Exploring Consistency of Reading Rate Among Text-to-Speech Tools to Optimize the Comprehension of Struggling Readers

• In Event: Poster Session 13
  In Poster Session: PS 13 Topic - Technology, Media & Child Development

Abstract

Introduction: Due to the COVID-19 pandemic students have more options to engage with digital platforms for remote learning. Within these platforms, text-to-speech tools are frequently embedded with various features (e.g., reading rate and voice type). These text-to-speech tools are critical for students with print disabilities to access text. Text-to-speech is used to translate written text into spoken language, enabling one to listen while reading along.

Research demonstrates that text-to-speech can positively impact comprehension but significant variability in its effectiveness exists. These differences may be due to variations in text-to-speech settings. Previous research shows that how users interact with the text-to-speech features, may impact overall academic performance (Wood et al., 2018). Students with little practice in manipulating the text-to-speech reading rate, may inadvertently set the text-to-speech speed at a less than optimal level which may effect their reading of course material.

Objective: This study will investigate the default reading rate settings offered in text-to-speech programs and determine if there is a significant difference in the default setting across text-to-speech tools. Are there significant differences in slow, medium, and fast reading rate settings? Students tend to manipulate these reading rates. If reading rates are drastically different among these platforms, a student may not understand these differences in rate and know how to maintain a consistent optimal rate for comprehension across various platforms. For example, the student may access text-to-speech in a textbook, handout or on their phone, with each program utilizing various text-to-speech reading rates. Therefore, we will explore if there are significant differences between the slow, medium, and fast reading rate settings across various text-to-speech tools.

Hypothesis: There are differences in speed among the default settings and inconsistent speeds of slow, medium, and fast across platforms/software programs (e.g., Chromebook, Macbook, Kurzweil, Microsoft Word, and Texthelp). The hypothesis is that the more specialized software for text-to-speech will have a higher default setting speed than built-in text-to-speech programs.

Methods: A set of 22 5th grade standardized passages from the Gates-MacGinitie Reading Test will be used to investigate the differences under these three different conditions (i.e., slow, medium, fast). Every passage will be read by each text-to-speech program in slow, medium, and fast settings. The reading rate (words per minute) will be calculated by the total number of words per passage and the total reading time of the text-to-speech software for that passage.

Analysis: General Linear Modeling will be used to explore if there are statistically significant differences between default reading rate settings of the text-to-speech tools.

Implications: The results of this study can help inform educators on how reading rate across different devices can impact a student’s reading experience and ultimate comprehension of the text.

Feasibility of Completion: Researchers have obtained all study materials including specialized text-to-speech software and standardized reading passages. Data collection and analyses can be accomplished remotely without pandemic interruptions. Data will be collected by January 1, 2021. Analyses will be run in February. The content of the poster including tables and figures will be completed in March.

Authors

• Sarah Wood, Florida State University
  Presenting Author

• Jennifer L Keelor, Mount St. Joseph University
  Non-Presenting Author