Conference Poster


Cloze probability in typed text: How predictability in cloze, copying and picture naming tasks affects response times

Abstract

The cloze probability (CP) task is used to investigate the way in which sentences are processed. Participants’ task is to complete a sentence that is presented without the final word; however, sentences vary in levels of CP. High CP sentences have a highly anticipated completion word (e.g. Bradley prefers cats over dogs) whereas, low CP sentences can be completed with several different words, none of which are highly anticipated (e.g. In the distance they heard the kettle). These tasks have previously utilised spoken and hand-written forms of language production, but technological advances mean CP tasks need to be adapted to include typed language as a production modality. In the present study, sentences were presented online to be completed through typed text. Three tasks were included. A standard CP task in which participants completed sentences by freely producing the final word, a picture naming task where the final word of the sentence was presented in picture form, and a copying task. High CP sentences resulted in quicker reaction times of latency and duration compared to low CP sentences. Copying tasks gained the fastest responses, followed by picture naming and then cloze tasks. The effect of language processing, shown through CP tasks, remains present when utilising typed text as modality of language production. Copying task results suggested prediction may not be required in all language production. These results assist in relating CP tasks to modern methods of language production.

DOI (Digital Object Identifier)

Permanent link to this resource: https://doi.org/10.24384/xe8s-1380

Attachments

Authors

Budenberg, Eleanor

Oxford Brookes departments

Faculty of Health and Life Sciences
Department of Psychology, Health and Professional Development

Dates

Year: 2021


© The author

All rights reserved.


Related resources

This RADAR resource is Part of Get Published!

Details

  • Owner: Daniel Croft
  • Collection: Research
  • Version: 1 (show all)
  • Status: Live
  • Views (since Sept 2022): 631