co-constitutive of each other” (p.1= 2).¹⁷⁹ Accordingly, bodies and their environments are always transacting, a= nd the mind cannot be artificially separated from its transactional body. Sullivan argues that the transacti= onal body can and should be used to “free ourselves of the rigidity and stagnation of the self that accompany gender binarism” (p.110), but w= arns that this is inherently a dangerous process: “To challenge the rigid confines of one’s gender is therefore to commit gender treachery̷= 0; it is to risk all the psychological, physical, emotional, financial, and ot= her punishments meted out to gender traitors in our society.”(p.90)¹⁷⁹ = The concept of a body and ones world being co-constitutive dovetails with resea= rch showing how oral narratives are a constructed phenomena. Researchers have sometimes dismiss= ed oral narratives having little validity¹⁸⁰. Recently there is increasing respect for the role of oral narratives within research studies but with an understanding that the stories told are themselves a creation of intention, memory, interpretation, and historical facts^181-184.

The phenomena of = mutually constitutive identities and transactional bodies are both particularly rele= vant to the population being studied here. This project explores social influences on a group of highly marginalized people, transgenders, about whom dominant messages are persistently and extremely negative. In this group, communal knowledge is aggressively built. Whether through sharing the names = of accepting health care providers, or sharing the horror stories of death by street violence, the rapid creation of communal knowledge is a survival ski= ll for these populations. In additio= n, the transgender population experiences a unique relationship between mental self-perception and physical body. As individuals transform their bodies in an effort to achieve congru= ence between mental and physical gender identity, their gender variance is often publicly exposed. This exposu= re impacts personal identity and mediates the level of external gender oppress= ion experienced. Understanding the experience of transgender people from the lens of the dominant social parad= igms is an exercise in futility, because from this perspective there is no prece= dent for understanding the group at all. As is key to standpoint theory, the lived experience of the people m= ust be given primacy. The explora= tion of their constructed knowledge can then highlight deficiencies in the assumptions inherent in dominant paradigms.

Conceptual and Empirical Challenges in Studying Transgender People

The small size of= the transgender population makes it challenging to study this group of people.<= span style=3D'mso-spacerun:yes'> In addition, the often extreme discrimination experienced by transgender people results in many individuals hiding this identification, or trying to pass as non-transgender. Among community members, there is notable concern over those who have “disappeared” or “gone stealth”; that is, transitioned into the opposite sex and then, throu= gh their ability to pass, severed ties with transgender people and submerged themselves in the gender normative world.&= nbsp; Several concerns emerge as a result of this relative isolation of stealth transgender people: how are their stories included in research; how= do social service agencies reach out to them with support services; what effect does it have to isolate oneself from people similar to you? One barrier researchers encounter = when studying transgender people is that potential study participants must acknowledge some level of gender variance.= Study recruitment is therefore very likely to overlook people who ha= ve gone stealth, and as a result, not include their experiences. The design of the present study co= uld not overcome this limitation.

Another barrier is language. As has been discuss= ed, there is no single word that adequately captures the range of gender variant people. While “transgen= der” has been deliberately used here as the most adequate term, it is a heuristic tool that has distinct limitations. One author amusingly refers to it as a “spongy neologism”= ;¹⁸⁵; a direct reference to the dubious nature of gathering disparate groups under one term.= Many people who present as gender variant would never self-identify = as transgender. Some of this gro= up does acknowledge that the word is intended to include them, and will respon= d to it when it is used by others (e.g. responding “yes” to a survey question about being transgender). For others, not only will they not identify with the word transgende= r, they also may not even consider that the term would ever include them. An experience at the beginning of = my fieldwork highlighted this issue.

One of my first interactions with the transgender community in New York was as facilitator for an his= toric meeting of transgender activists from across the state. They had gathered to form a new al= liance to help pass statewide anti-discrimination legislation. The first order of business was to= agree on one or more identifying terms for the name of their alliance. After hours of debate, and a final consensus on using the controversial term ‘transgender,’ some participants expressed surprise that they had reached any agreement at all. This is how strongly the divisions over this term are felt within the community.

As is noted in Da= vid Valentine’s exploration of identity within the transgender communitie= s, the word “transgender” is less likely to be used in some communities of color.⁹ = This was a particular issue as I sought out an African American elder female butch f= or the study. Experience led me = to understand that few if any in this group would ever consider identifying as transgender, but I hypothesized that the gender variant experiences of these individuals would share commonalities with those of other participants. In recruiting the person interview= ed, I had to surmount the barrier of having the word “transgender” in= the study title and provide an explanation as to why she fit the study population. Ultimately the participant did not balk at the study title, and as her story shows her his= tory was profoundly influenced by her gender variance.

The final and most notable barrier in studying transgender people is the great diversity of pe= ople under this single category. Transexuals, gender queers, drag queens, butch women, effeminant mal= es, and cross-dressers of every possible social and cultural backgrounds all cluster under this grouping. The variation in identities directly reflects the variation in perso= nal experiences. Possibly, this i= s much too large a population to easily categorize and plumb for those desired com= mon experiences. Alternately, in = the words of C. Wright Mills, “Perhaps the variety is not as ‘disorderly’ as the mere listing of the small part makes it seem.” (p.131)¹⁸⁶ = In fact, this research takes as an assumption that research on this broad group will find common themes, reflecting a level of common experiences belied by the = diversity of personal identifying labels. This level of diversity begs for closer scrutiny, and future dissertations could easily be done, for example, on Latino male cross-dress= ers alone. But the nature of work= ing in a relatively unexplored area means that a broad spectrum of information abo= ut the transgender people is most valuable right now. To meet this need, this research p= roject deliberately uses a triangulation of qualitative methods to cull stories fr= om a variety of community perspectives. Congruent with this approach, the findings are intended to be illustrative, not generalizable. They are to be used as would the findings any exploratory research project, to help create a basis of knowledge about these populations, and inform more specialized research to follow. This said, there is also no = reason to believe that the stories contained herein are exceptional, and as the sampling and outreach sections describe, no efforts were made whatsoever to= recruit people with exceptional stories into the study.

The Researcher’s Role

Post-positivist f= eminist research acknowledges that research is an interaction between the subjects = and the researcher that creates situated knowledge. A researcher’s past experien= ces, opinions, and enculturation all act as a lens through which the incoming da= ta is filtered. The participant’s reactions to the researcher are influenced by the same factors.^{162, 163} = These principles are all taken as given in this study. As such, I will explicitly explore= my own social location in relationship to this research below.

