KWAL ignoring repetitions

[Brielle Stark]

Hi there,

I'm running KWAL for selected content words in a sample. I do not want to include any word specified as a target [: target] or as a coded error [* s] or [* phon]. For other analyses (in freq), we've used -sm@* and -sm** to specify this. I tried to use this same specification for KWAL analysis but the output still considered words within targets.

For instance:

kwal +d4 -sm** -sm@* +sfairy +sdress +t*PAR  @

and output line for a person's transcript was:

but the fowry [: fairy] [* phon] took [* s] a dress
%mor: conj|but det:art|the n|fairy v|take&PAST det:art|a n|dress .

Is there any way to tell KWAL to ignore what is in the target brackets and also to ignore any word followed by an error code? Many thanks,

Brie

--

Brielle Stark, PhD

Post Doctoral Fellow in Communication Sciences and Disorders, University of South Carolina
t: +1 803-777-9240, alternate email: sta...@mailbox.sc.edu

Aphasia Lab: http://web.asph.sc.edu/aphasia/
Center for the Study of Aphasia Recovery: http://web.asph.sc.edu/cstar/

[Davida Fromm]

Hi Brie,

Okay, just to confirm …

You want kwal to ignore “fowry” and “fairy” and “took” on the speaker tier in the example sentence you sent?

-Davida

--
You received this message because you are subscribed to the Google Groups "chibolts" group.
To unsubscribe from this group and stop receiving emails from it, send an email to chibolts+u...@googlegroups.com.
To post to this group, send email to chib...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/chibolts/CAEs2yToKJjXkzm-DJTOpBg9eSYT0XaOaGSxCvWuSq-AADRKGFw%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

[Leonid Spektor]

Brie,

The command you used is supposed to output your example tier, because the command line includes search for word "dress" and this word is not followed by [* ...] code or is in [: ...]. The same command for FREQ "freq -sm** -sm@* +s"n|fairy" +s"n|dress" +t*PAR @" would also output word "n|dress". The reason it did not output word "dress" using your command with "+sdress" option is because the word "dress" does not exist on %mor tier, so by pure luck your "+sdress" option did not find that word.

One thing to remember is that KWAL will output tiers even if only one of word specified with +s option is located on the tier. if you want to exclude tiers that have either one of the specified words, then you need a different command line.

As best as I understand you search criteria I think one of the following commands will work for you:

This command will still output example tier because the word "dress" is not an error or a target replacement.

kwal -s<\**> -r5 +sfairy +sdress +t*PAR kwal.cha

If you want to exclude tier if ether one of the words is found on it, then you need two sequential command:

kwal -2 +o@ -s<\**> -r5 +sdress +t*PAR +f +d @

kwal +d4 -s<\**> -r5 +sfairy +t*PAR *kwal.cex

If you have more questions about how to customize command best for your requirements, then please email to me directly.

Leonid.

[Brielle Stark]

Not sure why I wrote "repetitions" in the tag line... brain dead on a Thursday!

So, using KWAL, I'd be looking for correctly expressed "fairy" and "dress" by the participant, of which the only instance was "dress" in this example since fowry was coded as an error. Essentially, whenever I try to run the command, the analysis looks within the targets that we write after the errors [: fairy], which we write purely for ourselves so we know why we coded a certain way. We would like to exclude any word from within a target [: fairy] from the KWAL analysis.

In this example, my ideal KWAL command would output that the participant said one unit of "dress" and did not say "fairy."

PAR: but the fowry [: fairy] [* phon] took [* s] a dress 

%mor: conj|but det:art|the n|fairy v|take&PAST det:art|a n|dress .

I'm not sure if I follow Leonid's suggestion, so I'm going to try to summarize it here. 

The command "kwal -s<\**> -r5 +sfairy +sdress +t*PAR kwal.cha" will give me only "dress" as the output? I don't think I understand the second suggestion of 'exclude tier of either one of the words is found on it,' sorry.

Thanks for all the help,

Brie

On Thu, Aug 17, 2017 at 10:14 AM, Davida Fromm <fr...@andrew.cmu.edu> wrote:

Hi Brie,

Okay, just to confirm …

You want kwal to ignore “fowry” and “fairy” and “took” on the speaker tier in the example sentence you sent?

-Davida

On Aug 17, 2017, at 8:41 AM, Brielle Stark <briell...@gmail.com> wrote:

Hi there,

I'm running KWAL for selected content words in a sample. I do not want to include any word specified as a target [: target] or as a coded error [* s] or [* phon]. For other analyses (in freq), we've used -sm@* and -sm** to specify this. I tried to use this same specification for KWAL analysis but the output still considered words within targets.

For instance:

kwal +d4 -sm** -sm@* +sfairy +sdress +t*PAR  @

and output line for a person's transcript was:

but the fowry [: fairy] [* phon] took [* s] a dress
%mor: conj|but det:art|the n|fairy v|take&PAST det:art|a n|dress .

Is there any way to tell KWAL to ignore what is in the target brackets and also to ignore any word followed by an error code? Many thanks,

Brie

--

Brielle Stark, PhD

Post Doctoral Fellow in Communication Sciences and Disorders, University of South Carolina
t: +1 803-777-9240, alternate email: sta...@mailbox.sc.edu

Aphasia Lab: http://web.asph.sc.edu/aphasia/
Center for the Study of Aphasia Recovery: http://web.asph.sc.edu/cstar/

--
You received this message because you are subscribed to the Google Groups "chibolts" group.

To unsubscribe from this group and stop receiving emails from it, send an email to chibolts+unsubscribe@googlegroups.com.

To post to this group, send email to chib...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/chibolts/CAEs2yToKJjXkzm-DJTOpBg9eSYT0XaOaGSxCvWuSq-AADRKGFw%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "chibolts" group.

To unsubscribe from this group and stop receiving emails from it, send an email to chibolts+unsubscribe@googlegroups.com.

To post to this group, send email to chib...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/chibolts/02C808E2-813C-4920-ABFE-EF36C11CE8B5%40andrew.cmu.edu.

For more options, visit https://groups.google.com/d/optout.

[Leonid Spektor]

Brie,

KWAL outputs the whole tier and not just one particular unit of a word. The command you have does in fact output your ideal KWAL output. Perhaps you should use FREQ command instead of KWAL, since you are looking for word units?

Leonid.

To unsubscribe from this group and stop receiving emails from it, send an email to chibolts+u...@googlegroups.com.

To post to this group, send email to chib...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/chibolts/CAEs2yTpsBuzox__heCOpXrJL9qh5urDJCy3Aw9PwwEwxu6WDBw%40mail.gmail.com.

[Brielle Stark]

That makes since, Leonid. I'll use freq! Thank you very much.

Brie

Brie,

To view this discussion on the web visit https://groups.google.com/d/msgid/chibolts/CAEs2yTpsBuzox__heCOpXrJL9qh5urDJCy3Aw9PwwEwxu6WDBw%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "chibolts" group.
To unsubscribe from this group and stop receiving emails from it, send an email to chibolts+unsubscribe@googlegroups.com.
To post to this group, send email to chib...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/chibolts/C0EEE7E4-8B27-4F76-AA79-7EE118EC29CF%40andrew.cmu.edu.

For more options, visit https://groups.google.com/d/optout.

[Brielle Stark]

*sense, rather. Again, thanks!

Brie