I agree that 'reliability' is a complex issue for qualitative researchers,
and that exact matching of 2 people's codes does not confer reliability or
validity. But, as you say, the *process* of comparing two people's codes,
and then discussing disagreements, is very fruitful and can help the
resaercher to tighten their code definitions.
I shall be asking my PhD supervisor to code some of my interviews, precisely
so that we can do a comparison and discuss differences. As I see it, this
is one of the only ways I can demonstrate some semblance of (1) consistency
in my own efforts at interpretive coding (because disagreements may be
because I myself have been inconsistent), and (2) content validity, by
comparing my coding against the codes of a subject expert with the same
epistemological orintation as I have.
I'd be very interested in what others think. I know that I'll be skewered
at my viva if I can't demonstrate some efforts towards checking consistency
of my judgment, or content validity.
Karen
-----Original Message-----
From: Kath Checkland
To: [log in to unmask]
Sent: 5/25/01 9:53 AM
Subject: Re: Coding in Nvivo
Interesting question, this, and it suggests to me that you haven't fully
thought through your ontology/epistemology. What does "reliability" of
coding mean in a qualitative context? Surely, given the fact that we are
dealing with subjective interpretations of what people say (which may
itself
be very different from what they think...) there can be no "right"
answer as
far as coding goes. I would have thought that some of the most
interesting
data comes out of those instances where two different coders disagree -
simply asking the question why? and looking to see what each person is
seeing in the text will broaden and deepen the understanding of the
data.
What does anyone else think?.....
Kath Checkland
----- Original Message -----
From: "Christine J. McPherson" <[log in to unmask]>
To: <[log in to unmask]>
Sent: 23 May 2001 07:57
Subject: Coding in Nvivo
> Is it possible for two coders to independantly code the same text and
to
> check
> the reliability of the codings in Nvivo?
>
> Thanks
> Chris
|