3
Methodological Challenges in Conducting Research on Lesbian Health

Conducting research on lesbian health presents numerous challenges because lesbians represent a subgroup of women for which standard definitions of the population are lacking and lesbians are not readily identifiable. These challenges are further compounded because many in the lesbian community distrust research and researchers and there has been little funding support for conducting research on lesbian health topics. It is not surprising, then, that methodologically rigorous large-scale studies are lacking in this area. Furthermore, a number of methodological challenges for comparing findings across studies are consistently found in lesbian health research.

1.  

Inconsistencies in the way sexual orientation is defined make it difficult to compare findings across studies. Studies have not been consistent in how they define a lesbian sexual orientation, with some focusing on sexual behavior and others focusing on identity or desire. Studies have also used a range of time frames in which to capture reports of past or present sexual behavior, some for example looking at behavior during the past six months or a year and others looking at lifetime behavior. As discussed in Chapter 1 these can all be appropriate ways of assessing sexual orientation depending on the needs of the study. How-



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 97
--> 3 Methodological Challenges in Conducting Research on Lesbian Health Conducting research on lesbian health presents numerous challenges because lesbians represent a subgroup of women for which standard definitions of the population are lacking and lesbians are not readily identifiable. These challenges are further compounded because many in the lesbian community distrust research and researchers and there has been little funding support for conducting research on lesbian health topics. It is not surprising, then, that methodologically rigorous large-scale studies are lacking in this area. Furthermore, a number of methodological challenges for comparing findings across studies are consistently found in lesbian health research. 1.   Inconsistencies in the way sexual orientation is defined make it difficult to compare findings across studies. Studies have not been consistent in how they define a lesbian sexual orientation, with some focusing on sexual behavior and others focusing on identity or desire. Studies have also used a range of time frames in which to capture reports of past or present sexual behavior, some for example looking at behavior during the past six months or a year and others looking at lifetime behavior. As discussed in Chapter 1 these can all be appropriate ways of assessing sexual orientation depending on the needs of the study. How-

OCR for page 97
--> ever, researchers have usually failed to state how they define sexual orientation. 2.   The lack of standard measures, including measures of sexual orientation, makes it difficult to compare findings across studies. Studies of lesbian health have lacked standard measures of sexual orientation including its three components—behavior, identity, and attraction or desire—which makes comparisons among studies difficult. In addition, like much other research on health-related behaviors, studies have often lacked standard measures of such variables as alcohol consumption and drinking behavior, depression, and childhood sexual abuse. 3.   The use of small, nonprobability samples limits the generalizability of findings. Most lesbian health studies have relied on nonprobability samples. In particular, many studies have used convenience samples (e.g., from lesbian bars, music festivals, gay and lesbian organizations). As discussed later, these nonprobability samples are not likely to be representative of the population of lesbians. Further, most lesbian women sampled have been white, middle-class, well educated, and between 25 and 40 years old (Hughes and Wilsnack, 1997) and thus may not be representative of other socioeconomic, racial or ethnic, or age groups of lesbians. 4.   The lack of appropriate control or comparison groups makes it difficult to assess the health of lesbians relative to other groups of women. In many research designs it may be useful to compare lesbians to another subgroup of women (e.g., heterosexual women, women in general). However, few studies have allowed direct comparisons between lesbians and other subgroups of women by using the same sampling strategies to identify subjects across sexual orientations and including measures of sexual orientation. Some studies have used as a comparison findings from earlier studies of women randomly selected from the general population. Although this method is an improvement on having no comparison group, the two groups are often quite different in terms of several key demographic variables. For example, most studies of lesbian health have included women who are more highly educated, of higher socioeconomic status, younger, and more predominately white compared to probability samples of women in general. In addition, it is important to note that samples from the general population of women include some

