[Corpora-List] Distribution of errors in dictations

[Brett Reynolds]

In our EAP program, we include dictation as part of our listening  
assessment. One thing I've noticed is that errors seem to cluster  
together. I'd like to see if this is more than just a vague  
impression from marking too many tests. What kind of tools and  
statistics might be used to show whether something's happening here?

I suppose one thing I could do is simply count the number of error- 
free words that are followed by an error and compare that to the  
number of errors that are followed by an error. I remember though  
some ways for measuring "burstiness", and I wonder if that might be  
relevant. Any thoughts?

Best,
Brett

<http://english-jack.blogspot.com>

-----------------------
Brett Reynolds
English Language Centre
Humber College Institute of Technology and Advanced Learning
Toronto, Ontario, Canada
brett.reynolds at humber.ca



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/corpora/attachments/20090302/f1354ed7/attachment.htm>
-------------- next part --------------
_______________________________________________
Corpora mailing list
Corpora at uib.no
http://mailman.uib.no/listinfo/corpora