There has been mu= ch discussion about whether it is necessary, or even beneficial, for a researc= her to study a community of which they are a member.¹⁸⁷ = A level of perceived commonality can be very beneficial for getting hard-to-reach target populations to be accepting and forthright with a researcher. Additionally, lived experience as a community-member may be particularly valuable in studying highly stigmatized populations²⁷. But some argue that insider status impedes the ability to critically reflect on= the issues. As a transgender pers= on, I am an insider to the communities being studied. While this did result in easier ac= cess to community members, it did not mean that participants’ experiences = were reflective of my own, or that we were “matched” in a way that w= ould neutralize interviewer effect^{188, 189}. The differences in socio-economic status, cultural and ethnic identity or enculturation were often large. As Naples has argued, issues of gender, cla= ss, and racial-ethnic relations made the possibility of being a true community insider “mythical” (p. 373).¹⁸⁷ I am a Caucasian, middle-class, well-educated,= non-religious Midwesterner who was socialized as a female. At many points during the study I questioned my ability to be a competent listener to the stories, and wonder= ed how and when my cultural positioning would influence the decisions to highl= ight or ignore certain points? Whi= le this question is difficult to answer, it emphasizes the continued importanc= e of self-reflexivity in this and any other research of these populations. As part of this self-reflexivity a= nd in an effort to clarify my positionality within this research project, I will = now present a brief summary of my social location.

From some of my e= arliest memories, I have had a strong association with the dyke community. My mother came out when I was twel= ve, exposing me to a variety of dykes active in the politics of the late 1970s. The proposed Equal Rig= hts Amendment, civil disobedience at the state capital, women’s only land, and lesbian separatism were all living topics in my teenage home. For a period of that time, I react= ed to my mother’s sexual orientation very poorly, socializing with her frie= nds amicably on one hand, and reverting to fundamentalist religious ideology and active verbal queer-bashing on the other.&= nbsp; When I came out at the age of 16, it was followed by a yearlong stin= t in the mental hospital to reconcile these disparate behaviors. While there, I struggled against t= he negative perceptions of the medical community: I was told I was “demented”, my problem was that I “hated men”, and I earned the official diagnosis of homosexual (twelve years after it had been removed from the DSM). <= /o:p>

I have long assoc= iated early life antagonistic responses to my presented gender ambiguity with circumscribed life chances. A= t the age of 19, I was fired from my job in a sweep that included all other sexual minority or gender variant employees. At the age of 21, I stood si= lently as my manager quipped to other upper managers that it would have been smart= to “drop the bomb” on Washington D.C. during the last major = LGBT march I attended. The contrad= iction between getting praise from managers for my performance while simultaneously enduring jokes about their desire to kill me and everyone I knew was untenable. I grew to understa= nd that I could not work in most all traditional jobs, and retreated socially = and professionally into worlds almost wholly peopled with others who were also LGBT. During this time I stab= ilized my self-perception as a butch dyke, and saw myself as surrounded by lesbians and gay men, with the occasional highly visible and personally intriguing d= rag queen.

In 1992 I was ele= cted to a leadership role for the upcoming lesbian, gay, and bisexual march on Washington. In this capacity, others and I were given several education sessions about this new word “transgenderR= 21; and the previously fractionalized group of communities it represented. The message to me was different th= an the others, it included the addition: “And you are one of us.” While this codicil could have been simply presumptuous, it ultimately was a powerful message for me as I came = to understand its truth. This new paradigm greatly helped to clarify many of my historical experiences. As if being given a new sense to describe the world, I grew to understand how I was different than many of t= he gays and lesbians with whom I was working or socializing. I now had language for my long-felt commonality with drag queens.

While I had a his= tory of experiencing discrimination and threats of violence, it almost always based= on my gender identity, not sexual orientation. Often the most hostile threats were based upon a perception of me as a fag, or confusion over my gender. In the early 1990s, my lover and I= were attacked while participating in a Queer Nation demonstration. Surprisingly, the attack was immed= iately diffused by an undercover police officer and an arrest ensued. The subsequent prosecution highlig= hted the inability of the Chicago District Attorney’s office to successful= ly utilize recently passed hate crimes legislation (this event eventually led = to a reorganization in the office to correct this problem). But while all eyes were on the fir= st anti-lesbian and gay hate crime to be prosecuted under the newly passed legislation, the crime actually hinged on transphobia. As we were attacked directly after= the perpetrator shouted “fudgepacker”. While it took a second to process,= we quickly understood that the perpetrator thought us to be fags, not the fema= le identified butch dykes we were (at the time). This irrevocable mixing of gender presentation and sexual orientation is common in discrimination I have experienced. Separating the s= exual orientation and gender identity paradigms in my mind helped to clarify why = my experience was so different from other dykes, particularly those who could = more easily pass as gender-normative in the streets.

For years to foll= ow, I acknowledged being transgender, but did not often publicly pronounce this fact. Two areas of social ten= sion contributed to this reluctance: my enculturation in the dyke community with= the concomitant tension between many dykes and transgender people; and my perception that the transgender communities validated transexuals above oth= er gender variant people. There = did not seem to be a cultural space for the non-transexual transgender person.<= span style=3D'mso-spacerun:yes'> In late 1999, a series of conversa= tions with the noted trans activist, Leslie Feinberg, made me determine to be more public and active about being transgender.= In particular, I vowed to be more public about being a transgender person who did not conform to one of the two historical dichotomous choices= for sex, but lived somewhere in the gray area in between those two fixed points. At the same time, a n= ew generation of people was promoting the concept of “genderqueer”= , a deliberate positioning that is neither male nor female, but somewhere in-between. Ironically, it wa= s not long before a transexual friend of mine reported that the genderqueers were= the new hot item in San Francisco<= /st1:City>, and transexuals were being disparaged for being too conforming to historical gender stereotypes.

In my new more pu= blic acknowledgement of being transgender, I gave several “Transgender 101” workshops at LGBT health conferences and started to drop the pronouns from my biographical sketches.&nb= sp; I quickly found that being transgender was an asset professionally.<= span style=3D'mso-spacerun:yes'> Many working in the LGBT health communities were interested in including representatives from the less-understood bisexual and transgender communities, and my dual personal categorization as a dyke and transgender person was valuable.