OCR for page 97
--> unknown percentage of lesbians whose results affect the general population findings in undetermined ways. In selecting a sample comparison group, it could be useful to consider some of the matching strategies commonly used in epidemiological case-control studies where the intent is to sample a control group that is similar to the case group on one or more specified characteristics. In applying these general principles from epidemiology, the comparison group of women could be matched to the lesbians on a pairwise or groupwise basis. Some lesbian health studies are using this approach by defining the comparison group of women as a sister, work colleague, or neighbor of the lesbian research participant. 5.   The lack of longitudinal data limits understanding of lesbian development and its implications for how to define and measure lesbian sexual orientation. Most existing studies portray cross sections of experience at one point in time, rather than development over time. Although discontinuity and change characterize the lives of many lesbians, the available cross-sectional data cannot address compelling questions of behavior and identity across time. Prospective, longitudinal studies are essential for understanding vulnerability, resilience, and well-being of lesbians across their life span. In the following sections, several key methodological issues in conducting research on lesbian health are discussed briefly: defining and sampling the study population, developing instruments to assess being lesbian, and eliciting disclosure of information.1 1   The research design and data collection issues involved in conducting research on sexual behavior were addressed more comprehensively by the National Research Council Committee on AIDS Research and the Behavioral, Social, and Statistical Sciences (NRC, 1989, 1990). The reader is also referred to Bradburn et al. (1979) for a more detailed discussion of response effects to threatening questions in survey research and strategies for improving interview methods and questionnaire design. An in-depth discussion of the methodological issues in conducting research on lesbian health is beyond the scope and resources of this workshop study. The intent of the committee is to highlight the range of issues involved in doing research with lesbians and to suggest some approaches for addressing them. Numerous books on research design and methodology are available that provide more detailed and technical analyses of these issues.

OCR for page 97
--> Defining the Population A critical initial step in conducting research is to clearly specify the target population you wish to study. To sample and identify the lesbian study population, researchers must thus clarify how they have defined sexual orientation in terms of identity, behavior, attraction or desire, or some combination of these. The type of question (e.g., identity versus behavior versus attraction) should be driven by the hypotheses being assessed. Thus, if one wishes to do a study of women who describe themselves as lesbian, perhaps to explore the process of coming out to family members, it may be sufficient to ask a single question such as, Do you consider yourself to be a lesbian? Women who answer yes can then be considered eligible respondents or subjects. In another instance where one wishes to study, for example, woman-to-woman transmission of a sexually transmitted disease, this question would not be a very appropriate strategy for identifying eligible participants. Some women who self-identify as lesbians and would thus qualify for the first study would not fit parameters of the transmission study because they have never had sex with a woman, they may be in an active sexual relationship with a man, or they are not sexually active. Definitions of the lesbian population and assumptions about its composition and behavior have varied in this body of research, a characteristic also of research on gay men and homosexuality in general. A review by Sell and Petrulio (1996) of 152 public health articles published between 1990 and 1992 that included gay men or lesbians revealed that only four of the studies reported the conceptual definition of sexual orientation employed to identify the population sampled (e.g., indicating that they defined homosexuality in terms of sexual behavior or attraction). The remaining studies typically relied on self-identification to define subjects or defined their sample based on the setting from which it was obtained (e.g., gay and lesbian organizations, bars, clinical settings). Table 3.1 presents information about the way in which lesbian sexual orientation has been defined in a wide range of studies. As illustrated in the table, specific questions have varied across studies, and studies have focused on different components of sexual orientation.

OCR for page 97
--> We found that there were just so many different varieties of combinations of self-identification, of attraction to women and men, and sexual behavior, either current sexual behavior or past sexual behavior, that well over 30 different groups were defined by the combination of those variables Alice Dan, Public Workshop, October 6-7, 1997 Washington, D.C. Instrumentation There are no agreed-upon standard questions with which to assess whether or not a woman is a lesbian (see Chapter 1). Once researchers wishing to assess sexual orientation determine which dimensions or aspects of sexual orientation are most relevant to the study, they must then decide how to measure the chosen dimensions. Researchers have used an array of questions to identity lesbians, focusing on the different components of lesbian sexual orientation: self-identification, sexual behavior, or sexual attraction or desire. Further, there is variation in the time periods during which the different components were assessed (e.g., lifetime or recently). However, there are no standard questions for measuring these dimensions. Thus, in one study, women might be identified as lesbian if they had had only female sex partners during their lifetime; in another study, women might be identified as lesbian if they had had any female sex partners during the past five years. This lack of standardization has made it very difficult to compare results across studies. Sell and Petrulio (1996) found that self-reported lesbian, homosexual, or bisexual identity was by far the most common method used to categorize lesbians in public health research in the studies they surveyed (see Table 3.2). The committee does not believe that enough information is available at this time to determine what, if any, particular wording is best in questions designed to elicit information about various aspects of sexual orientation. Methodological research is needed to improve measurement

