Evaluating Collaborative Search for a Learning-Oriented Search Task

Abstract

Web search has become a convenient option for seeking information related to learning, therefore understanding how to facilitate human learning through a search engine has the potential to improve the quality of informal education and online learning. One less understood aspect of search as learning is the effect of collaboration in a search session focused on learning. Given the benefits of collaboration during both searching and learning, this thesis aims to evaluate the benefits of collaborative search for a learning-oriented search task.

We have developed a collaborative search system---SearchX---which adopts essential features of collaborative search as found in the literature. In the design process, we focused on providing support for research needs (such as running crowd-sourced experiments and fast prototyping). We deployed SearchX in a crowd-sourced user study on collaborative and single-user search for a learning-oriented search task with a total of 76 participants. Our results show that workspace awareness helped users in understanding the search task at the early stages of the search session and in selecting useful documents. However, we found little evidence that collaboration brought significant changes to the way users search and to users' learning outcomes. We found though that a good rapport between collaborators resulted in higher learning gains, highlighting the importance of social presence for online learning. Finally, we found indications that having a less difficult search topic results in less time spent on understanding search results and more exploration of the topic domain.