Extended communication samples of augmented communicators II: Analysis of multiword sequences

K. M. Yorkston, D. R. Beukelman, K. Smith, R. Tice

Research output: Contribution to journalArticlepeer-review

12 Scopus citations


The three primary purposes of this project are (a) to identify those word sequences that occur frequently across a group of 10 linguistically intact augmented communicators, (b) to determine the communality with which the various augmented communicators use specific word sequences, and (c) to evaluate the usefulness of word sequences in providing keystroke savings for users of augmentative and alternative communication (AAC) devices as compared to letter-by-letter spelling or single-word retrieval. Communication samples were collected for 14 consecutive days from 10 subjects who used AAC devices. These samples were analyzed with custom software to determine the frequency of occurrence of one-, two, three, four, and five-word sequences. The results revealed that three-, four-, and five-word sequences occurred very infrequently within and across communication samples. Two-word sequences occurred much more frequently than the longer sequences. The authors concluded that clinicians should continue to encourage linguistically intact clients to select long messages (three words or more) for their AAC devices based on message content and the urgency with which messages need to be delivered rather than relying exclusively on statistical analysis. However, statistically based multiword analysis should not be abandoned. Frequently occurring one- and two-word sequences may be an important source of messages that will increase communication efficiency when stored and retrieved in an AAC device.

Original languageEnglish (US)
Pages (from-to)225-230
Number of pages6
JournalJournal of Speech and Hearing Disorders
Issue number2
StatePublished - 1990

ASJC Scopus subject areas

  • Otorhinolaryngology


Dive into the research topics of 'Extended communication samples of augmented communicators II: Analysis of multiword sequences'. Together they form a unique fingerprint.

Cite this