OCR for page 97
--> TABLE 3.1 Summary of Measures of Sexual Orientation Used in Studies of Lesbian Health   Measures Used to Assess Sexual Orientation Study Identity Behavior Desire Focus of Study (sample size) Berger et al., 1995   Sexually active during lifetime with one or more female partners; no sexual intercourse with men during past 12 months   Whether bacterial vaginosis occurs in lesbians and, if it does, whether it is sexually transmitted (n = 103) Bevier et al., 1995   Sex of sexual partners since 1978; engagement in vaginal intercourse, oral-penile sex, and/or anal intercourse in more than 50% of sexual encounters   Characteristics, behaviors, and HIV infection of women attending an STDa clinic who reported same-sex contact (n = 135) compared to women who had sex only with men (n = 1,383) Bloomfield, 1993 Whether women report being primarily heterosexual, primarily or exclusively lesbian, or bisexual     Drinking patterns of self-identified heterosexual women and self-identified lesbian or bisexual women in San Francisco (n = 844) Brand et al., 1992 Scale from 1 = exclusively heterosexual to 5 = exclusively gay or lesbian     Comparison of lesbians, gay men, and heterosexuals on weight, dieting, preoccupation with weight, and exercise activity (n = 124) Buenting, 1992     Sex of preferred sexual partner Survey of health lifestyles of lesbian and heterosexual women (n = 79) Carroll et al., 1997 Self-identification as lesbian, bisexual, heterosexual, or other Sex of partner asked with respect to participation in nine specific sexual practices; years since last male sexual encounter; lifetime number of male and female partners   STD testing, diagnosis, and sexual practices among self-identified lesbian and bisexual women (n = 421) Chicago Women's Health Study (Hughes et al., 1997) Self-identification: "only heterosexual, mostly heterosexual, bisexual, mostly homosexual, or only homosexual" Sexual behavior in the past year: "only men, mostly men, equally men and women, mostly women, only women" Current sexual interest or attraction: "only men, mostly men, equally men and women, mostly women, only women" Indicators of mental health in lesbians (n = 284) and a comparison group of heterosexual women (n = 134)

OCR for page 97
-->

OCR for page 97
-->   Measures Used to Assess Sexual Orientation Study Identity Behavior Desire Focus of Study (sample size) Chu et al., 1990   ''After 1977 and preceding the diagnosis of AIDS, did this patient have sexual relations with a male partner?"   Assess demographic characteristics and behavioral risk factors in lesbians in national surveillance data for reported cases of AIDS (n = 79)     "After 1977 and preceding the diagnosis of AIDS, did this patient have sexual relations with a female partner?" If only female partners, then classified as lesbian; if both male and female partners, then bisexual     Deren et al., 1996 "Do you consider yourself to be heterosexual [straight], lesbian, bisexual, other?" Sex of sexual partner(s) during past 30 days (men only, women only, women and men)   Relationship of sexual orientation to HIV risk behavior and serostatus in a multisite sample of drug-injecting and crack-using women (n = 830 lesbians; n = 5,791 heterosexual women) Einhorn and Polgar, 1994 Self-defined as lesbians, bisexual, or undefined Sex of sexual partners since 1978   HIV risk behavior among lesbians and bisexual women (n = 1,086) Turner et al., 1998b   Whether engaged in woman-to-woman sex during period between consecutive interviews or blood samples (approximately 6 months)   Assess HIV status in male and female out-of-treatment drug injectors to determine risk factors for HIV seroconversion in high- and low-seroprevalence cities (n = 6,882) Gómez et al., 1996 "What do you consider your sexual orientation to be?" "How many women, in total, have you had sex with in the past 3 years?"   Risk for HIV and other STDs in order to develop relevant prevention strategies (n= 461)     "In the past 3 years, how many men have you had sex with?"    