Personally, I und= erstand that my progress along a gender identity continuum is continuing to change,= and it is difficult to see where it will stop.= Early life alliances with the lesbian separatist community create tension with my ongoing masculine identification, which is probably the rea= son I have not found an earlier settling point with my gender identity. My immediate future might well inc= lude hormones or even surgery. Cur= rently I have no body modifications. Like many others, I rarely pass as anything consistently, instead living in that world of constant public confusion about my gender. At my request colleagues refer to = me by the male pronoun, but I currently eschew identification as a man. I once thought this to be a compli= cated position to take but now, as I have made more friends in the transgender communities, I realize this balancing of apparent contradictions is quite common. The close connection between pronouns and gender has been loosened for me, and for many of my friends. If anything, my lived experience has taught me not to be critical of gender identity, but simply accept a person as they present.

My social locatio= n within the transgender communities is one factor that impacts my interpretation of study data; another is my social location within the larger context of our society. As previously mentio= ned, I was encultured as a white middle-class female. My early years were mostly spent o= n the outskirts of Chicago, in a relatively ethnically-homogeneous community. As is the case with many white you= th, my parents actively taught us to “white out” our position within t= he racial and ethnic diversity of society.⁵⁶ = This left me with a personal perception that I was not biased or discriminatory,= but also with a critical lack of awareness of racial and ethnic disparities or = my role in perpetuating them. In= my late teens, I moved into downtown Chicago. As I started to hang with a group = of friends from the predominantly Black and Latino working class dyke bar I frequented, my cultural grounding started to broaden. In the years since I have both wor= ked and socialized in a wider range of racial and ethnic communities, as well as formally studied the mechanisms of discrimination within our society. These experiences have provided a partial counter to the earlier ignorance, but they also taught the depth to which we are each a product of our enculturation. I understand that as a result of my social programming, I carry internal messages and display external ones rel= ated both to my position of relative privilege within our society and to the prevalent discrimination against people by such factors as race, class, or ability. I also have personal experiences that lead me to believe that my current cultural presentation to others is not easily categorizable, as variations in language and styles of communication likely convey some diversity of experience. This has previously had the effect= of easing communication between me and people with disparate cultural backgrounds. It was my hope t= hat this phenomenon might be an asset in this project.

Data Sources: A Study in Finding Hard to Reach Populations

Low-Incidence and Low Profile

Finding hard-to-reach or “hidden” populations has been the focus of inc= reased study over the past few years^{28-30, 190-19=
4}. As can be expected, those populati= ons who less often interact with the health system in predictable, systematic w= ays are less accessible to both the researcher and intervention programs. For the lowest income populations, accessing healthcare often means interacting with governmental representati= ves who gatekeep entitlement programs. Often a group of people must assiduously avoid some governmental representatives to avoid incarceration (e.g. undocumented immigrants, intravenous drug users, prostitutes, etc.). Sometimes avoidance is caused beca= use governmental representatives have a history of violence against a certain c= lass of people, as is the case with young black men. These factors are all issues for t= he transgender communities. Transgender people often carry many stories of discrimination, negle= ct, and abuse at the hands of government officials, and each story potentially gives another reason to avoid social services or health services. Such wariness may discourage trans= gender people from participating in a study such as this one. The relatively small size of the transgender population makes this group even more difficult to reach. Considering all these issues the challenge is obvious, how to reach small groups of people who are proficien= t at hiding?

Approaching the fieldwork for this study I w= as faced with three facts. First= , I was a member of the target populations and therefore had general community knowledge. Second, this insider-status might result in easier access to people. But third, I had moved to the area= relatively recently and thus had almost no detailed knowledge about the local transgen= der communities. Thus, I had very little insider knowledge about how to gain access to people. To overcome this challenge, I pull= ed from a significant personal history of related projects. Previously, I had worked on fieldi= ng large-scale surveys for HIV-positive people in Penns= ylvania, Baltimore, Virginia= , and central Florida. Demographic profiles demonstrated = that these surveys succeeded in accessing a range of hidden populations. I determined the lessons-learned f= rom those studies to be broadly applicable to hidden populations. I reviewed these projects, distilling the lessons learned into an outreach plan.

Findings related to a secondary research ques= tion

As stated in the introduction, one secondary research question was “What is the feasibility and acceptability of t= he proposed methodology in relation to these target populations?” The information in this section of= the methods section delineates both these methods and the positive community response. As a result, this s= ection will be covered with great detail. With this detail I hope to augment the available information about h= ow to gain access to hidden populations.

The sampling plan

As in any qualitative study, sampling presen= ted a specific challenge. Without t= he power of random sampling, the data cannot be considered representative. In the words of Howard Becker, = 220;We don’t want, simple-mindedly, to assume that some feature contained in= our example is just ‘naturally’ there in every class member and thus does not require explanation.” (p.71)¹⁹⁵ This challenge was made more diffi= cult by the heterogeneity of the target population. Some groups within the population = rarely overlapped socially or culturally. For example, MTF cross-dressers might only rarely intersect with FTM people. In light of this deci= sion to target the full transgender communities, I shared the common concern of a social scientist in desiring to depict “the full range of variation in some phenomena” (p. 71).¹⁹⁵ Admittedly, this goal is difficult= in a study of this size. Twenty interviews were planned for the study, and a possible 16 additional people would be included through the two focus groups. In total, only about 36 people were expected to be enrolled in the study. To best address the needs for both diversity and representation, purposive sampling was used. = In purposive sampling, the key is to choose “information-rich cases for study in depth” (p. 169)¹⁶⁶ Within the purposive sampling para= digm, there is one method that is particularly suited for highly heterogeneous populations, maximum variation sampling.&n= bsp; Through this method, a researcher deliberately chooses cases that represent a wide variation of experiences.= Then logically, “Any common patterns that emerge from great variation are of particular interest and value in capturing the core experiences and central, shared aspects or impacts.” (p.172)¹⁶⁶ Thus, having a heterogeneous target population became a strength of the study, as common themes that emerge are that much more compelling as a result.&nbs= p; In structuring the sampling plan, it was first necessary to consider= the categories with which this variation might occur. Considering the findings from the = needs assessments and current theories on gender oppression, I enumerated two sampling categories specific to being transgender: gender vector and degree of gender variance. So, under this cons= truct, the sample needed to include both MTF and FTM people and a range of interpretations of gender variance. Since traditional literature on health disparities showed significant population variations in health by race, age, and socio-economic status I a= lso wanted diversity along these spectrums to be represented in the sample. The final consideration was the ty= pe of data being collected from each person.&nbs= p; The methodology included life history interviews, in-depth key-infor= mant interviews, and focus groups. After considering issues of practicality, I decided it was feasible to apply this sampling plan to both life history and focus group participants but differe= nt categorizations were needed for the key informant group. Thus, outreach and inclusion for d= ata collection from transgender people was governed by the need to obtain information-rich interviews with people representing diversity on each of t= he five key categories: gender vector, visible gender variance, race, age, and socio-economic status.

