OpenConf

Full Program »

Word commonality for High Frequency Words comparing an Individual to the Composite samples

File
View File
pdf
271KB

This study investigated the degree to which the high frequency (HF) words of an individual’s language sample overlapped with the HF words from a spoken language corpus. The goal was to demonstrate a strategy to support vocabulary selection for adults who rely on augmentative and alternative communication (AAC). Using the British national corpus (BNC), 66 individual samples gathered from non-disabled English speaking adults were analyzed. The analysis consisted of 5,000 words from each individual for a composite sample of 333,000 words. A grouped frequency distribution was used to identify the HF words. Then the percentage of overlapping HF words between the composite sample and an individual’s sample was computed.
The results found 736 HF words from the composite sample using a frequency range from 0.1‰. In addition, 276.48 HF words on average (SD = 14.90) were identified using a frequency range from 0.5‰ across the 66 individual samples. An analysis revealed that an average of 88.12% HF words in the individual samples overlapped with the HF words of the composite sample (SD = 4.64%). In conclusion, this study showed that well-defined HF words from corpus data can be used as one source of external evidence to support AAC vocabulary selection for an individual to guild the design of a treatment protocol.

Author(s):

Sangeun Shin    
University of Pittsburgh
United States

Katya Hill    
University of Pittsburgh
United States

 

Powered by OpenConf®
Copyright ©2002-2014 Zakon Group LLC