Question about flucalc in bilingual samples

12 views
Skip to first unread message

Shelley Brundage

unread,
Apr 4, 2018, 4:15:56 PM4/4/18
to chib...@googlegroups.com
hello Chibolts,
I am analyzing some data from bilingual children. The CHAT files have both English and Spanish words in them. In the Spanish files, the English words have been tagged with "@s". I have two questions about running flucalc for Spanish:

1. the MOR words output in the flucalc spreadsheet seems to be counting both Spanish and English words in the samples, rather than just the Spanish ones. I thought that tagging the English words in the Spanish files meant that they would be ignored in Spanish counts, but that does not appear to have happened (I cross-checked word count with FREQ). Is there a way to count just the Spanish words for flucalc?

2. In some instances the flucalc output says that words per minute is N/A. Does this mean that there were not enough utterances to calculate WPM? if so, what is the necessary number of utterances for this?

Thank you

Shelley

--
Shelley B. Brundage, Ph.D., CCC, BCS-F, Fellow-ASHA
Associate Professor
Speech, Language, and Hearing Sciences 
George Washington University
2115 G street NW, Suite 226
Washington, DC 20052
Office 202) 994-5008 

*she/her pronouns

Leonid Spektor

unread,
Apr 4, 2018, 4:39:51 PM4/4/18
to chib...@googlegroups.com
Shelley,

1. at this time FLUCALC does not distinguish between words of one language versus another. FLUCALC considers all text as one.

2. means that there were no bullets in a file from which total time value could be computed. If you files does have bullets, then please email it to me for testing.


Leonid.

--
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/CAOooO9YXsyN_YA748CXU-UF6piYgVfN1AA-a22ozxafBP_k%2B-g%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

Shelley Brundage

unread,
Apr 4, 2018, 4:45:26 PM4/4/18
to chib...@googlegroups.com
hi Leonid,
I have a follow up question
I am getting the same number of words from FLUCALC and FREQ. so does that mean that freq doesn't distinguish words of one language, either?

On Wed, Apr 4, 2018 at 4:39 PM, Leonid Spektor <spe...@andrew.cmu.edu> wrote:
Shelley,

1. at this time FLUCALC does not distinguish between words of one language versus another. FLUCALC considers all text as one.

2. means that there were no bullets in a file from which total time value could be computed. If you files does have bullets, then please email it to me for testing.


Leonid.
On Apr 4, 2018, at 16:15, Shelley Brundage <brun...@gwu.edu> wrote:

hello Chibolts,
I am analyzing some data from bilingual children. The CHAT files have both English and Spanish words in them. In the Spanish files, the English words have been tagged with "@s". I have two questions about running flucalc for Spanish:

1. the MOR words output in the flucalc spreadsheet seems to be counting both Spanish and English words in the samples, rather than just the Spanish ones. I thought that tagging the English words in the Spanish files meant that they would be ignored in Spanish counts, but that does not appear to have happened (I cross-checked word count with FREQ). Is there a way to count just the Spanish words for flucalc?

2. In some instances the flucalc output says that words per minute is N/A. Does this mean that there were not enough utterances to calculate WPM? if so, what is the necessary number of utterances for this?

Thank you

Shelley

--
Shelley B. Brundage, Ph.D., CCC, BCS-F, Fellow-ASHA
Associate Professor
Speech, Language, and Hearing Sciences 
George Washington University
Office 202) 994-5008 

*she/her pronouns

--
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.

--
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/4AD17735-51C3-4523-A9AF-D4BBC3B88672%40andrew.cmu.edu.

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

Leonid Spektor

unread,
Apr 4, 2018, 4:48:45 PM4/4/18
to chib...@googlegroups.com
That is correct. Non of CLAN commands try to infer user's intentions or to second guess user's wishes. Instead it is up to user to specify exactly what he or she wants from CLAN command.


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.

Brian MacWhinney

unread,
Apr 4, 2018, 4:49:42 PM4/4/18
to ChiBolts
Shelley,

Yes, FREQ counts all words.  If you want FREQ to ignore the words in a second language, you can add -s"*@s".  However, for FluCalc, it would seem to me that excluding a single word from the other language would skew the analysis.  Of course, it makes sense to exclude whole utterances in the other language that are marked with codes like [- spa] and Leonid agrees that it makes sense to add this facility to FluCalc.

--Brian

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.

Shelley Brundage

unread,
Apr 4, 2018, 4:56:54 PM4/4/18
to chib...@googlegroups.com
hi Brian, 
One more question related to KIDEVAL, which I am also trying to use
The totals for FREQ Tokens in KIDEVAL do not match the FREQ total tokens. Is this because KIDEVAL works off the MOR line and FREQ off the main line?




--Brian

--
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/5B9446A8-F131-4A78-919C-3EA58CDF6E19%40cmu.edu.

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

Leonid Spektor

unread,
Apr 4, 2018, 5:02:38 PM4/4/18
to chib...@googlegroups.com
Shelley,

KIDEVAL computes lammas from %mor tier. To get the same result with FREQ you need to add to FREQ options " +sm;*,o%  *t*CHI ".


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.
Reply all
Reply to author
Forward
0 new messages