Key informant sampling was also conducted us= ing purposive sampling, but the key categories of participants were different.<= span style=3D'mso-spacerun:yes'> In this group the information being sought from key informants was not necessarily personal. In some cases a persons’ key-informant status derived wholly from their professional experiences (i.= e. that of the doctors experienced in transgender health care). Some key informants would undoubte= dly be community-members as well, but in those cases their primary value to the st= udy was still as professionals working with the transgender community. After consideration, the sampling categories identified for key-informants were: medical care provider, allied healthcare provider, social service provider, and community ally.

The outreach plan

Using the lessons learned from the prior HIV surveys, I designed an outreach plan with four basic steps: establish value, identify initial community contacts, destribute a verified solicitation, and build inreach. In the paragra= phs below, these steps will be explained and the dissertation fieldwork will be presented as a case study of how they can be implemented.

1.&n= bsp; 3D"Text Establish value

Establishing value first requires the resear= cher to identify and prepare to convey why this project is of value to the target populations. This can be done= in a variety of ways: Does the project lead to better services? to more service dollars? to better policy decisions? to a better living environment? Or alternately, is the project not closely connected with participant benefit, but is not detrimental and also carries intrinsic benefit, such as a desirable stipend?

Establishing value is a process that must be= done in conjunction with community members to ensure that the value is real, not just perceived. In this case,= the dissertation research had only a tertiary link to better services or better policy decisions. Also, commu= nity members reported negative experiences with researchers in the past and a hi= gh level of community-wide mistrust of researchers. After informal discussions with pe= ople about these issues, it became clear that the negative research experiences = were linked by projects that had no perceived value, at least in part because th= ey were conducted by community outsiders, with little or no within-group credi= bility. The first hurdle to establishing v= alue was to establish credibility of the research. Despite previous bad experiences, = the community-members I spoke with demonstrated high altruistic motivation to participate in research. They understood that a research project could be valuable as long as it was conducted credibly, even if the findings did not obviously impact their lives. I would further characterize these discussions by saying that people seemed eager for credi= ble research to be conducted on transgender populations. I considered both the eagerness and history of mistrust in considering how to establish value for the project.<= span style=3D'mso-spacerun:yes'> In this case, it became most important to establish credibility, which would then combine with the high level of altruistic community-based desire for research to create overall value. The credibility of this study was established through three facts:= first, it was conducted by a within-community member, alleviating concerns of sensationalizing or deliberately twisting data; second, it emer= ged from a reputable local university; and third, a dissertation is a well-known research framework, with a level of scientific oversight built into the process. As such, with high c= redibility, even considering the minimal direct benefit to participants, I considered myself prepared to successfully establish the value of the research project= to community-members.

How can a project that does not have those f= actors establish credibility and thus value? A piece of credibility can be established through institutional affiliation alone, but the best mix to establish credibility is community involvement and institutional affiliation. Have community members been involved in the research design, the methodology plans, or instrument devel= opment? Credibility is often undermined be= cause the instrument or methodology strikes community members as illogical or ill-informed. In this study, community member involvement in research design provided some insurance (but not proof) against this possibility. All of these concepts are inherent in the model of participatory act= ion research, where community members work side-by-side with researchers from research design to presentation of findings.¹⁹⁶ The lessons from participatory act= ion research are important, but they become especially crucial when dealing with hidden populations. The dissertation format (i.e. sole author) restricted but did not eliminate the applicability of this method for this project. In particular, a range of community members commented on the research design and assisted in actual outreach.

Another piece of value is any monetary recom= pense for participation in the project. In the case of this project, the personal value of the stipend to participants varied widely. Stipends were not large, participants received $40 for a multi-hour = life history interview, and $20 for a shorter focus group or key informant interview. Nonetheless, there= were participants for whom this was a significant amount of money. For one, the $40 stipend represent= ed half of her financial ability to sleep indoors that night. In some cases, the food that somet= imes accompanied the interviews was also valuable, as illustrated by the partici= pant who reported usually eating out of dumpsters. As in any research study, a stipend is intended to be motivational, without being coercive. While a $20 might be considered co= ercive for a person who was eating out of dumpsters, another element of the total value of the project arose here. Specifically, as the interviewer I = was often left with the distinct impression or explicit comment that the proces= s of telling their story was of benefit to the participants. One participant mentioned that it = was like a therapy session, one said that they really needed to tell this to someone, and most displayed eagerness in the interview that revealed their enthusiasm. Other researchers= have written about the reward of a listening process that “takes the participants seriously, values what they say, and honors the details of the= ir lives.” (p. 83)¹⁹⁷ This “reciprocity” can be key component in the value a participant derives from a research project. This phenomena stands in accord with resea= rch on trauma and the role of narrative in the healing process.^{198, 199 In this situation, the personal
experiences of the participants had been so strongly laced with isolation or
negative opprobrium that having the opportunity to tell their story to a
sympathetic listener carried an
unexpected benefit.}

Identify initial community contacts<= /o:p>

The second step of outreach was to identify = all links into the communities. W= hile I had moved to New York City fairly recently, my work and social connections created a series of access points. At the beginning of t= he dissertation fieldwork, I was engaged (in part because of my outsider statu= s) to be the facilitator for a statewide transgender meeting. At this historic meeting, transgen= der leaders from a variety of statewide organizations were being brought togeth= er to see if they could create a single statewide coalition. Current political exigencies drove= this need, as the possibility of passing statewide anti-discrimination legislati= on loomed but advocacy organizations were then supporting a version of the legislation that only included LGB and not T.

