syntactic measures for a growth curve analysis

[Yuriko Oshima-Takane]

Hi  Brian and colleagues,

 

We would like to do a growth curve analysis on some syntactic measures

using the functional data analysis. MLU is not a good syntactic measure

for this analysis because it sometimes goes down like 1.56 -> 2.20 ->1.60 ->2.15.

We are wondering if MLUs excluding one morpheme utterances could be

used for this analysis.  But we are not sure how to exclude all the one

morpheme utterances from the transcripts and calculate MLUs.  Could you

please let us know how to do that? Also if you know any other syntactic

measure which would be good for a growth curve analysis, please let us know.

Thanks.

 

Yuriko

 

[Brian MacWhinney]

Dear Yuriko,

     This is a recurring question for child language researchers.  The idea of calculating

MLU based on the exclusion of all one-morpheme utterances is a nice one.  In theory, I would have

thought that I could get KWAL to search for all utterances that contained a space using a switch such

as +s"* *"  and then send that all to a new file or to MLU, but this doesn't work.  Maybe Leonid knows how to 

get this to work.

  Another (probably more useful) approach is to use DSS to measure later syntactic growth.

-- Brian

-- 
You received this message because you are subscribed to the Google Groups "chibolts" group.
To post to this group, send email to chib...@googlegroups.com.
To unsubscribe from this group, send email to chibolts+u...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/chibolts?hl=en.

[Leonid Spektor]

Yuriko and Brian.

The +s"* *" option doesn't work because kwal and most of the CLAN commands are word oriented. This means that they do not know about space characters between words.

We have implemented a different way to achieve the same result. Here is the command that will do the trick:

maxwd +x2 +d1 +g2 +c50000 +f*.cha

That is assuming that there are fewer than 50,000 utterances per file. If your have more, the just increase that number.

After that run command:

mlu -t%mor *.mxwrd.cex

Leonid.

[Leonid Spektor]

Yuriko and Brian.

I apologize for sending the same email again, but there was a typo in my original email.

The +s"* *" option doesn't work because kwal and most of the CLAN commands are word oriented. This means that they do not know about space characters between words.

We have implemented a different way to achieve the same result. Here is the command that will do the trick:

maxwd +x2 +d1 +g2 +c50000 +f *.cha

That is assuming that there are fewer than 50,000 utterances per file. If your have more, the just increase that number.

After that run command:

mlu -t%mor *.mxwrd.cex

Leonid.

On Feb 15, 2012, at 22:11, Brian MacWhinney wrote:

[Leonid Spektor]

Yuriko and Brian.

I just found an even simpler command to do the same thing:

kwal +x2w +d +f +t%mor *.cha

mlu *.kwal.cex

Leonid.

On Feb 15, 2012, at 22:11, Brian MacWhinney wrote:

[Yuriko Oshima-Takane]

Hi Leonid and Brian,

 

Thank you for your great help!

We will try the command to see if MLU excluding one morpheme utterances would work for

our growth curve analysis. If not, we will try DSS.

 

Yuriko

 

From: chib...@googlegroups.com [mailto:chib...@googlegroups.com] On Behalf Of Leonid Spektor
Sent: Thursday, February 16, 2012 12:01 AM
To: Brian MacWhinney; chibolts
Subject: Re: syntactic measures for a growth curve analysis

 

Yuriko and Brian.

 

I just found an even simpler command to do the same thing:

 

kwal +x2w +d +f +t%mor *.cha

mlu *.kwal.cex

 

 

Leonid.

 

 

 

On Feb 15, 2012, at 22:11, Brian MacWhinney wrote: