A: Conducting research with hard-to-reach or "hidden" populations is difficult, and that difficulty increases when dealing with stigmatized hidden populations, such as gays and lesbians. Changes in societal attitudes towards gays and lesbians have made it easier to reach some of them, but it is still hard to reach those located in rural areas and areas without established gay communities.

Advances in technology such as increasing internet access and improved networking tools can provide researchers with more opportunities to locate and recruit participants from these kinds of populations. For example, one of our studies with gay adoptive fathers included recruitment from web-based communities designed for gay and lesbian parents as well as the development of a private, voluntary web-based group to encourage communication among participants.

A: One advantage is the relative ease of getting the message out about the research. We posted advertisements on listservs and in the "classified" or "public notice" section of appropriate web-based communities. Also, we searched through available profiles in these communities and sent targeted invitation letters, which was highly successful for us.

A potential disadvantage in recruiting through an established community is that the researcher will be viewed as an outsider and not trusted by the community. In our case, one of us had a profile on the main web-based community that was used for recruitment, and that profile truthfully presented the researcher as a community member with similar attributes to other community members. That enhanced the credibility and trustworthiness of the researcher.

You may also have to field questions about whether or not you're able to get a representative sample, but this wasn't an issue for us. While some folks will attempt quantitative studies with hidden populations, the majority will more likely seek to complete qualitative projects or cross-sectional studies using surveys, etc.

Also, the reason one wants a representative sample is in hopes of generalizing findings to a larger population, and this is nearly impossible with research with hidden populations as we don't have good ideas of the parameters of the population such as size, demographics, dispersion, etc. For example, how many GLBT folks are in the US? You can guess, but you really have no true idea. From the beginning, someone attempting quantitative studies with a convenience sample is at a disadvantage as critics will argue that you cannot generalize the findings outside of the sample from which its results are drawn.

A: One possibility is using videoconferencing tools. This can be done for free using Windows Messenger and does not require expensive software or equipment. We've found that videoconferencing in this way is user-friendly. Also, you can use Microsoft Office speech recognition to simultaneously produce a rough transcription of the interviews. The quality varies on how clearly the individuals speak, cadence, pitch, etc. Also, as voice recognition software is "trained" to understand the primary speaker, it only follows that that the computer will pick up more (and get it more accurate) of what the researcher says.

The other method of preserving interviews is through voice recording software. This can be a way for the researcher to preserve the interview in digital format as a sound file, which they can then transcribe later if it is so desired. Also, don't simply assume that it's good to record interviews rather than use expanded field notes, because that is a debatable assumption (Rodwell, 1998; Denzin & Lincoln, 1998). Many researchers feel that field notes that are immediately expanded post-interview work just as well as transcriptions.

A: You should consider issues of accessibility. Accessibility can pose a problem both for people with disabilities, and for people living in geographically isolated areas. Become familiar with the Web Content Accessibility Guidelines. In addition, people who access the internet in public locations such as a library or internet cafÈ, may have concerns about sensitive content in situations where others can easily see the computer screen, or they may worry about sharing information through public terminals not under their control.

In our study with gay adoptive fathers, the only concerns about privacy centered on the standard concerns of participants that they would not be identified in the research report. This was especially pronounced in cases where individuals subverted state laws such as a participant discussing how he adopted even though he lives in state where adoption by a gay person is illegal. Standard assurances of confidentiality served well in this case.

If you are considering web-based groups as a possible method for data collection, then you need to think carefully about your role in the group. Will you be an active voice in the discussions, or will you approve membership only without participating yourself? In terms of confidentiality, you can think of a web-based group as a type of focus group, and while you can assure your own adherence to confidentiality and encourage the participants to do the same, you can't make promises about each participant's behavior.