This meeting is in itself illustrative of challenges in working with the transgender communities and the environment I entered to begin the fieldwork. A short profile of the event will help demonstrate this point. Several factors contributed to the meeting’s success, all of which might well have been necessary instea= d of optional. First, a social ser= vice provider with a long history of community involvement convinced a funder of= the value in underwriting the event. As a result, an extensive number of travel stipends were offered. Second, a variety of people respon= ded to the call for participants, crossing the boundaries that might have been cre= ated if one politicized figure took the lead.&n= bsp; Third, a facilitator was used, one who was within-community but perceived as neutral due to being relatively unknown locally. All these steps helped convene a strongly divided group of people. Some participants were personally adversarial to each other, as a re= sult of unexplained historical interactions.&nb= sp; Some were politically adversarial, especially since a recent dispute= had cleft a leadership organization into two opposing factions and supporters of each were present. In short, = the meeting began with a mix of intentions, some collegial and collaborative and others adversarial and suspicious. This was highlighted at the onset, where the first order of business after introducing me as the facilitator became an objection to using any facilitator chosen by a representative of a particular political faction wi= thin the group. This is the role in which I first met many of the transgender leaders in New York State. It was not a role that was directly related to my fieldwork but it highlighted the issues around establishing personal credibility in a highly politicized and contentious environment. Ultimately, and to my benefit, the= group decided to continue using me as a facilitator.

At the end of the event I made an announceme= nt about the dissertation research, asking if I could contact New York City-ba= sed people to help me identify participants.&n= bsp; The response was overwhelmingly positive, and with only a few exceptions, people were willing to be contacted to help identify participants. As = will be demonstrated below, this was an excellent example of distributing a veri= fied solicitation.

In addition to this list of community-leader= s from the statewide meeting, I had a few additional contacts to plumb. For about a month or so I had been attending a transgender support group run by the Gender Identity Project. The participants and leaders were clearly becoming friends and would likely help with outreach. There were also a few transgender = people my friends knew and were willing to help me contact. My partner at that time was workin= g at the Gender Identity Project and would be able to provide me with some connections there, as well as an outline of community venues or gathering places. Lastly, I was doing a consulting job with a CDC research project that was studying the House Ball community in town. The House = Ball community is a tight organizational structure of “houses,” with= in which are “families”, all of whom interact through inter-house competitions, or “Balls”. The whole structure provides a social support system that pulls in m= any Black and Latino gays, lesbians, and transgender people, providing them with role models and mentors to help them avoid the hazards of life. The House Ball communities were mo= stly widely documented in the documentary Paris is Burning, and given fur= ther public attention through Madonna’s song Vogue, modeled on the competitions at the House Balls. This consulting job was fortuitous because many transgender people of color in town were members of a House and participants in this community. As a consultant to the project, I = worked with a team of people who were building an outreach project that would prov= ide HIV testing and interviewing outside of House Ball events. While my role was to help create t= he interview instrument, I came in contact with many other members of the communities and became friends with several of them. So I assembled the list of all pos= sible community contacts mentioned above, and moved onto the next step of the outreach plan, distributing a verified solicitation.

2.&n= bsp; Distribute verified solicitation=

3D"Text The crux of the outreach plan was to distribute information only in a manner that would be seen as credible by t= he target population, a process I’ve named distribute verified solicitations. This is mu= ch like when a stamp is added to a website, showing an independent party has judged them to be safe for online financial transactions. The goal was to make sure that eac= h time a request for participants or information about the project was distributed= , it was done through a medium that carried an implicit “stamp of approval”. This “stamp” was achieved when the request was carried forward by community insiders, by community key influencers, or in places that are vie= wed as community-owned or run.

Information about the project was distributed through three mediums, either through a flyer posting, a personal solicitat= ion, or via email. Verifying the m= essage on the flyer itself was important, as was verification through the flyer placements. The flyer headlin= ed with a reversed box of text broadcasting that the study was being conducted from a within-group researcher: “Transgender health researcher is looking for people to interview for a study about the social issues affecting transgender health.”

The first step in distributing a verified solicitation was to place this flyer at as many within-community sites as w= as reasonable. This required emb= arking on a small public relations campaign about the project. The steps followed are presented below.

Verification of a message relies heavily on individual communication and personal interaction. As noted in the inset above, mass communications should be avoided unless there is a previous relationship, a= nd even then only used sparingly.

Sometimes a verified solicitation can be set= up through a cold call or unreferred contact.= For example, early in the project I wanted to contact the person who fielded the local needs assessment. Since I had no referral, credibility was that much harder to establish. In this example, I established the needed credibility through insider knowledge: I talked about my knowledge of her= work, of the similar body of work that had been conducted nationwide, and referen= ced the other community leaders with whom I was in communication. The contact, Kelly McGowan, was ve= ry receptive and immediately provided leads to help move the research forward.=

By the time I had completed the steps in the= box above, I had at least three agencies with fliers posted, a series of commun= ity leaders who were informed about the project, and most importantly, a long l= ist of places, people and events I should access in order to reach potential participants. The next step i= n the outreach plan was to Build Inreach.=

3.&n= bsp; Build inreach

Building inreach followed the same format as distributing the verified solilcitations, but carried the process further as new people are added to the outreach plan.&nbs= p; The steps followed are presented here.

As inreach was built, the principles of soci= al networks started to carry the project forward, and the number of people who know about it started to build quickly.&nb= sp; Importantly, each of these people received the project information through a verified message. Whenever they carried the message forward, it was de facto verified = as well. As I stepped through the process of this last outreach step, I found myself going to community event= s, arranging for others to solicit participants at their community events, and speaking to a host of people I did not know personally. Inevitably, I could not pursue all= the leads that I was given and simply stepped through as many of them as possib= le in the time available. Then, I started to build my list of people who might want to be interviewed, and was starting to contact them directly to finalize the interview times and locations.

The general call for participants was very effective. In less than a mon= th, long before I expected, I had access to many of the people needed to fill my sampling plan. One byproduct = of the outreach plan was that I was also contacting many of the people who were logical key informants, and any missing categories of key informants were r= eadily available through direct referral. Sometimes people had to be told there was no place for them in the study. But more often, the diversity needed by the sampling plan was pursued by me specifically asking= for a type of respondent, or focusing my solicitations on people or places that intersected with the category of respondent I wanted. For example, referrals from the Ho= use Ball community were much more likely to reach yield people of color. The diversity of the respondent po= ol was good, and I arranged interviews with people who fit most of the categories required from the sampling plan. Within 30 days, general outreach had to be shut down and people call= ing in to participate were put on a list for participation in the focus group.<= span style=3D'mso-spacerun:yes'> It was unexpected that the outreach would yield so many participants in such a short time. There clearly was much more moment= um to be built, and were it not for the limiting resources of time and money, this project could clearly have been many times as large.

