Informatics Dissertation Defense: Mikko Tuomela

Informatics PhD candidate Mikko Tuomela will defend his dissertation, "Individual reading types and the effects of automated annotation." Tuomela's committee consists of Professor Michael Twidale (chair); Professor Emeritus Bertram "Chip" Bruce; Affiliate Professor Karrie Karahalios (Computer Science); and Catherine Marshall (Texas A&M).  

Abstract: In today's digital world, there is an overwhelming mass of data that is available. This "data deluge" poses a number of challenges – in general: how to deal with such an overflow of information? The fact that certain data exists does not mean that it is accessible; and if it is accessible, it is not necessarily usable. The information explosion will continue; therefore, our ability to analyze must be improved and enhanced.

In this work, noting the prevalence of the web interface to read content online, an augmented web solution is presented to help with the information explosion: a browser extension producing automated annotation based on dictionaries to assist the reader in finding relevant parts of the document more quickly and comprehending them more easily. What kind of an effect does this kind of an extension have on readers? How do readers employing different reading styles benefit (or not benefit) from its use? Does the extension also make the reader to pay more attention to their own reading types and style? A study of twenty-four participants using the browser extension is conducted. Transcripts from the test sessions and following interviews are analyzed under five themes: reading types; changing relationship with the document; own reading and annotation habits; accuracy, problems, and errors; and suggestions.

From the analysis of the data, three types of readers with different characteristics emerge: "Careful reader", "Jumper" and "Searcher". They are found to have different motivations for reading; using the extension to enhance the reading experience has different effects for each group as well. The annotations (highlightings) produced by the extension are found to attract readers’ attention; especially the “Searchers” found added visual information valuable. Descriptions of how the readers experienced the annotations are analyzed in detail. Several surprises are also noted: many participants seem to sometimes treat highlighted words as something of a concern; and some participants use text highlighting with mouse as a transient annotation practice which creates an interesting conflict with the annotations produced by the extension.

This study also contributes by offering ideas for future studies about reading, web augmentation, and digital annotation: several possible directions for future research are identified. An earlier experiment with sentiment analysis on Wikipedia is described as well.