Comment Evaluation and Revision in a Bilingual Electronic Meeting

Abstract

Translation accuracy continues to be a major problem in multilingual electronic meetings and a significant reason for such systems lack of acceptance and use. One way of increasing accuracy is detecting potential errors before a comment is added to the discussion. By comparing the original message text with a round-trip translation (RTT) and correcting any wording mistakes, misunderstandings might be prevented in the overall conversation. In this study, one group used an electronic meeting system with automatic translation that detected differences between a participants typed comment and a RTT. If there was a significant change, the group member was then given an opportunity to modify the text before submitting it to the transcript. Another group, serving as the control group, used an equivalent electronic meeting system without error detection. Results show that there was only a slight, non-significant increase in comprehension when comments were revised and translated to German with an 80% similarity threshold, but there would have been a significant increase in comprehension if a 50% threshold had been chosen.