OCR for page 97
-->

OCR for page 97
-->   Measures Used to Assess Sexual Orientation Study Identity Behavior Desire Focus of Study (sample size) Johnson et al., 1987 Categorize self as lesbian, bisexual, or heterosexual Lifetime history of heterosexual intercourse (0, 1-9, 10-100, or>100 episodes)   Reproductive system problems of lesbian and bisexual women in a nonclinical setting (n = 1,921 lesbians; n = 424 bisexual women) Krieger and Sidney, 1997 "Have you experienced discrimination, been prevented from doing something, or been hassled or made to feel inferior in any of the following situations because of your sexual preference (heterosexual, bisexual, homosexual)? (yes or no) a. In your family; b. at school; c. getting a job; d. at work; e. at home" Lifetime number of same-and other-sex sexual partners   Prevalence of self-reported experiences of discrimination based on sexual orientation among black and white adults with same-sex sexual partners, and health-related consequences of discrimination (n =1,724; n = 1,031 women; n = 114 lesbian or bisexual women) Lemp et al., 1995 Self-reported lesbian or bisexual identity Sex of sexual partners since 1978 (men only, men and women, women only); engagement in specific sexual behaviors (e.g., oral sex with women)   HIV seroprevalence and risk behaviors among lesbians and bisexual women in San Francisco and Berkeley (n = 550) Michigan Lesbian Health Survey (Bybee and Roeder, 1990) 7-point scale: "Circle the number below that best describes how you think of yourself" Lesbian = 1; 4 = bisexual; heterosexual only = 7 Whether ever had sexual contact with a woman; number of female sexual partners in the past year; whether ever had sex with a man; time since most recent sexual contact with a man; age at first sexual experience with a woman.   General health and HIV risk of lesbians across the state of Michigan (n = 1,681)

OCR for page 97
-->

OCR for page 97
--> drawing the sample or by weighting responses in the analyses to adjust for the chances that names on one list are duplicated on another. Multiplicity or Network Sampling The goal of multiplicity or network sampling is to reduce the amount of screening necessary to identify members of the population subgroup by capitalizing on linkages to relatives or neighbors. Its most likely use is to estimate the size of a hard-to-identify population subgroup. 7 With this strategy, an interviewed subject is asked to identify others (e.g., relatives or neighbors) who are members of the population subgroup and, in some cases, to provide data on them. This method requires that respondents be able and willing to report whether people linked to them are members of the population subgroup. Further, if the characteristics of the population subgroup are being studied, the respondents must be able to report on these characteristics accurately, or the individuals they identify must be contacted in person and interviewed. If multiple routes of sample selection are used, responses must be weighted in the analyses to compensate for differential probabilities of selection. There are clearly both ethical and logistical problems in using this technique to estimate the number of lesbians or to obtain samples of lesbians. First, lesbians' sexual orientation may not be known to all others in their networks and so they would not be identified. Second, even if a lesbian's sexual orientation is known to the respondent, she may not want the respondent to disclose this information to others. 7   Killworth et al. (1998) describe the use of a social network approach to estimate the size of hard-to-count populations. In their study, a representative sample of the U.S. population was asked how many people they knew in 29 populations of known size, and how many people they knew in 3 populations of unknown size whose size the researchers wanted to estimate (people who are homeless, women who were raped in the past 12 months, and people who are HIV-positive). Responses were then used to calculate a maximum likelihood estimate of the number of people in the respondent's social network, and the patterns of all respondents' responses about the populations of unknown size were used to estimate the size of the unknown subpopulations. The method accurately estimated the size of 20 of the 29 known populations, and the calculated estimates of the unknown populations were consistent with other published estimates. To the committee's knowledge, this approach has not yet been used to estimate the size of the lesbian population, but it is a strategy that might yield useful information.

OCR for page 97
--> Snowball Sampling In snowball sampling, members of a population subgroup are identified and asked to identify or report on other members of the population subgroup. The chances of being identified thus depend to some extent on the level of involvement in various social and community networks. Once a list or frame is obtained that contains all or a large proportion of the population subgroup, it is possible to draw a probability sample from the list. It is important that the snowball technique not simply be stopped when a desired sample size is reached. Rather, snowball sampling should be continued until a full roster of names believed to contain a large percentage of the population of interest is developed; then individuals from this list can be randomly sampled into the study. Although snowballing can be useful for constructing a sampling frame of all members of some population subgroups, given the hidden nature of the lesbian population and the potential risks of disclosure, it is not likely to be a feasible strategy for use with this population. Location (Convenience) Sampling In location sampling, sometimes referred to as convenience sampling, members of a rare population are sampled in places where they congregate, for example, gay bars or bookstores, where there is a high expectation of sufficient lesbian density to make the sampling efficient. Unfortunately, this strategy, too, results in significant sampling bias because bars are more likely to be frequented by young women who are old enough to drink and by women who drink more frequently, just as lesbian bookstores might be visited more frequently by feminists, and so forth. One suggested strategy for improving location sampling is first to map all of the possible sites and then to draw samples from within these sites in ways that cover multiple time periods and geographic areas. There are, nevertheless, significant problems with this methodology. First, only people who go to these particular sites (e.g., bars or bookstores) will be sampled. Second, people who go to these sites more frequently have a greater probability of being selected into the sample. Thus, the location

