« Previous | Contents | Next »
Listen
Sexual Orientation Research Phase 1: A Review of Methodological Approaches
6. Diversity and inclusion in the research process
6.1 Community consultation and participation
There is a lot of variation across sexual orientation research projects in the extent of LGBT consultation and participation in the research process. While researchers are often dependent on local LGBT organisations or venues for construction of a sample frame, few stated that LGBT groups provided input into determining the research agenda, developing topic guides or questionnaires, or in the analysis of the findings. While the importance of community consultation and participation in the research process is becoming more clearly acknowledged, it is also important to balance this against avoiding research fatigue; overburdening individual participants and the wider LGBT community; and biasing data by being influenced primarily by the input of the few participants who have the most free time or are the most highly motivated.
The Brighton and Hove's LGBT community needs assessment stated that the contact made with a wide variety of LGBT groups "proved a very useful strategy not only for generating questions, but also publishing the project and developing ownership within LGBT groups and services."(Webb and Wright, 2001) One key structure for such input is through membership on an appointed and involved steering committee to establish priorities for, guide and review the research progress.
6.2 Groups excluded from research
There are two key types of 'exclusion' in sexual orientation research which researchers need to actively address in the design of studies: people who are rendered invisible because they have not participated in the research (for whatever reason) and people who are invisible because they are unidentifiable in the data (for example, because sexual orientation was not asked).
Reasons for not having participated in the research include being:
excluded from the sample frames used, for example those who are not 'on scene', who live in rural areas, or do not access LGBT media or websites;
excluded by the research method employed, for example because of literacy or eyesight problems they can not complete a postal questionnaire or they do not own a land line telephone and so are excluded from telephone surveys;
excluded by the research because they perceive the participation to be too great a burden, an intrusion or not of relevance to them.
Each of these areas of non-participation can impact on the representativeness, nature and size of the sample, as well as on the response rates and the findings of a study, as those who do not take part may differ from those that do in ways which are relevant to the research subject.
As discussed in Chapter 3, much sexual orientation research achieves samples biased towards young, white, articulate, middle class, well-educated men. Gonsiorek and Weinrich argue that "like most scientific discourse, research on homosexuality has been filtered through societal biases. Among its other effects, this has meant that white, middle class and above, adult, English speaking males have been studied more than their numbers warrant."(1991) In Scotland, the groups most absent from research include older and disabled LGBT people and those living in rural areas, ethnic minority and transgender people, lesbian and bisexual women. Inclusion of these groups is something that needs to be actively incorporated into the research design.
6.2.1 Sexual orientation
As discussed in Chapter two, definitions of sexual orientation when they are used to determine eligibility for a study may exclude potential respondents who are in fact relevant to the subject. Recruiting only LGBT identified respondents may exclude men who have sex with men, women who have sex with women, or others who are homosexually active but do not self-identify as LGBT.
6.2.2 Transgender and transsexual people
There is a severe lack of research with identifiable transgender respondents and into transgender issues. Transgender respondents are often not included in sexual orientation study samples, because their needs and experiences are perceived to be so different from those of other LGB respondents. Morrison and Mackay state that transgender people were not included in their sample because "the study team recognised at the outset that the experiences of transgender people were outwith our current knowledge base and required considerable examination before any early assumptions about experiences or the methodology required to research these could be made."(2000) Grossman and Kerner also excluded transgender (and bisexual) youth from their study of lesbian and gay young people's social networks. "Many of the transgender youth at [the Drop-In Centre] tend to cross-dress and seem to receive less support from the adults in their lives."(1998) Any bisexual or transgender youth who expressed an interest in participating in the research were informed that since their needs were "unique, it would be necessary to design a specific study for them." It is important therefore to recognise that research into 'LGBT' issues is unlikely to redress the lack of work with transgender respondents, and that a more targeted approach is required.
Attempts to measure the prevalence of medically defined transsexuals are usually based on those reporting to psychiatrists for treatment, which is also likely to undercount overall prevalences as those that do not seek treatment, are not medically defined as transsexual, or seek treatment abroad, will not be included in national prevalence rates. The extent to which rates are an underestimate will depend on each country's availability, accessibility and affordability of services, as well as cultural factors.
6.2.3 Gender
Even prior to the 1980s shift in research focus onto AIDS and sexual health, there has been an emphasis in sexual orientation research on gay (and to a lesser extent bisexual) men. Smailes in her study of lesbian' experience of housing states that her work is situated within "the general oppression experienced by women in a patriarchal society. Some of the issues are common to both lesbian and gay men; however, men do not experience sexism or the issues faced by women living a life without reference to male approval."(Smailes, 1994) The danger of arguments which place the experience of lesbians and women into a separate frame to that of gay and bisexual men is that it justifies their future exclusion from other studies, which is dangerous when lesbian orientated research attracts so little funding. Morrison and Mackay cite Von Schulthess as their reason for not including women in their study of violence in Edinburgh. However Von Schuthess states that her goal in that research project is "to develop a better understanding of the unique features of anti-lesbian violence," and therefore it might be expected that she should focus on differences rather than similarities, particularly given the emphasis in previous research. A similar argument for excluding ethnic minorities from the Edinburgh study because an attack may be prompted by their ethnicity would be absurd, and yet not entirely distinct from Morrison and Mackay's rationale for excluding female respondents.
