Herbert I. Schiller Room (MCC 201)
Thursday, October 9, 12-2 pm

An informal workshop to share notes, experiences and ideas regarding the use of various tools / methods / technologies in support of qualitative social research and writing.

Here are some notes and links coming out of the discussion...

Why the workshop? ~

This is an aspect of social scientific research that is rarely squarely addressed, but constitutes a signficant chunk of what many of us actually do on a day-to-day basis. This needn't be seen as a dry, boring question left to the techies (beneath the dignity of us serious intellectuals, fixated on the uber-tool of library and book:)) - ideally, sustained and creative thinking about tools can be a way of opening up one's research in interesting and productive ways. There are also new hardware and software options for the support of qualitative research that may or may not be useful, depending on one's individual research needs, but at the very least are worth exploring and thinking about. The purpose of the workshop and the page is to get people thinking cooperatively along these lines, and to share experiences and tips arising from people's own research experiences. It may be particularly valuable for those at the outset of research, or an academic career; because tool and data decisions (conscious or otherwise) have a way of getting locked in as information systems grow, making smart up front decisions, taking some time early on to think about a suite of tools that's appropriate to your particular research area, style and needs is likely to pay off down the road.

This is intended as a document in progress - additions, revisions, suggestions, etc. are most welcome -

Recording tools ~

Audio cassette: this worked well for many; variable speed models allow one to play back at a slower speed, making transcription easier. Some limitations with sound quality. Storage may also be bulky.

Microcassette: some have used this in past, though difficulties were also reported in replaying tapes after recording. The thinness of the tape may lead to stretching, raising long-term sound quality and durability issues.

MiniDisc Recorders: these are nice for sound quality and storage, BUT, contrary to what you might expect, simple transfer from disc to computer is currently difficult (basically requires real-time uploading) -- ** this is a significant limitation - check carefully on this before buying **

Digital Voice Recorders: these produce comparable sound quality to standard audio recorders, but have the advantage of being easily and directly uploadable to your hard drive through a USB port (unlike current MiniDiscs). Most record in proprietary file formats, but are easily converted, through a Save As feature, to standard audio formats (.wav, .mp3, etc.)Typical models can record for around 3 hours before needing to be dumped to a computer - some come with removeable memory sticks that can extend that time (in general, price varies directly with memory capacity for DVRs - the low end models ($50) will typically have around 3 hours of storage), while high-end models ($200) will have much larger storage capacities, memory sticks, etc. The memory restriction makes DVR's less useful for extended research trips which require you to do many hours of interview while away from your computer; they work particularly well with laptop systems, where you can bring the computer with you and dump as you go. Starting out digital has some strong advantages: easy non-bulky storage, ease of back-up (one CD will hold hundreds of interview hours), ease of movement into other digital formats (e.g. Powerpoint presentations, web-based file-sharing (a useful collaborative tool?), etc.).

DAT (Digital Audio Tape): This is the top of the line option, producing better sound quality than any of the options listed above - certainly the best option for recording conferences, or large, noisy, high-end, and multi-voice settings (none of the options above are terribly good in such situations, though an external powered mike, ~$60, may help considerably). Quality costs, though: $800-2000.

Depending on your research needs, digital cameras and digital video may also be useful tools. In some cases, linking digital pics of research sites, settings, artefacts, etc. may be a useful complement to your audio files, or simply a useful heuristic for capturing and reminding yourself of bits of 'texture' that wouldn't come through in sound alone. Similarly but more so for digital video, which can be an important resource for some kinds of ethnographic and conversation analysis work. Digital cameras run anywhere in the $100-1000 price neighborhood, and digital video cameras from $400 to as much as you want to pay. [Apart from research needs, you should also consider the sociotechnical dynamics of the interview situation - bulky and instrusive technologies can change the interview dynamic considerably, with effects that may vary wildly depending on who you're interacting with and how you introduce the technology...]

At the most mundane level, telephone interviews are easily recorded by a phone-to-mike wire (telephone jack on one end, microphone jack on the other) that should be available at any electronics store for about $10. If you want to record and talk out of the same wall jack (i.e. not have to plug the recorder into a separate phone jack), you can pick up a plastic jack splitter for about $5, again at just about any electronic/hardware store. This would work for anything with a standard headphone-sized 'mic in' jack, which means just about anything described above, analog or digital.

Transcription tools ~

Voice recognition software: the two main brands currently on the market are Dragon and ViaVoice. Unfortunately, the ethnographer's dream of fully automatic transcription seems some distance away. Voice recognition is coming along, and most reports now give them about 90-95% accuracy, under carefully controlled conditions. Their biggest limitation for interview / ethnographic purposes is that they require training to a single voice, meaning you can train the software to recognize your voice, but it won't be able to do much with anyone else you're speaking with. (Current versions under development are configurable for multiple user profiles, kind of like the Windows XP interface, but this doesn't really solve the problem.) One workaround that was mentioned in the workshop is to essentially redictate your interviews - listen to the interview through headphones while repeating them into the computer. In any case, most transcripts produced through voice-transcription will require some clean-up work to catch the (frequently bizarre) errors.

