Automated grammatical tagging of child language samples

Recent studies of the automated grammatical categorization ("tagging") of w ords using probabilistic methods have reported substantial levels of accura cy-over 95% agreement with manual tagging for words from a variety of texts . However, the texts with which this method has been tested were written by adults and edited by publishers. The present study examined the accuracy w ith which such methods could tag transcribed conversational language sample s from 30 normally developing children. On a word-by-word basis, automated accuracy levels ranged from 92.9% to 97.4%, averaging 95.1%. Accuracy at co rrectly tugging whole utterances was lower, ranging from 60.5% to 90.3%, wi th an average of 77.7%. Probabilistic methods of coding language samples ho ld potential as a viable tool for child language research. Further study an d improvement of automated grammatical tagging is warranted and necessary b efore widespread use can be made of this technology.