Making Sense Of Text Responses - The Sequel

Abstract

Lectures can be made more engaging and interactive through response technology. While dealing with answers from multiple choice question is quite easy, the open text questions prove to be much more challenging.

This paper will build on the previous experiences related to processing the text responses gathered with response technology (i.e. one2act iLike) as described in Stoica & Thorseth (2016). Some of the challenges that were identified and which were related to the processing and visualization of text responses have been addressed and implemented in a new version of the iLike software.

The new iLike version has improved ways to deal with common and frequent words which do not bring much meaning by themselves and can clutter the word clouds used for visualizations. In addition a system to detect and remove repeated words (which we call “shouting”) has been created and integrated into the software.

Experiencing with the software indicated that the word cloud alone may not be enough to build the discussion on the spot thus we have created an interactive version of the word cloud that allows accessing the context in which a word appears.

The paper concludes with a discussion about our experiences using the presented features and future directions for further work remaining to address the still existing challenges.