Abstract:

The presentation will provide results from a study designed to evaluate three different speech to text systems: (1) CART; (2) C-Print; and (3) Automatic Speech Recognition (ASR) for accuracy, students’ comprehension of lecture content, ease of use by the presenter and students, and cost benefit of each system. Three methods of presenting text and non-text content in close visual proximity were evaluated; a heads up display of generated text; a composite PC/laptop based system that allows text streaming with non-text content, e.g. Powerpoint; and display of the text on a wireless hand held mobile device which provides flexibility.

Students who are deaf or hard of hearing as well as students with other special needs such as vision inpairment, learning disabilities (LD), or attention deficit hyperactivity (ADHD), are often denied full participation in formal and informal educational activities due to the lack of access services that are flexible, easy to use, cost effective, available and portable. Continuing advances in technology have the potential to expand availability of real-time display text and non-text information that could reduce communication and educational barriers that still impede effective learning. Our study was initiated to help provide expanded strategic options and data-based guidance for students, teachers, employers, service providers and educational administrators who are trying to reach the goal of full participation in the educational environment and the world of work for students who are deaf or hard of hearing or have other special needs.

Data from our testing procedures with 108 students from diverse educational and linquistic backgrounds, and access service preferences, clearly demonstrate that ASR is an effective access service, both educationally and cost wise for providing access to information in formal and informal learning environments and access to the world of work. In addition, recent advances in ASR software has greatly reduced the training time for users, and greatly increased the accuracy of the output. Another benefit of ASR, often overlooked, but crtical for success in science, technology, engineering and mathematices (STEM disciplines) is that ASR provides direct access to the vocabulary of science and technology, which I have found to be very important for understanding the concepts behind the words, an understanding that is frequently missing. Based on 15 years of experience teaching students who are deaf or hard of hearing in mainstream classes at the university level, this is one area that is holding them back from achieving greater success in their studies, As good as our interprers are, signs used in STEM classes are often ambiguous, and can apply to different words or concepts, and may be "homemade" signs negotiated by the student and interpreter locally.

The results from this study will be presented in an interactive presentation where we share what we have learned about what works or doesn’t work; what can be done to modify off the shelf software/technology to optimize effectiveness in individual classrooms; what we learned about displaying multiple inputs of text and non-text information to enhance access to learning; and we hope to learn from the participants what they have learned from their experiences – the good, the bad and what to avoid so we can improve teaching and learning for all.

Research Supported by the National Science Foundation, Award # 0622854

Keywords:

deaf and hard of hearing students, speech to text, display modes, stem disciplines.