Transcription software: This is a nice option for people working in digital, and does away with the old strategy of window swapping (i.e. digital voice software opened in a window on top, word processor in a window on bottom, and flipping between the two). There's a nice bit of free software available for PC users called ExpressScribe, which allows you to listen and transcribe all in the same program. Has some nice side features too: variable speed control (you can set the file to play at whatever speed your fingers can keep up with), automatic time insertion, forward / back keys that can be set at whatever intervals you find most useful. Really a useful tool. Similar products for Mac users?

Transcription hardware: If you're transcribing from audio tape, you really, really need a foot pedal to stop, start and control speed. The department has one available for use. Digital transcription pedals are also available which do essentially the same thing. These run anywhere from $60-200, depending on whether you buy a proprietary (i.e. designed to work with one company's standard) or a universal (i.e. works with all digital file types) model.

Data management / analysis tools ~

There are a number of software packages on the market that bill themselves as useful tools in thinking through and managing your qualitative data. In general, these allow you to search, code, and link between various sorts of text-based data (they can also link to external files, such as digital picture and video files, though they don't really have any 'in-house' tools for working on these yet). This might not mean much to those in the early stages of research or dealing with relatively small amounts of data, but for those sifting and sorting large quantities of qualitative data (field notes, interview transcripts, open-ended survey tools, etc.) these may prove useful alternative to cutting-and-pasting (the old-fashioned way) and 3 am pile sifting sessions. Programs of this type commonly in use include NVivo (latest product in the Nudist family line), AtlasTI (built explicitly around the principles of grounded theory), HyperResearch, and the Ethnograph, though others exist. Prices on these run in the $200-500 range for an educational license. Each of the four mentioned above have free limited-functionality demos available for download - if you're interested, play around with them a bit and see what feels comfortable before you buy.

For the quantitatively-oriented, SPSS (Statistical Program for the Social Sciences - used by 99.9% of quantitative social scientists [margin of error of +/- .1%]) is now mercifully available in a user-friendly interface.

For bibliographic software, EndNote would seem to be the academic standard. This is a very useful tool for building up your own library of references and notes, and automatically generates bibliographies, footnotes, etc. according to whatever format is required. Like many of these tools, the value of a good bibliographic reference set grows with time - the initial investment of money and effort may well be worth it. Student prices run around $100, or the department may have some older versions available.

Design principles ~

Some more general issues to consider when making tool decisions:

Interoperability - how well do the various aspects of the larger research system fit together? Are there strong barriers which might impede one from moving between tools (e.g. from recording, to transcription, to analysis)? What are the import / export capabilities of these systems?

Scale - is the tool/system in question appropriate for multiple data scales? Is a system that works well for small data sets (e.g. re-reading notes, 'just remembering') going to work as you dataset grows? Can 'big' data be brought down to the scale of local manipulation?

Learning curve - how much time will you invest to learn the tool/system, and is it worth it? (bear in mind there are both immediate and longer-term considerations here)

Robustness / durability - how solid is your system? Is it crash-prone (software that dies, tapes that stretch)? Is it easily backed up to protect against this?

Sustainability - when archaeologists from some future civilization uncover your data, will they be able to access it? Closer to home, will the tools by which you currently code and access your data be reliably available during the potential lifetime of its usefulness? Tapes and CDs deteriorate, laser-printed paper fades, software and hardware specs come and go - how standard is your standard? What provisions might you make for 'migrating' data as systems change?

Data protection - in cases of sensitive information, how secure is your data? What happens if you leave your backpack on the bus? What happens if someone swipes your laptop? (IRB guidelines currently stipulate that ALL subjects must be informed if the integrity of data is compromised...)

Collaboration - is collaboration a value you want your data system to support? With other researchers? With the groups you're studying?

Price - this is obviously a factor - what is the tool/system worth to you, and how much can you afford to spend? (These are hard constraints - but when making cost decisions, bear in mind that if you're thinking about building a data system over the long haul, your monetary investment in the system will eventually pale in comparison to your time investment - in this sense, going with a lesser system likely to present future obstacles in order to save a few bucks may be counterproductive in the long-run…)

Other links ~

Qualitative Data Analysis research page (comprehensive set of links to various QDA software resources, some of them obscure):

A good page with links to a variety of qualitative and quantitative research packages:

The Content Analysis Resource page:

A slightly dated (1998) comparative academic review of AtlasTI and Nudist software:

An article by Alan Stockdale, "Tools for Digital Audio Recording in Qualitative Research":

Other suggestions?