As the number of interviewees neared the goa= l, the profile of actual versus desired respondents was compared and a list was ma= de of missing categories of respondents. The same principles used in the other outreach were used to select t= hese final respondents, just with a much more focused, rather than broad-based effort. When no crossdressers presented themselves, I obtained information about the local meeting of Cross-Dressers International. While I did not have a direct referral, I did have the name of the organization&#= 8217;s president. I called on a meet= ing night, spoke to the president and used the same strategies discussed previo= usly to verify my status as a legitimate community researcher with her. As in other cases, she was very responsive. She was able to i= dentify a potential participant from the meeting attendees and an interview was literally scheduled within 15 minutes of initiating the call.

The last category of person I wanted to incl= ude presented a challenge of a different sort.= It hinged on my personal experience with elder dyke communities, with African American dyke communities, and on information arising from a survey conducted by Jessica Xavier in Washington, D.C. In that survey, the majority of Af= rican American FTM vector respondents identified not as transgender but as “Doms”, a shortening of “dominant lesbian.”² I suspected the existing outreach = was not going to access this group because their self-identity did not overlap = with the newly-minted word “transgender”. Despite this distinction in self-identity, I theorized the underlying issues were congruent with the participant group. As a resul= t I wanted to interview an elder Black FTM vector gender variant person, who wo= uld likely not identify as transgender. To reach people in this category, I identified one of my friends who= was in the group herself, and therefore might know of others. I called this friend, explained the study, and provided an additional explanation of why I wanted to interview someone who would not self-identify with the word transgender. The friend was understanding and suggested I go to a Griot = Circle picnic occurring later that week. This is an elder LGBT group, with active participation by elder Black lesbians. At the event, the f= riend introduced me to a variety of elder butch or male-acting lesbians. While I explained my research proj= ect to these people, I carefully gave a caveat that I did not think they were transgender but just wanted to interview one of them on the theory that som= e of the issues were similar. By t= he end of the picnic I had five names and numbers of five people who were interest= ed in doing interviews. As can b= e seen in the findings section, my initial hypothesis was confirmed. The life history from the elder Bl= ack lesbian participant was one of the most dramatic stories and vital expressi= ons of a lifetime of gender variance.

At this stage in the fieldwork, the only interviews that remained were the focus groups. Ultimately both were successful, b= ut outreach for each was conducted in a different fashion. By this time, the interviews had g= iven me firsthand experience with the fact that the MTF participants were likely= to include the least stable people, as MTFs in general received the greatest amount of social opprobrium. = FTM participants were slightly more likely to have houses, phones, email or an appointment book, all of which facilitate scheduling a focus group. Conside= ring this information, I structured each group differently.

The FTM focus group used more conventional outreach methods. I talked to= my FTM support group members, some FTMs I’d met at community events and = the FTM key influencers I’d previously contacted. I drafted an email from the flyer = text, and sent it to key influencers to send on.= The night of the group, I was a bit concerned that not enough people= had RSVPd, but the turnout was sufficient.&nbs= p; There were six participants. Respondents ranged in age from 26 to 49. Self-reported race data was only available for 4 of 6 participants, of these 75% were white and one was black. My perception was that= at least one-third of the participants were people of color. Self-identities spanned a ra= nge of gender variance (as marked by use of terms “genderqueer” versus “male”). My perce= ption was that participants also varied widely in ability to pass, some would nev= er be questioned as male while others clearly were more androgynous in appeara= nce.

The MTF group could not reliably be filled i= n the same manner. Already some of = the people who expressed interest previously were difficult to confirm due to l= ife stability issues. For example= , I would get calls from people who wanted to be interviewed, but they were cal= ling from a social service agency and if I was not able to pick up the call, the= re was no way to return it after they walked out of the office. To overcome this barrier, I contac= ted a local social service provider about conducting the group in conjunction with their weekly transgender drop-in group. While the drop-in group was = for anyone, people with an active relationship with the agency and its social s= ervices were most likely to attend. T= his would bias the respondent pool towards lower SES participants. I decided this was not a problem, especially as interviews had already documented how many MTF transgender pe= ople were pushed towards the lowest SES through discrimination, and how the need= s of this group were little understood and very high. So, the agency, Housing Works, ann= ounced that this focus group was happening a week later, and I referred any new MTF people interested in being participants to this event. Eleven participants showed up to t= his focus group. This participant group ranged from 31 to 57 years old. Self-identity labels were again very diverse but trended towards transexual. Self-reported race/ethnicity was reported for all but three participants. Of this subset, 4 were black, one = was Native American, one was white Hispanic, and two were mixed Native American= and black. To supplement these da= ta with my perception, the group appeared to be predominantly people of color = with only one or two possible white participants.

This second focus group was conducted almost exactly one month after outreach started, and it officially marked the end = of data collection. Within days I would leave town on a planned trip, and I would only ultimately spend only a few more days in New York City= before moving to another state. Personally, it was an abrupt end to a very stimulating experience. For thirty days I was led on a ser= ies of adventures: both through the places I went in New York City, and through the places I= was taken to in the stories that were shared.&= nbsp; I remain honored that so many people chose to help me with the proje= ct, or more, chose to share the intimate details of their lives with me. One day not long before the last interviews I realized one of my personal lessons learned from this fieldwor= k, that this was one of the most exciting times of my life.

Data Collection

Interviews with p= otential participants were scheduled in a variety of locations. The foremost considerations in set= ting location were confidentiality and convenience to the participant. If there was suitable privacy in a participant’s house or place of work and it was determined to be convenient, that was often the first choice. For many, this was not an option, = at which time an office at GMHC that had been made available to me was the fir= st choice for life-history interviews. For key-informant interviews, which had a lower need for confidentiality, restaurants convenient to the participant were sometimes u= sed. In hindsight, I would avoid restau= rants in the future due to the difficulty in transcribing interviews with high ambient sound. For one life h= istory interview, I offered my apartment as the interview site as it was the most convenient to the person’s place of work.

