Krippendorff, K. (2004, October 23). CONTENT [Electronic discussion list].
From: Klaus Krippendorff <kkrippendorff@ASC.UPENN.EDU>
To: CONTENT@BAMA.UA.EDU
Subject: Re: [CONTENT] Software for Reliability
Date sent: Sat, 23 Oct 2004 00:30:29 -0400
bob and all those who have been
asking similar questions:
yes, the skeleton program for computing alpha that i have so far
will eventually have various unique features. i am saying
"unique" with some confidence as the extension of the well known
formula for alpha (for many coders, any metric, missing data,
etc.) required quite extensive mathematical work, completed a
couple of years ago, for example, for the computation of:
* confidence intervals
* the probability of failure to achieve a required level of reliability
* the reliability for individual coders
* the accuracy of coders (relative to
a known standard)
as distinct from reliability
* the reliability for partitions of
units of analysis, comparing different
classes of units with each other
* the reliability of conditional
variables, variables that
depend on what happened on another variable
* the reliability for situations in
which one unit of analysis is
assigned to multiple values, not one
* the reliability for unitizing, for
example of textual matter.
and more.
i do not know of any other agreement coefficient that is capable of answering
these
questions.
normand peladeaui of provalis research, known as the creator and
owner of wordstat, is programming the software. progress is slow
but normand has written more complex content analysis software in
the past, as we all know, and he intends to bring the project to
fruition soon.
unfortunately, i cannot give you a definite date of completion.
if you continue to be interested, ask me in a few months for a
progress report. thanks for your continued interest.
klaus krippendorff
-----Original Message-----
From: Bob Green [mailto:bgreen@DYSON.BRISNET.ORG.AU]
Sent: Friday, October 22, 2004 4:56 PM
To: CONTENT@BAMA.UA.EDU
Subject: Re: [CONTENT] Software for Reliability
In a recent e-mail Klaus Krippendorff referred to a skeleton
program that researchers have used for the purpose of computing
content analysis reliability.
I would be interested to know if this program has any additional
features to other programs that are available and an indication
of when the program might be available.
Any assisatnce with the above is appreciated,
regards
Bob green
At 03:11 PM 13/10/2004 -0400, you wrote:
>** i am working on one and have a skeleton version that i do
>not want to
>make publicly available. but i have helped several researchers
>to compute alpha for their data.
>** PRAM is working on a very basic version. i worked with the
programmer
>of that software, but it is not yet implemented as far as i
>know. ** richard craggs of the u.k. is working with me on a
>version that calculates alpha for assigning units to multiple
>categories. its special
>case is the basic alpha for assigning units to single categories
>of a nominal variable. it is almost done. i don't want to
>publish his email
>address for fear he would be bombarded by requests.
>** kristine n braaten found "a sas macro for calculating
>intercoder agreement in content analysis" published in the
>journal of advertising (1993). it calculates several indices,
>including the basic alpha. klaus krippendorff
>
>-----Original Message-----
>From: Paul Dingle [mailto:paul_s_dingle@YAHOO.COM]
>Sent: Wednesday, October 13, 2004 12:44 PM
>To: CONTENT@BAMA.UA.EDU
>Subject: [CONTENT] Software for Reliability
>
>
>Hi All,
> Just wondering if there are any freely available
>software programs/scripts for calculating the
>Krippendorff Alpha ?
>(preferably Linux but Windows will do too). Thanks,
>
>Paul
>
>
>__________________________________________________
---------------------------------------------------------
CONTENT is the Internet mailing list for news and discussion of
content analysis. For additional information (including
information regarding "signoff" procedures), visit the Content
Analysis Resources web site, at http://www.car.ua.edu.