Predicting readers’ domain knowledge based on eye-tracking measures
| Published date | 10 December 2018 |
| Pages | 1027-1042 |
| Date | 10 December 2018 |
| DOI | https://doi.org/10.1108/EL-05-2017-0108 |
| Author | Quan Lu,Jiyue Zhang,Jing Chen,Ji Li |
| Subject Matter | Information & knowledge management,Information & communications technology,Internet |
Predicting readers’domain
knowledge based on
eye-tracking measures
Quan Lu and Jiyue Zhang
Department of Information Management, Wuhan University, Wuhan, China
Jing Chen
Department of Information Management, Huazhong Normal University,
Wuhan, China, and
Ji Li
Department of Information Management, Wuhan University, Wuhan, China
Abstract
Purpose –This paper aims to examine the effect of domain knowledge on eye-tracking measures and
predictreaders’domain knowledge from these measures in a navigational tableof contents(N-TOC) system.
Design/methodology/approach –A controlled experiment of threereading tasks was conducted in an
N-TOC system for 24 postgraduates of Wuhan University. Data including fixation duration, fixation count
and inter-scanning transitionswere collected and calculated. Participants’domain knowledgewas measured
by pre-experiment questionnaires. Logistic regression analysis was leveraged to build the prediction model
and the model’s performancewas evaluated based on baseline model.
Findings –The results showed that novices spent significantly more time in fixating on text area than
experts, becauseof the difficulty of understanding the information of textarea. Total fixation duration on text
area (TFD_T) was a significantly negative predictor of domain knowledge. The prediction performance of
logistic regression model using eye-tracking measures was better than baseline model, with the accuracy,
precisionand F(
b
= 1) scores to be 0.71,0.86, 0.79.
Originality/value –Little research has been reportedin literature on investigation of domain knowledge
effect on eye-tracking measures duringreading and prediction of domain knowledge based on eye-tracking
measures. Most studies focus on multimedialearning. With respect to the prediction of domain knowledge,
only some studies are found in the field of information search.This paper makes a good contribution to the
literature on the effect of domain knowledgeon eye-tracking measures during N-TOC reading and predicting
domain knowledge.
Keywords Reading, Domain knowledge, N-TOC, Eye-tracking, Prediction
Paper type Research paper
Introduction
Nowadays, personalization has been increasingly attracting research attention, for it is
an optimal way to improve reading performance and experience. Personalization tails
reading service to particular readers, which can be characterized according to readers’
attributes, including domain knowledge (Hwang et al., 2013), interest (Wang et al., 2016),
learning style (Hwang et al., 2013), etc. To perform personalization better, reading
systems need to learn about readers’attributes. Doing prediction through some implicit
feedback is a good way to gain the attributes, because of its advantage of not interrupting
readers.
Predicting
readers’
domain
knowledge
1027
Received17 May 2017
Revised9 September 2017
18November 2017
20February 2018
Accepted26 April 2018
TheElectronic Library
Vol.36 No. 6, 2018
pp. 1027-1042
© Emerald Publishing Limited
0264-0473
DOI 10.1108/EL-05-2017-0108
The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/0264-0473.htm
Domain knowledge has been well studied and found to have an influence on reading
performance. Researchers found that readers with high level of domain knowledge achieved
better performance than those with low level of domain knowledge (Surber and Schroeder,
2007). Additionally, research has demonstrated that domain knowledge is an important
factor whenproviding reading personalization support (Hwang et al.,2013).Therefore, it is of
great value to predict domain knowledge during reading.As one can imagine, personalizing
reading service for readers with different levels of domain knowledge may facilitate their
readingperformance and improve readingexperience.
