Evaluation of social media collaboration using task-detection methods

Abstract

Collaboration using social media is a good way of jointly constructing knowledge. This study aims at better understanding collaborative knowledge construction processes by applying innovative (micro-)task detection approaches. We take a closer look at the interactions of a user with a shared digital artifact by analyzing the captured interaction data. The goal is to identify domain-independent interaction patterns, which can serve as indicators for knowledge development (operationalized as accommodation). We designed an empirical study under laboratory conditions that used our method. The applied task detection approach identified accommodation with a rate of 77.63% without resorting to textual features. This result instantiates an improvement as compared to a previous study in which the text in focus was identified as the feature with best discriminative power. We discuss our hypothesis that our method is independent from the used knowledge domain. © 2011 Springer-Verlag Berlin Heidelberg.