questions about Kideval

[Li Zeng]

Hello. 

I ran Kideval on the sample of the child L2 English data, and came across some problems. 

1) I got the excel spreadsheet of Kideval results, which look very helpful. However, I want to know more about the nature (the formula) of calculating the following indexes, 

verb/utterances:  does it mean verb tokens or verb types divided by the number of utterances?

TTR: what is the formulae of measuring the type/token rate in the spreadsheet result of Kideval?

utterance errors:  In my transcription of the data, I postcode erroneous utterances with [+ gram], but I'm not sure whether the Kideval knew that. How does Kideval calculate "utterance errors"?

2) After running Kideval, I've got a warning message, shown as below: 

WARNING: FOUND 2 %mor: TIERS THAT DO NOT LINK IN SIZE TO THEIR SPEAKER TIERS.

THIS WILL EFFECT THE RESULT OF IPSyn ANALYSES.

What does it mean by "do not link in size to their speaker tiers" ? and how can I fix it?

Thank you very much. 

Li 

[Brian MacWhinney]

Li,

Item #13 in KIDEVAL is about verb tokens.  

TTR is simply the ratio of types divided by tokens.

KIDEVAL is not designed to track utterance errors.  However, I believe that it has a methiod for excluding utterances if you use the [+ exc] postcode.  

If you get the error message about some %mor tiers not linking in size to their main tiers, this indicates that you probably ran MOR and then changed something after that.  If you run MOR again, it will correct this error.

--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/1b6b4111-9aab-4cde-8674-fcdd34a5eec6%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Li Zeng]

thank you very much Brian. 

Li