Date of Award:

8-2023

Document Type:

Thesis

Degree Name:

Master of Science (MS)

Department:

Mathematics and Statistics

Committee Chair(s)

Alan Wisler

Committee

Alan Wisler

Committee

Brennan Bean

Committee

Stephen Walsh

Abstract

People with communication disorders often experience difficulties being understood by unfamiliar listeners or in noisy environments. A common strategy for effectively communicating in these scenarios is to use simpler and more predictable language. Despite the prevalence of this strategy, there has been little to no research to date focused on the effectiveness of language simplification as a communication strategy. This study seeks to begin filling that gap by using natural language processing to determine whether speakers with early-stage Parkinson’s disease and age-matched neurotypical speakers are able to successfully simplify their language while still maintaining the original message.

Simplification was measured by several lexical diversity and lexical sophistication measures. Natural language processing methods were deployed to automatically compute the above metrics for text transcriptions of a story simplification task by each participant. A similarity score was also calculated to measure how closely each retelling mapped to the original story.

Each measure indicated that both groups showed statistically significant reductions in the complexity of their language in the rephrased passage relative to the original text and were overall able to maintain the original story’s message. These results provide strong preliminary evidence for the efficacy of language simplification as a communication strategy.

Checksum

b75c91478796b26261ddf03bb6b340dc

Share

COinS