From lackaff at gmail.com  Mon May 14 12:26:48 2007
From: lackaff at gmail.com (Derek Lackaff)
Date: Mon, 14 May 2007 12:26:48 -0400
Subject: [Weft QDA users] Inaccurate code review
Message-ID: <243f58150705140926n6e9ab9b5s92b46ac7a68db95b@mail.gmail.com>

Hello,

I'm using Weft 1.0.1 (Win) to code a series of focus group transcripts, but
seem to be getting inaccurate results when using the Code Review tool.

Each transcription is a different document. Each document contains several
passages that have been coded in multiple categories. The entire text of
each document has been categorized with a title (FG1, FG2, etc.) When I run
the coding review, I setup the content categories as rows and the title
categories as columns, and set the "Display Options" to "Number of
Passages".

Th problem seems to be that only the first coded example of the row category
is found in each document - multiple instances of categorized content in a
document are not counted, so the table contains only 0s and 1s. Switching to
"Number of Documents" does not change the table, and switching to "Number of
Characters" displays the number of characters in the single passage that was
found.

I'd greatly appreciate any suggestions on how to correct this counting
issue. I've been very impressed with Weft's simplicity and ease-of-use thus
far!

Cheers,

Derek
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/weft-qda-users/attachments/20070514/f7bdb601/attachment.html 

From branson.patrick at gmail.com  Wed May 16 09:52:45 2007
From: branson.patrick at gmail.com (Patrick Branson)
Date: Wed, 16 May 2007 21:52:45 +0800
Subject: [Weft QDA users] Inaccurate code review
In-Reply-To: <243f58150705140926n6e9ab9b5s92b46ac7a68db95b@mail.gmail.com>
References: <243f58150705140926n6e9ab9b5s92b46ac7a68db95b@mail.gmail.com>
Message-ID: <57c35dba0705160652v48abc73ci523931ad119af42@mail.gmail.com>

Derek,
I'm also using Weft. I checked out your problem using my data. The Code
Review worked for me. This is how I used it.
I have transcripts of interviews and journals. I have developed a number of
categories and coded the transcripts according to the categories. I know
some of the categories are nested one in the other. It's obvious to me from
reading the transcripts, but I don't know accurately to what extent they are
nested.
Because I'm still only part way through my data analysis, I haven't worried
too much about the Code Review feature of the programme. Some months ago, I
played around with it but it didn't make any sense to me. When I received
your email, I decided to find out for myself just what it would do.
I worked with one category at a time. I set the display option according to
the number of passages and minimised the Code Review box. Then I went to the
Documents and Categories drop down menu and selected my first category. When
I restored the Code Review box the category was highlighted in the
Categories box. I clicked on Add as row and also on Add as column. The
number of passages containing the category appeared in the cell. Then I
minimised the Code Review box again, brought up the Documents and Categories
menu, selected the next category and restored the Code Review. I clicked on
Add as row and Add as column. It then did its "spreadsheet" caper and gave
me the number of passages that contained both categories and the numbers of
passages that did not contain both categories. I kept going with this
procedure. It produced a nice little spreadsheet with some interesting
features that should keep me going for hours. The export function works
well. I brought it up in Excel. If I ever need to drop it into my thesis -
and I hope I don't ever do that - it should work okay.
There are features in the Code Review that I haven't checked out yet, but
I'm sure they will work when I come to use them.
I hope my rambling has been of help.
Good luck with your work.
Pat Branson


On 5/15/07, Derek Lackaff <lackaff at gmail.com> wrote:
>
> Hello,
>
> I'm using Weft 1.0.1 (Win) to code a series of focus group transcripts,
> but seem to be getting inaccurate results when using the Code Review tool.
>
> Each transcription is a different document. Each document contains several
> passages that have been coded in multiple categories. The entire text of
> each document has been categorized with a title (FG1, FG2, etc.) When I run
> the coding review, I setup the content categories as rows and the title
> categories as columns, and set the "Display Options" to "Number of
> Passages".
>
> Th problem seems to be that only the first coded example of the row
> category is found in each document - multiple instances of categorized
> content in a document are not counted, so the table contains only 0s and 1s.
> Switching to "Number of Documents" does not change the table, and switching
> to "Number of Characters" displays the number of characters in the single
> passage that was found.
>
> I'd greatly appreciate any suggestions on how to correct this counting
> issue. I've been very impressed with Weft's simplicity and ease-of-use thus
> far!
>
> Cheers,
>
> Derek
>
> _______________________________________________
> weft-qda-users mailing list
> weft-qda-users at rubyforge.org
> http://rubyforge.org/mailman/listinfo/weft-qda-users
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/weft-qda-users/attachments/20070516/b4515868/attachment.html 

From alex at pressure.to  Thu May 17 06:21:09 2007
From: alex at pressure.to (Alex Fenton)
Date: Thu, 17 May 2007 11:21:09 +0100
Subject: [Weft QDA users] Inaccurate code review
In-Reply-To: <243f58150705140926n6e9ab9b5s92b46ac7a68db95b@mail.gmail.com>
References: <243f58150705140926n6e9ab9b5s92b46ac7a68db95b@mail.gmail.com>
Message-ID: <464C2C95.90305@pressure.to>

Derek

Thanks for your message; I'm sorry for the delay in replying.
> I'm using Weft 1.0.1 (Win) to code a series of focus group 
> transcripts, but seem to be getting inaccurate results when using the 
> Code Review tool.
>
> Each transcription is a different document. Each document contains 
> several passages that have been coded in multiple categories. The 
> entire text of each document has been categorized with a title (FG1, 
> FG2, etc.) When I run the coding review, I setup the content 
> categories as rows and the title categories as columns, and set the 
> "Display Options" to "Number of Passages". 
I set up a small project and did some coding in a similar way, then 
tested it in the code review. I'm afraid I'm getting similar 
inaccuracies - it appears to be a bug in the calculations of number of 
passages.

Thanks for reporting this - it will be fixed asap. In the meantime, a 
manual technique may be possible if the number of categories is not 
massive. You could view the 'real' coding categories, then use the Find 
tool to search for occurrences of each document's title. It will find 
the headings that precede each passage in the viewer.

Alternatively, you might try exporting the category to a plain text 
file. If you then open that in a word-processor, it might have 
facilities for counting the number of occurrences of the document's 
title as a pre-passage header.

sorry for the trouble
alex