GALOP criticism of Mason and Palmer's 'Queer Bashing' (1996) survey for assuming that lesbian and gay male experiences were sufficiently alike that they could be addressed within one research approach, seems to neglect the fact that filtering in questionnaires has always enabled the questions asked to coincide with individual respondents' experiences.
6.2.4 Ethnicity
Grossman and Kerner argue that LGBT youth have many stress-related experiences to contend with that their heterosexual counterparts do not "and the challenges may be more so for members of racial and ethnic minorities. These include: reconciling sexual orientation and ethnic and racial identity, facing discrimination among gays and lesbians, and having difficulties coming out to family members."(1998) Despite this, there is very little sexual orientation research with ethnic minority respondents, particularly in Scotland.
As with sexual orientation and gender, ethnicity is a complex, composite concept.(Ndofor-Tah, 2000) What often happens however is that because too few participants in particular groups were identified for analysis, all minority ethnic respondents may be grouped together (e.g. Pilkington and D'Augelli, 1995; Von Schulthess, 1992) or else findings are reported on small cell sizes (e.g. GALOP study reports on 7 Irish respondents, 1998). Comparing 'black' and 'white' respondents is not a useful distinction given the great diversity within these groups. Even amongst 'Black Africans' or 'south Asians' there is a wide range of different cultural, religious and regional diversity, but it is accepted in much research practice that it is not viable to usually break this group down further.
Obtaining a sample of LGBT respondents of minority ethnic origin is repeatedly described as problematic, particularly for quantitative research that requires sufficiently large and representative samples before generalisations can be made. Webb and Wright's comment was a recurrent one: "we also wanted to look at differences with regard to ethnicity, but…the sub-sample of respondents from minority ethnic communities was too small to allow for this."(Webb and Wright, 2001) Large scale studies in the US and Australia have recruited respondents from particular minority ethnic groups through 'Black Pride'(Battle
et al., 2002) or 'Asian Pride'(Prestage
et al., 2000) events. The lack of such events and the low prevalence of ethnic minorities render this approach currently untenable in Scotland. Cultural taboos and culturally inappropriate language and survey instruments also effect researching sexual orientation in some ethnic minority communities (see Elam for discussion, 1999). The National Survey of Sexual Attitudes and Lifestyles included an ethnic boost, with the questionnaire in translation and interviewers available who could speak Punjabi and Urdu, however only a small minority of these ethnic minority respondents resided in Scotland. (Johnson
et al., 2001)
In Von Schulthess' study of anti-lesbian harassment, she notes that many of the minority ethnic respondents "commented on the connections among their triple minority status and their experiences with violence."(1992) While the white lesbians in her sample were more likely to report a verbal assault, "lesbians of color" reported higher rates of physical violence, threats, vandalism, and rape.
6.2.5 Disability
There has been little sexual orientation research which has sought to be inclusive of disabled respondents or which has tackled disability issues. Mobility impairments may reduce some LGBT people's level of access to LGBT venues where research recruitment may be taking place. Sensory disabilities, such as hearing and sight impairments, and learning and cognitive difficulties can require different research methodologies to improve access. Some of these approaches, depending on the individual situation, might include:
large print on any documents or self-completions,
face to face instead of self-completion or telephone interviewing,
availability of help-line facilities,
clear and unambiguous instructions and questions that have been cognitively tested for clarity,
interviewers who can sign, or use of interpreters,
reading out of show cards,
asking what help could be provided in advance of an interview,
use of minicoms and talk-text,
interviewer sitting in the light, so mouth is clear and well lit,
taking rest breaks during an interview,
the use of specially trained interviewers or facilitators where appropriate,
Braille format questionnaires,
especially designing questionnaires and/or show cards to make greater use of pictures and symbols,
and using the telephone for following up non-respondents and individuals with special needs.
6.2.6 Age
Sampling strategies involving recruitment of respondents from commercial venues, mailing lists, STD clinics and so forth tend to sample highly community active individuals, which also serves to exclude both the youngest and oldest LGBT people.
Younger respondents
There are inherent methodological limitations to conducting research with younger LGBT (and questioning) people - these are an almost invisible group, with major issues and no adult community to identify with (Martin, 1996). Pilkington and D'Augelli (1995) in their study of victimisation of LGBT youth, argue that the "most critical limitation" to the few studies they could find on the issue was the "reliance on adults' retrospective accounts of events that had occurred in adolescence." They state that the few instances where youth have been surveyed contemporaneously these studies have tended to focus on white gay male youth. Studies have often been restricted to university campuses or particular areas, further reducing the generalisability of findings. "Another weakness of these studies has been a failure to define 'youth' consistently. For instance, nearly three-quarters of the 'youth' in [a 1993] report were 21 years old or older."(Pilkington and D'Augelli, 1995)
Grossman and Kerner's (1998) study included respondents as young as 14, but they do not discuss any ethical or consent issues arising out of this, increasingly an issue for research in Britain.