OCR for page 97
--> sampling method produces a probability sample of visits rather than visitors. Combining Strategies for Sampling Lesbians Many lesbian health research studies have used nonprobability sampling that goes far beyond convenience sampling, employing multiple methods in an attempt to obtain more broadly representative samples in a geographic area (e.g., sampling from newsletter rosters, women in gay and lesbian bookstores, lesbian community organizations, attendees at women's music festivals). Each of these methods casts the net wider and wider, attempting to sample participants from outside the social networks of the researcher or from outside institutional settings such as clinics and hospitals. Nonetheless, no matter how wide the net becomes, these methods are all still non-probability-based sample designs and are subject to criticism for selection bias even if they represent the most sophisticated methods now used in lesbian research. Given the difficulties in conducting population-based probability sampling of lesbians and the cost-efficiency of non-probability-based designs, an important research question is whether techniques can be employed to improve the usefulness of non-probability-based samples. One strategy might be to examine the consistency of findings across studies that use different sampling methods across different settings, with the expectation that results might center around the true population parameter. Confidence in reported estimates can increase as replicated well-designed studies are published on lesbian health using similar variables. Adding Sexual Orientation Questions to Existing Studies One suggestion for increasing the amount of information on lesbian health is to add sexual orientation items to large sample surveys or other large studies. This would add a significant amount of analytical capability at little additional cost. Adding appropriate questions about sexual orientation to large cross-sectional or cohort studies strengthens the ability to understand variation in health status by sexual orientation. However, in-

OCR for page 97
--> vestigators have been reluctant to add such questions to their studies for fear that women would not answer them and, in addition, might be so alienated by the questions that they would not participate in the study. There is evidence, however, that sexual orientation items can be used successfully in studies and in surveys (NRC, 1989). Two examples of ongoing studies that have included sexual orientation items in their data collection efforts are the Nurses' Health Study II8,9 and the Women's Health Initiative. In September 1994, NHS-II investigators began to receive requests for information on sexual orientation and health from other researchers and, more importantly, from both lesbian and heterosexual participants in the study itself. In response to these requests, investigators decided to add a question to gather information about sexual identity, hypothesizing that the social experience of women with a lesbian or bisexual sexual orientation might put them at higher risk for a number of conditions, particularly chronic disease and cancer. Questions were first pilot-tested on a sample of 1,050 nurses for inclusion in the June 1995 questionnaire. A variety of questions were asked in order to compare the willingness of respondents to answer different types, including questions about sexual identity 10 and about the sex of sexual partners over the past five years and during one's lifetime. The researchers found that women were willing to return the form (response rate = 78.4%) and answer the questions on sexual orientation, that they considered these questions acceptable, and that there was little negative feedback.11 As a result, investigators added a question on sexual identity to the 1995 Nurses' Health Study cohort questionnaire, which was mailed to 116,000 women. These results must be interpreted with some caution because nurses 8   NHS-II is a prospective cohort of 116,680 female registered nurses established in 1989 as a companion study to the long-running Nurses' Health Study I. Participants are mailed questionnaires every two years that include comprehensive health questions. 9   The description of the NHS-II is based on the workshop presentation of Dr. Patricia Case. 10   The item categories included heterosexual; bisexual; lesbian, gay, or homosexual; none of the above; or prefer not to answer. 11   In the NHS-II pilot study, 98.0% of the respondents reported that they were heterosexual; 0.9% that they were lesbian, gay, or homosexual; and 0.1% that they were bisexual (Case et al., 1996).

