Comparative Analysis of Text Data in Successful Face-to-Face and Electronic Negotiations

Various combination of Natural Language Processing and Machine Learning methods offer ample opportunities wherever texts are an important element of an application or a research area. Such methods discover patterns and regularities in the data, seek generalization and in effect learn new knowledge. We have employed such methods in learning from a large amount of textual data. Our application is electronic negotiations. The genre of texts found in electronic negotiations may seem limited. It is an important research question whether our methods and findings apply equally well to texts that come from face-to-face negotiations. In order to confirm such more general applicability, we have analyzed comparable collections of texts from electronic and face-to-face negotiations. We present our findings on the extent of similarity between these two related but distinct genres. In this study we have analyzed similarities in the text data of electronic and face-to-face negotiations. The results show that – in certain conditions – vocabulary richness, language complexity and text predictability are similar.This is an expanded version of a paper published in the Proceedings of FINEXIN 2005 (Workshop on the Analysis of Formal and Informal Information Exchange during Negotiations), 31–42, Ottawa, Canada, May 2005.