In previous studies, a variety of methods such as logistic regression,stepwise regression
and classification and regressiontrees (CART), have been used (Liu et al., 2016;Wang et al.,
2016;Hu et al., 2014). With respect to predicting domain knowledge, behavioral measures
are widely used. For example, Zhang et al. (2011) chose search behavioral measures to
predict domain knowledgefrom searching process. But, actually, it is better to considereye-
tracking measures when predicting readers’domain knowledge. Because, although
behavioral measures are more easily obtained, eye-tracking measures are superior to
behavioral measures in terms of technology and methodology, when tracing information
processing (Rayner, 1986). Therefore, eye-tracking measures are regarded to trace
information processing during reading, by reflecting visual behavior quantitatively and
objectively (Just and Carpenter, 1980).Meanwhile, previous studies have provided evidence
for the domain knowledge effect on eye-tracking measures. Ho et al. (2014) revealed the
positive correlation between domain knowledge and inter-scanning transition during
reading. Tsai et al. (2012) suggested fixation duration reflected on individual’s domain
knowledge. With respect to fixationcount, Jie et al. (2014) found that low domain knowledge
participants fixated more frequently than high domain knowledge participants, when
learning musical scores of complex rhythm. However, in spite of the advantages of eye-
tracking measures, studies on predicting readers’domain knowledge from eye-tracking
measures have rarelybeen found.
A document reading system based on the navigationaltable of contents (N-TOC system)
is broadly used in reality, such as AdobeDigital Editions and Foxit Reader. A navigational
table of contents is the automatictable of contents, which generates a site map that points to
various parts of the document, thus providing a quick way to jump to the desired section.
Therefore, it is valuable to investigate how to provide a personalization reading service in
this system. Given this background information, we explore predicting domain knowledge
based on eye-tracking measures when reading in an N-TOC system. The purpose of this
study is to examine the effect of domain knowledge on eye-tracking measures and predict
readers’domain knowledge from these measures in an N-TOC system. Specifically, we
examined the followingresearch questions:
RQ1. Does readers’domain knowledgeaffect their eye-tracking measures when reading
in an N-TOC system? If so, how?
RQ2. By using logistic regression analysis, can readers’eye-tracking measures predict
their level of domain knowledge? If yes, what are the significant predicting
measures and how do they predict it?
To explore these questions, a control experiment was conducted in this study. According to
their domain knowledge, participants were divided into two groups: expert group and
novice group. All participants were asked to finish three reading tasks in a document
auxiliary reading system which supported reading the document with an N-TOC and two
questionnaires. Theireye-tracking measures, including fixation duration, fixation count and
EL
36,6
1028
Get this document and AI-powered insights with a free trial of vLex and Vincent AI
Get Started for FreeStart Your Free Trial of vLex and Vincent AI, Your Precision-Engineered Legal Assistant
-
Access comprehensive legal content with no limitations across vLex's unparalleled global legal database
-
Build stronger arguments with verified citations and CERT citator that tracks case history and precedential strength
-
Transform your legal research from hours to minutes with Vincent AI's intelligent search and analysis capabilities
-
Elevate your practice by focusing your expertise where it matters most while Vincent handles the heavy lifting
Start Your Free Trial of vLex and Vincent AI, Your Precision-Engineered Legal Assistant
-
Access comprehensive legal content with no limitations across vLex's unparalleled global legal database
-
Build stronger arguments with verified citations and CERT citator that tracks case history and precedential strength
-
Transform your legal research from hours to minutes with Vincent AI's intelligent search and analysis capabilities
-
Elevate your practice by focusing your expertise where it matters most while Vincent handles the heavy lifting
Start Your Free Trial of vLex and Vincent AI, Your Precision-Engineered Legal Assistant
-
Access comprehensive legal content with no limitations across vLex's unparalleled global legal database
-
Build stronger arguments with verified citations and CERT citator that tracks case history and precedential strength
-
Transform your legal research from hours to minutes with Vincent AI's intelligent search and analysis capabilities
-
Elevate your practice by focusing your expertise where it matters most while Vincent handles the heavy lifting
Start Your Free Trial of vLex and Vincent AI, Your Precision-Engineered Legal Assistant
-
Access comprehensive legal content with no limitations across vLex's unparalleled global legal database
-
Build stronger arguments with verified citations and CERT citator that tracks case history and precedential strength
-
Transform your legal research from hours to minutes with Vincent AI's intelligent search and analysis capabilities
-
Elevate your practice by focusing your expertise where it matters most while Vincent handles the heavy lifting
Start Your Free Trial of vLex and Vincent AI, Your Precision-Engineered Legal Assistant
-
Access comprehensive legal content with no limitations across vLex's unparalleled global legal database
-
Build stronger arguments with verified citations and CERT citator that tracks case history and precedential strength
-
Transform your legal research from hours to minutes with Vincent AI's intelligent search and analysis capabilities
-
Elevate your practice by focusing your expertise where it matters most while Vincent handles the heavy lifting