OCR for page 97
--> may be more willing to respond to questions about sexuality than general population samples. Nonetheless, these results are promising and demonstrate that careful pilot testing of questions in advance can help to reveal patterns of response and ease investigators' fears, often legitimate, of negative effects on the study. The WHI,12 a longitudinal study of about 100,000 postmenopausal women, has included one question on its baseline data collection forms about gender of adult lifetime sexual partners (i.e., men, women, both men and women, or never had sex). Questions on sexual orientation could also potentially be added to large national sample surveys such as the Bureau of Labor Statistics National Longitudinal Surveys, NHIS, the NHANES, and the National Survey of Family Growth. Only very large population-based surveys such as these can be expected to yield sufficient sample sizes of lesbians without substantial changes in the existing research design. Consequently, these surveys represent an important opportunity to attempt population-based sampling of lesbians if they were to include questions of sexual orientation. The committee does not believe, however, that it is feasible at this time to add questions regarding sexual orientation to the U.S. Census of Population and Housing for a variety of reasons, including the facts that questions are often asked of one individual in a household who serves as a proxy for the others; that efforts are made to minimize the number of questions, making it difficult to have any kind of item added to the questionnaire; and that the census questions require the approval of the U.S. Congress. It should be noted that there are also potential drawbacks to the strategy of adding questions to existing surveys. For example, including questions on sexual orientation does not guarantee that they will be considered in the analyses. Alternatively, the data may be analyzed and interpreted by persons who are unfamiliar with the nuances of conducting research with and defining the lesbian population. Another potentially very significant problem is the issue of confidentiality (see Chapter 4). Because these large-scale surveys are confidential but not anonymous, the extent to which lesbians will disclose their sexual orientation in this setting is unclear. Indeed, the degree to which sexual orientation will be 12   See information on WHI in footnote 8 in Chapter 2.

OCR for page 97
--> disclosed even when anonymity is ensured is unclear for all studies that inquire about sexual orientation. Areas for Further Methodological Research in Studying Lesbian Health The committee identified several strategies for increasing the statistical power of research on lesbians. These include pooling data across studies using similar measures and methods; meta-analysis of methodologically sound studies, including those with smaller sample sizes; multisite studies; and fostering national and international collaborations. Nonetheless, the field would benefit greatly from additional research focused on improving research methodologies. This type of research also has significant potential to benefit the conduct of health-related research with other populations, especially other hard-to-identify and/or rare population subgroups. The committee suggests the following areas for additional research: Research is needed on how techniques for sampling other hard-to-identify and/or rare population subgroups might be applied to obtaining probability samples of lesbians. Research is recommended on the validity and reliability of questions measuring the different dimensions of sexual orientation: identity, behavior, and attraction or desire. Research is needed on technologies to elicit disclosure in order to determine their usefulness in studies with lesbians. It is particularly important to consider how factors such as race and ethnicity, socioeconomic status, age, and region of residence affect the use of such technologies. Qualitative research is needed to increase the depth of understanding of lesbians' lives and to better inform other research. Consideration should be given in larger studies to including experts in qualitative methods in the design team for an integrated approach to seeking information. Existing databases from federal and other large surveys should be reviewed to identify those with potential for analysis by sexual orientation. In the absence of the ability to conduct targeted population-based

OCR for page 97
--> studies of lesbians, efforts to create more effective non-probability-based designs should be undertaken. Methodological research is needed on the feasibility of using matched control designs for research directed toward assessing the health status of lesbians compared to other women. There may be innovative ways of sampling that have not been widely used. One suggestion made at the workshop (Gruskin, 1997) is to work with managed care organizations, many of which have research divisions, to obtain samples of lesbians and nonlesbian matched controls. The membership of a managed care organization usually is quite large and captured in a medical care delivery system, and access to information-rich medical records may be possible. Methodological research should be done to investigate the utility of this approach and to examine the ethical issues that it raises, with careful attention to how to maintain confidentiality. References Berger BJ, Kolton S, Zenilman JM, Cummings MC, Feldman J, McCormack WM. 1995. Bacterial vaginosis in lesbians: A sexually transmitted disease. Clinical Infectious Diseases 21(6):1402-1405. Bevier PJ, Chaisson MA, Hefferman RT, Castro KG. 1995. Women at a sexually transmitted disease clinic who reported same-sex contact: Their HIV seroprevalence and risk behaviors. American Journal of Public Health 85(10):1366-1371. Binson D, Moskowitz TM, Anderson K, Paul J, Stall R, Catania J. 1996. Sampling men who have sex with men: Strategies for a telephone survey in urban areas in the United States. Proceedings of the Section on Survey Research Methods, American Statistical Association. Alexandria, VA: American Statistical Association, pp. 68-72. Bloomfield K. 1993. A comparison of alcohol consumption between lesbians and heterosexual women in an urban population. Drug and Alcohol Dependence 33:257-269. Bradburn NM, Sudman S, Blair E, Locander W, Miles C, Singer E, Stocking C. 1979. Improving Interview Method and Questionnaire Design. San Francisco: Jossey-Bass. Bradford J, Ryan C. 1988. The National Lesbian Health Care Survey: Final Report. Washington, DC: National Lesbian and Gay Health Foundation. Brand PA, Rothblum ED, Solomon LJ. 1992. A comparison of lesbians, gay men, and heterosexuals on weight and restrained eating. International Journal of Eating Disorders 11(3):253-259. Buenting JA. 1992. Health life-styles of lesbian and heterosexual women. Special Issue: Lesbian health: What are the issues? Health Care for Women International 13(2):165-171. Bybee D, Roeder V. 1990. Michigan Lesbian Health Survey: Results Relevant to AIDS. A Report to the Michigan Organization for Human Rights and the Michigan Department of Public Health. Lansing: Michigan Department of Health and Human Services.