Every effort was = made to ensure a focused, confidential, and secure environment for each interview.<= span style=3D'mso-spacerun:yes'> But the reality sometimes strayed = from this principle. During one li= fe history interview I learned someone was sleeping in the top of the bunk bed= I was sitting on. Later in the = same interview a third person came, sat down and started to play video games at = the foot of the table in front of me. While I became uncomfortable, the participant did not. At each juncture, I confirmed that= were fine discussing intimately personal matters in this situation and they were. In another example, lac= k of foresight raised a security issue, although not for the participant, instead for me as the researcher. I h= ad met a participant at her apartment for an interview, but I did not plan on her desire to do the full interview in one sitting. After hours of talking I found mys= elf in a situation that was outside of my level of personal comfort, heading alone down the stairwell and across the yard of a large Br= ooklyn housing project late in the evening. A summary of interview location is presented in the table below.

Locations<= /o:p>	# of Interviews=
Participant house<= o:p>	7
Participant work	5
GMHC	5
Restaurant	2
Phone	2
Other	1

Life history inte= rviews were conducted in an unstructured format.&= nbsp; People were asked to give their age, race, and gender identification= at the beginning of the interview. After that, they were encouraged to tell their life story as they ch= ose. Occasionally prompts were used to elicit additional information. Questions were most often used whe= n the narrative appeared to skip swathes of time, and then were used to redirect towards filling in the missing information. Occasionally questions were used to ask health outcomes related questions. But most often, the health information that emerged was at the choice of the participant. Key informant and focus group inte= rviews also followed this unstructured format.&nb= sp; Instead of life stories, I encouraged participants to talk about the issues they felt affected transgender health outcomes. Sometimes if a major theme was not brought up (i.e. addictions), I asked it with a prompt. <= /p>

As a community-ba= sed interviewer, I approached the participants from the standpoint of a collaborative relationship. A= t the beginning of the interviews, we arranged ourselves on chairs as we chose, a= nd if possible I would get the participant a snack or meal to eat during the time. The informed consent wa= s read and filled out first. Then a = portable tape recorder was set up and tested. In all but the first few interviews, a detachable microphone was also used to enhance sound quality. After the consent, the taping started. I began by reciting the date and t= he interview number then we proceeded to the content of the interview. The flow was interrupted as needed= to change tapes. After the inter= view ended, the participant was thanked, and either given their stipend or set up for their final interview. Li= fe history interviews ranged in length from 45 minutes to five hours. As the interviews proceeded, more = became single session events, even if they were a life history. Despite the length, participants o= ften preferred finishing the interview in a single sitting.

A process log was= kept during the month of interviewing. The process log documented my thinking about the next steps needed, = the outreach conducted, and any thoughts about the interviews. One issue that arose was stress re= lated to hearing the interviews. Wh= en engaged in a life history interview, I experienced a level of immersion that had been absent from previous shorter interviews. I often would find myself transcen= ding the traditional parameters of listening, and feel as if I had experienced a shadow version of the story. = This phenomena was very interesting. Sometimes it was very positive.&nbs= p; There were benign experiences that were relayed of common cultural phenomena which I had long thought I understood. But only after hearing them throug= h the experiential standpoint of a life history interview I realized that my understanding and acceptance of them as common phenomena was limited by my outsider status. Hearing the = life history interview immersed me in an insider’s perspective for long en= ough to expand my acceptance of the particular topic. For example, in my personal cultur= al experience (as a white middle class person), it is rare or novel to have relatives on drugs or in prison. While I had long understood that this was common fact of life for ma= ny working class inner-city people of color, my understanding was academic. In contrast, I realized after a fe= w of the life history interviews that my awareness of this phenomena became much more experientially-based, creating a greater level of acceptance of it that incorporated less personal novelty. Sometimes the experience of immersion into an interview was negative= , and I would often find myself dazed and overwhelmed by the interviews that recounted the greatest level of violence, or that included unexpected viole= nce in the narrative. I routinely used my existing therapy sessions to alleviate the personal stress that emerged as a result of the interviews.

Life History Participant Pro= file
Gender Vector	Pseudonym*	Age**<= /p>	Race/Ethnicity
F= TMs	C= arrington	30	w= hite
<= o:p>	C= raig	30	b= lack
<= o:p>	C= hris	20	w= hite
<= o:p>	D= erek	30	b= lack
<= o:p>	J= enna	50	b= lack
M= TFs	D= arlene	40	w= hite
<= o:p>	I= sabella	20	w= hite/hispanic
<= o:p>	M= arnie	40	w= hite
<= o:p>	A= nna	50	w= hite
<= o:p>	N= irvana Rose	30	A= API/hispanic
<= o:p>	R. Francine Bailey	50	b= lack/nat am
<= o:p>	S= affron	20	w= hite
<= o:p>	S= hashatika	20	b= lack/hispanic
*= Francine has asked that her real name be used.
= Ages are rounded to nearest decade for anonymity.

Data Transcription, Cleaning, and Verificatio= n

Upon completion of the interviews, transcrip= ts were created by staff at the Survey and Evaluation Research Laboratory at <= st1:place w:st=3D"on">Virginia Commonwealth Univer= sity. Upon return, I cleaned and verifie= d all transcripts as follows. Each transcript was first read thoroughly. Some obviously were deficient due to either problems interpreting the words of the participant or due to poor recording quality (e.g. as in the c= ase where an air conditioner provided obtrusive background noise). In these cases the tapes were eith= er reviewed in needed areas, or in the most extreme cases, reviewed in full to check against the written transcripts.&nbs= p; Only one tape was reviewed in full.= This was necessary since problems with interpreting the participant’s dialect had resulted in many misinterpretations on the transcript. After this proces= s was completed, almost all problems with transcription were remedied, and only a= small number of comments proved indecipherable.&= nbsp; While cleaning each transcript, special care was paid to identifying= and systematically changing any names inadvertently used in the interview. A record of the original names and= the correlating pseudonyms was kept in a password protected computer file to en= able me to keep narrative threads about other people intact. For those transcripts that appeare= d to be clear and without omissions, I still checked the written words against t= he source tapes for at least two passages, to verify the accuracy of the docum= ent.

