Questions about CHIP

[Jenna]

Hello again, 

I am using the CHIP utility for a research project. I am interested in looking at how much mothers imitate and expand upon child utterances and whether that facilitates language development.

I have a few questions about some of the summary measures included in CHIP. I plan to use %_Overlap and %_EXPAN in my analysis, but I am not sure I understand Avg_Dist, Rep_Index, and %_IMITAT. These are the descriptions included in the manual:

“$DIST: the distance the response utterance is from the source”

“Avg_Dist: The sum of the DIST values divided by the total number of overlapping utterances.” 

“Rep_Index: Average proportion of repetition between the source and response utterance across all the overlapping responses in the data.” 

“%_IMITAT: ?? The numerator is the number of imitations and the denominator is ???”

My questions are: 

1. How is distance measured? Is this how many utterances there are between source and response?
2. Is the Rep_Index how closely the source and response match (e.g. 3 out of 4 words match)?
3. Does %_IMITAT measure percent of responses that exactly imitate the source? What is the denominator?

Any further explanation you could provide on these measures would be very helpful. Thank you again for all of your help and for this wonderful program!

Jenna K. Poland

M.A Speech-Language Pathology Candidate, May 2015
Graduate Assistant to Nan B. Ratner, Ed.D
University of Maryland, College Park

[Brian MacWhinney]

Dear Jenna,

    Page 68 of the manual explains that the nature of the source and the target varies depending on whether the report is in the %chi, %adu, %csr, or %asr tier.  On page 70, the manual explains that distance refers to distance in terms of utterances.  Yes, the Rep_Index reflects the number of matching words.  Regarding %_IMITAT, you might notice that I was a bit unclear about this index when I wrote the documentation in the manual and so I put question marks.  I assume the denominator is the number of utterances.  If this is important to you, we could set up a large test corpus to see exactly what this is doing.  There is a chip.cha file in the /examples folder that we could extend to check on this.

—Brian MacWhinney

--
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/bcfa6537-5981-4ffc-a4f4-356baac83272%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.