OCR for page 97
--> Carroll N, Goldstein RS, Lo W, Mayer KH. 1997. Gynecological infections and sexual practices of Massachusetts lesbian and bisexual women. Journal of the Gay and Lesbian Medical Association 1 (1):15-23. Case P, Spiegelman D, Hunter DJ, Manson JE, Willet WC. 1996. Sexual Orientation in Relation to Behaviors in the Nurses' Health Study II: Selected Results from a Pilot Study. Presentation to the American Public Health Association. New York: National Development and Research Institutes. Chu SY, Buehler JW, Fleming PL, Berkelman RL. 1990. Epidemiology of reported cases of AIDS in lesbians, United States 1980-89. American Journal of Public Health 80(11):1380-1381. Cochran SD, Mays VM. 1988. Disclosure of sexual preference to physicians by black lesbian and bisexual women. Western Journal of Medicine 149(5):616-619. Cochran SD, Bybee D, Gage S, Mays VM. 1996. Prevalence of HIV-related self-reported sexual behaviors, sexually transmitted diseases, and problems with drugs and alcohol in 3 large surveys of lesbian and bisexual women: A look into a segment of the community. Women's Health Research on Gender, Behavior, and Policy 2(1-2):11-33. Deren S, Goldstein M, Williams M, Stark M, Estrada A, Friedman SR, Young RM. 1996. Sexual orientation, HIV risk behavior, and serostatus in a multisite sample of drug-injecting and crack-using women. Women's Health: Research on Gender, Behavior, and Policy 2(1-2):35-47. Einhorn L, Polgar M. 1994. HIV-risk behavior among lesbians and bisexual women. AIDS Education and Prevention 6(6):514-523. Gómez CA, Garcia DR, Kegebein VJ, Shade SB, Hernandez SR. 1996. Sexual identity versus sexual behavior: Implications for HIV prevention strategies for women who have sex with women. Women's Health: Research on Gender, Behavior, and Policy 2(1-2):91-109. Greene B. 1994. Ethnic-minority lesbians and gay men: Mental health and treatment issues. Journal of Consulting and Clinical Psychology 62(2):243-251. Gruskin E. 1997. Presentation before the Institute of Medicine Committee on Lesbian Health Research Priorities. Washington, DC, October 6-7, 1997. Hendershot TP, Rogers SM, Thornberry JP, Miller HG, Turner CF. 1996. Multilingual audio-CASI: Using English-speaking field interviewers to survey elderly Korean households. In: Warnecke R, ed. Health Survey Research Methods. Hyattsville, MD: National Center for Health Statistics. Hughes TL, Haas AP, Avery L. 1997. Mental health concerns of lesbians: Preliminary results from the Chicago Women's Health Survey. Journal of the Gay and Lesbian Medical Association 1(3): 137-148. Hughes TL, Wilsnack SC. 1997. Use of alcohol among lesbians: Research and clinical implications. American Journal of Orthopsychiatry 67(1):20-36. Johnson SR, Smith EM, Guenther SM. 1987. Comparison of gynecologic health care problems between lesbians and bisexual women. A survey of 2,345 women. Journal of Reproductive Medicine 32(11):805-811. Kalton G. 1993. Sampling considerations in research on HIV risk and illness. In: Ostrow DG, Kessler RC, eds. Methodological Issues in AIDS Behavioral Research. New York: Plenum Press. Pp. 53-74.