Data Analysis

The process of qualitative data analysis is not governed by strict procedural formulas, Crabtree notes “there are nearly as many analysis strategie= s as there are qualitative researchers.” (p.17)²⁰⁰ According to Gibbs, there are two hallmarks of qualitative data analysis, that it is an interpretive philosop= hy, and that it takes a holistic view of the issue, “the proper understan= ding of people’s words can only be achieved if these are related to the wi= der context in which they have been used.” (p. 2)²⁰¹ This reflects Denzin’s outli= ne of steps in analyzing qualitative data. In these steps, a researcher first loses themselves in the data and reduces phenomena to their essential elements; this step is named intense involvement and bracketing= . Next the researcher explores tentative interpretations of the data. This is followed by a p= rocess of contextualization of the dat= a, where the data is relocated into the lives of the participants. Then, the researcher makes an examination of deviant cases, adva= nces alternative explanations, and fina= lly presents the analysis.¹⁶⁸ This process was used to frame and structure the data analysis for this project.

Coding and data reduction were accomplished = through the qualitative analysis software program Nvivo, version 5. 0. While the a priori framework of he= alth determinants was available, the initial process was closer to open coding, whereby the immersion in the data guided the development of coding nodes. ADDIN EN.CITE <EndNote><Cite><Author>Strauss</Author><Year>= 1990</Year><RecNum>37</RecNum><record><rec-numbe= r>37</rec-number><ref-type name=3D"Book">6</ref-type><contributors><author= s><author>Strauss, Anselm L.</author><author>Corbin, Juliet M.</author></= authors></contributors><titles><title>Basics of qualitative research : grounded theory procedures and techniques</title></titles><pages>270</pages><ke= ywords><keyword>Social sciences Statistical methods.</keyword></keywords><dates>= <year>1990</year></dates><pub-location>Newbury Park, Calif.</pub-location><publisher>Sage Publications</publisher><isbn>08039325020803932510<= /isbn><call-num>HA29 .S823 1990300/.72HA29Social Work HA29 .S823 1990= </call-num><urls></urls></record></Cite></= EndNote>²⁰² From this eight major nodes emerge= d.

As coding continued, subcategories were iden= tified for these major nodes, until all data were placed in either a major or subcategory as was applicable. Then, in the process of tentative interpretations, these categories = were reflected back upon the framework of social determinants. As I contexualized the data I star= ted to make choices on which issues emerged as a higher priority, for example, the downward spiral on the social gradient was profound in some cases, but in t= he full story, how did that relate to the experience of extreme violence? Clearly I could say that all social determinants affected the life of transgender people, but as that was a description, not an explanation. Providing only a description would abdicate my responsibility to critically analyze the impact of these data. In further exploration of both dat= a and issues related to social determinants, a cohesive explanation began to emer= ge, and was sustained even after consideration of deviant cases and alternative explanations. A phenomenon I eventually labeled ‘compromised survival’ was the true hallmark= of these stories, and that phenomena was primarily driven by extreme underlying experiences of stress/violence, social support, and social exclusion. Additional issues such as addictio= ns and health outcomes also deserved particular note, but they were mediated by the primary factors listed above. These foci then developed into the context of the findings, and I proceeded to organize them into the content for this report.

Ethical Considerations

Ethical considerations are of increasing importance in participant research. As Institutional Review Boards tighten their guidelines, governmental oversight bodies still find ethical violations in even large well-regarded studies, so we are better reminded that it is the researchers’ responsibility to maintain all appropriate ethical standards. As the researcher for this study, = I was trained in ethical study practices and operated under a study protocol that= was approved by the Columbia University Institutional Review Board. Under this protocol, all promotion= al materials were approved by the IRB to ensure they were not unduly coercive = or misrepresentative of the study. Participants went through the process of providing a written informed consent before their first interviews.&nbs= p; As is standard, this document delineated the study process, personal risks and benefits, and an outside contact for any concerns. To reduce concerns about literacy, participants were offered the option of having the consent read aloud to them. The informed consents for key influencers allowed the option that an interview could be on the record or confidential.&nb= sp; Ultimately, all but one chose to have their interview on the record. Focus group and= life history interviews were automatically confidential. Despite this, one participant deliberately went out of her way to state that she wanted her interview to = be on the record. To ensure the integrity of the data, informed consents were separated from the interviews= and kept in a locked filing cabinet at all times. Tapes from the interviews were only available to the transcriptionist and the researcher. During the process of cleaning the transcripts, any individual names mentioned during confidential interviews = were replaced with pseudonyms in the transcripts.

Plan for Study Results

This dissertation has been prepared to parti= ally fulfill the requirements for graduation from the Sociomedical Sciences prog= ram at Columbia University. Fulfilling my degree requirements = is the first of three planned uses for this study data. The second planned use is to refas= hion the information to introduce it into the academic arena as quickly as possible. To this end, I will prepare a series of journal articles and conference presentations from this work. The body of this disser= tation has been crafted to be easily split into a series of shorter articles on the social determinants highlighted in the Findings section. To supplement these articles, I wi= ll prepare a conference presentation on my findings. Funds permitting, I will submit th= is presentation to applicable conferences, in particular the Transgender Health Conference and that of the American Public Health Association. The third use to which these data = have been committed is the creation of a community report of the study findings. This will follow the precedent of the ‘Research Highlights’ publications first introduced to me by Virgi= nia Commonwealth University’s Survey and Evaluation Research Laboratory staff. This style presents study findings= in a short newsletter-format, emphasizing the use of plain language and those ar= eas of most interest to the participants and service providers. While it will be available to the = most people through online publication, I will publish a minimum of 100 hardcopi= es for distribution back through the service providers and community-access po= ints utilized in this study. These= three outlets are planned for the study findings.

While the analysis presented here will be us= eful to other researchers and community members, the stories themselves are also powerful documents. In the wo= rds of Howard Becker, perhaps the most important service performed by a life histo= ry for sociology is that of being a “live and vibrant message…. Telling us what it means to be a kind of person we have never met face to face.” (p.70)²⁰³ It is my hope that I wi= ll be able to disseminate the stories collected in this research through some med= ium as stand-alone narratives. Re= ading the life history of another creates a level of affinity for their experience which cannot easily be imitated by any other means. Even as a relatively aware communi= ty insider, I was profoundly affected by immersing myself in the life historie= s of the study participants. On re= ading these stories, I think few could fail to respect the amazing tales of survi= val. This respect is a valuable tool to combat the poisonous effects of anti-transgender stigma.