34 In Davies and Feldman's 1997 Welsh study of 15 to 23 year old sex workers (the interviewer sometimes noted that the respondent looked younger than he claimed to be) half of the respondents lived with their parents. Despite this, no mention is made of any ethical consent procedures having to be met. A related issue Grossman and Kerner discuss that particularly impacts on research with LGBT youth is that definition is likely to an even greater impact on a younger sample, given that, for example, "self-identified youth may be different from those who have yet to identify themselves to others as gay or lesbian." The Stonewall study of youth homelessness in south east Scotland (2000) had respondents as young as 13, but also identified the difficulty of monitoring younger people's sexuality and the impossibility of gathering data from those who have not yet 'come out'.
Various subject areas are particularly pertinent to younger respondents, including realisation of sexual identity, 'coming out,' and violence (teenagers are 2 _ times more likely to be the victims of violent crimes than people over the age of 20 (Pilkington and D'Augelli, 1995)). Mason and Palmer (1996) also found the young to be the group reporting the highest levels of violence and abuse. "This seems to us no accident. Our schools are particularly constrained from dealing with issues of homosexuality. Section 28 of the Local Government Act, which prohibits local authorities from intentionally promoting homosexuality, along with attacks on sex education, and the unequal age of consent for young gay men, have created a situation in which many young lesbian and gay men grow up in a climate of fear and abuse which teachers are either unaware of, or feel powerless to deal with."(Mason & Palmer, 1996)
See Lee(2000) for a substantive review of (US and medically orientated) literature on young people, covering school-related problems, suicide, substance misuse, homelessness, antigay violence and prostitution.
Older respondents
Very little of the research on LGBT issues has specifically focused on older people, and most studies either deliberately impose a cut-off age or simply fail to identify or recruit many older respondents. In one of the earliest major studies of ageing amongst gay men,
Gay and Gray, the American researcher Raymond Berger explained he had to abandon his original intention to include women because he found that he could not identify a sufficient sample aged over 40.
35 In 1988 Kehoe conducted a broad-ranging study of ageing amongst lesbians sixty or over living on the east coast of America; broad studies of this kind do not seem to have been conducted in the UK.
36 A few recent studies have addressed the specific housing needs of older LGBT people, however, in one more than 60% of the sample was aged under 45 (the North British Housing Association, 1999). Hubbard and Rossington point out in their study of the support needs of older lesbians and gay men that they found virtually no existing research in the field: "older lesbians and gay men have been ignored in this country, both by the gay and lesbian community and by statutory providers of community care and housing services."(1995) Hubbard and Rossington point out that although there has, relatively speaking, been a wealth of LGBT research carried out in the United States, there are 3 key problems with attempting to translate these to a British context. Firstly, US approaches tend to emphasis individualistic and therapeutic approaches; secondly, the social policy contexts in Britain and the US are so different; and thirdly the abstract focus of US work on ageing.
From a sampling point of view, a number of articles cite the lack of networks and organisation specifically for older lesbian and gay men and their invisibility on the commercial pub and club scene. Very elderly respondents may be less able to cope with some methodological approaches, such as self-completion questionnaires or anything requiring small print to be read.
Where older LGBT people are a particular crucial and rich resource is in the collection of oral histories of the situation of LGBT people in the last century. Over the past decade a number of US and UK publications have emerged using a 'life story' approach, including Kennedy and Davis (1993); Brighton Ourstory Project (1993); and Porter and Weeks (1991).
6.2.7 Rural coverage
The urban biases of empirical research on gay men, women, and families have resulted in minimal knowledge about gay people in rural settings. This is an area that the consultation with LGBT community representatives highlighted as of particular concern for Scottish sexual orientation research.
There has been a long tradition of urban bias in research on sexual behaviour and research addressing LGBT issues, for example, all Kinsey's reports drew on urban-living samples. In addition, analysts of rural communities have tended to ignore the existence of rural gay people. (D'Augelli and Hart, 1987)
Perry notes that much previous research on the experiences of young LGBT people has focused on big cities in the US and UK, her study identifies issues relevant to young people living in smaller areas where "information and services could be more problematic."(1999) Another study of young people included smaller areas in order to assess the role of the internet in places with no 'off-line' community.(Gray, 1999)
6.3 Conclusions
There is a need to balance providing a role for LGBT consultation and participation in the development of research agendas and processes, with avoiding overburdening participants and biasing data by being unduly influenced by the input of participants who have the most free time or are the most highly motivated. Wide dissemination of research findings should be a priority for sexual orientation research in Scotland. This should make use of LGBT media, but also the Internet and 'mainstream' media where appropriate, to increase the likelihood of those who are not LGBT identified being able to access research findings.
Particular sections of the Scottish LGBT community tend to be excluded from sexual orientation research, for a variety of reasons. The access issues for transgender people, lesbians and bisexual women, minority ethnic people, people with disabilities, younger and older respondents, and those living in rural areas all need to be tackled. Inclusiveness in research is something which researchers and funders need to actively incorporate into the first stages of developing a project's sample design and research methodologies.
« Previous | Contents | Next »