home *** CD-ROM | disk | FTP | other *** search
- Comments: Gated by NETNEWS@AUVM.AMERICAN.EDU
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!howland.reston.ans.net!paladin.american.edu!auvm!MSU.BITNET!21602JMC
- Message-ID: <QUALRS-L%92122115184250@UGA.CC.UGA.EDU>
- Newsgroups: bit.listserv.qualrs-l
- Date: Mon, 21 Dec 1992 15:13:00 EST
- Sender: Qualitative Research for the Human Sciences <QUALRS-L@UGA.BITNET>
- From: "Josie.Csete" <21602JMC@MSU.BITNET>
- Subject: Advice on reliability
- In-Reply-To: The letter of Sunday, 20 December 1992 1:48am ET
- Lines: 27
-
- I have also been struggling with the issue of subjectivity in qualitative data
- coding and analysis. Todd mentioned triangulation is not possible because he
- doesn't have access to his participants again, but I would suggest other ways
- oftriangulating. First, (perhaps this is not possible), but my participants
- responded to an instrument and in my study I'm going to"guess"about their
- responses before I look at them, a sort of measure of how well I I know my
- participants. But there are also some forms of triangulation that you can do
- without ever seeing your participants again. For instance, have you
- consideredintra-rater reliability? I'm new at this stuff, so I'm going to
- code things, leave it for a while, and then code fresh clean copies again to
- see if I code the same stuff the same way over time. Second, I plan to try a
- bit of inter-rater reliability. After I'm happy with my coding(If ever) I plan
- to train a colleague on my categories with examples, and then give him a
- portion of my data to code, to see how we agree.
-
- I'd be very open to hearing what others think about these methods. Messages
- have been flying about "prostituting ourselves", and granted I'm new at
- qualitativeresearch, but I think finding out whether and to what degree people
- can see what I do in the data has important implicationsfor the usefullness of
- my findings.
-
- However, I'd hasten to add, that all these attempts at "triangulation" are
- ways of checking myself. I strongly support what Kelly said, in terms of
- providing some of the data as well as how you coded, and let your readers make
- their own determination as to validity and importance.
-
- Josie
-