checking files
Brian MacWhinney
macw at cmu.edu
Thu Feb 18 02:16:34 UTC 1999
Dear Info-CHILDES,
We have recently finished re-checking all of the CHILDES data. The
checked files are on childes.psy.cmu.edu in both Mac and PC format.
The major things we changed were:
1. Whereever possible, files now have an @ID field for the
Target_Child. This allows people to run STATFREQ easily. We will also
use this field for other things in the future. If you are currently
working with a copy of your own data that does not have these fields,
you may want to get these copies.
2. To further facilitate cross-file analysis, children in the role of
Target_Child are now always coded as *CHI. Adults in the role of
Target_Adult are always coded as *ADU.
3. Check also caught some errors for redundant delimiters which we now
fixed.
In the process of going through the data, I noticed a tendency for
people to use forms like &cause when forms like (be)cause are more
helpful for analysis, MLU, and even readability. In general, I would
recommend trying to use the second type of form wherever possible.
--Brian MacWhinney
More information about the Info-childes
mailing list