OCR for page 97
--> Killworth PD, McCarty C, Bernard HR, Shelley GA, Johnsen EC. 1998. Estimation of seroprevalence, rape, and homelessness in the United States using a social network approach. Evaluation Review 22(2):189-308. Krieger N, Sidney S. 1997. Prevalence and health implications of anti-gay discrimination: A study of black and white women and men in the CARDIA Cohort. International Journal of Health Services 27(1):157-176. Laumann EO, Gagnon JH, Michael RT, Michaels S. 1994. The Social Organization of Sexuality: Sexual Practices in the United States. Chicago: University of Chicago Press. Lemp GF, Jones M, Kellogg TA, Nieri GN, et al. 1995. HIV seroprevalence and risk behaviors among lesbians and bisexual women in San Francisco and Berkeley, California. American Journal of Public Health 85(11):1549-1552. Miller HG, Gribble JN, Rogers SM, Turner CF. In press. Abortion and breast cancer risk: Fact or artifact? In: Stone A, ed. Science of Self Report (provisional title). Mahwah, NJ: Lawrence Erlbaum Associates. Mosher W, Duffer A. 1995. Innovations in the 1995 National Survey of Family Growth. Paper presented at the annual meeting of the Population Survey of America, San Francisco, CA. Norman AD, Perry MJ, Stevenson LY, Kelly JA, Roffman RA. 1996. Lesbian and bisexual women in small cities—At risk for HIV? HIV Prevention Community Collaborative. Public Health Reports 111(4):347-352. NRC (National Research Council). 1989. AIDS: Sexual Behavior and Intravenous Drug Use. Washington, DC: National Academy Press. NRC. 1990. AIDS: The Second Decade. Washington, DC: National Academy Press. NRC. 1997. Evaluating Genetic Diversity. Washington, DC: National Academy Press. Rankow EJ. 1995. Lesbian health issues for the primary care provider. Journal of Family Practice 40(5):486-496. Robertson P, Schachter J. 1981. Failure to identify venereal disease in a lesbian population. Sexually Transmitted Diseases 8(2):75-76. Rossi PH, Wright JD, Fisher GA, Willis G. 1987. The urban homeless: Estimating composition and size. Science 235:1336-1341. Russell JM, Azadian BS, Roberts AP, Talboys CA. 1995. Pharyngeal flora in a sexually active population. International Journal of STD and AIDS 6(3):211-215. Sell RL, Petrulio C. 1996. Sampling homosexuals, bisexuals, gays, and lesbians for public health research: A review of the literature from 1990 to 1992. Journal of Homosexuality 30(4):31-47. Shaffer D, Fisher P, Hicks RH, Parides M, Gould M. 1995. Sexual orientation in adolescents who commit suicide. Suicide and Life-Threatening Behavior 25(Suppl):64-71. Smith EM, Johnson SR, Guenther SM. 1985. Health care attitudes and experiences during gynecologic care among lesbians and bisexuals. American Journal of Public Health 75(9):1085-1087. Trevathan WR, Burleson MH, Gregory WL. 1993. No evidence for menstrual synchrony in lesbian couples. Psychoneuroendocrinology 18(5-6):425-435. Trippet SE. 1994. Lesbians' mental health concerns. Health Care for Women International 15(4):317-323. Turner CF. 1997. Presentation before the Institute of Medicine Committee on Lesbian Health Research Priorities. Washington, DC, October 6-7, 1997 .

OCR for page 97
--> Turner CF, Forsyth BH, O'Reilly J, Cooley PC, Smith TK, Rogers SM, Miller HG. 1998a. Automated self-interviewing and the survey measurement of sensitive behaviors. In: Couper M, et al., eds. Computer-Assisted Survey Information Collection. New York: Wiley and Sons. Turner CF, Ku L, Rogers SM, Lindberg LD, Pleck JH, Sonenstein FL. 1998b. Adolescent sexual behavior, drug use, and violence: Increased reporting with computer survey technology. Science 280:867-873. Turner CF, Lessler JT, Devore J. 1992. Effects of mode administration and wording on reporting of drug use. In: Turner CF, Lessler JT, Gfroerer JD, eds. Survey Measurement of Drug Use: Methodological Issues. DHHS Pub. No. 92-1929. Washington, DC: U.S. Government Printing Office. Women's Health Initiative Study Group. 1998. Design of the Women's Health Initiative Clinical Thai and Observational Study. Controlled Clinical Trials 19:61-109.

OCR for page 97
This page in the original is blank.