Date of Original Version



Working Paper

Rights Management

All Rights Reserved

Abstract or Description

We conducted a series of empirical experiments in which users were asked to enter digit strings into the computer by voice or keyboard. Two different ways of verifying and correcting the spoken input were examined. Extensive timing analyses were performed to determine which aspects of the interface were critical to speedy completion of the task. The results show that speech is preferable for strings that require more than a few keystrokes. The results emphasize the need for fast and accurate speech recognition, but also demonstrate how error correction and input validation are crucial for an effective speech interface.


Later published in Proceedings of the Third DARPA Speech and Natural Language Workshop, San Mateo: Morgan-Kaufmann, 1990, p. 219-224.