Please use this identifier to cite or link to this item:
Title: Can we detect English proficiency through reading behavior? A preliminary study
Authors: Silva,IG
Carla Lopes
Issue Date: 2016
Abstract: If it were possible to automatically detect proficiency in languages using data from eye movements, new levels of customizing computer applications could possibly be achieved. An example in case is web searches where suggestions and results could be adjusted to the user's knowledge of the language. The objective of this study is to compare the reading habits of users with high and low English language proficiency, having in mind the possible automatic detection of the English proficiency level through reading. For this purpose, a study was conducted with two types of user, those with a high level of proficiency (Proficient Users), and those with low proficiency (Basic Users) in the English language. An eye-tracker was used to collect users' eye movements while reading a text in English. Results show that users with high proficiency engage in more careful reading. In contrast, low English proficiency users take more time to read, revisit sentences and paragraphs more often, have more and longer fixations and also a higher number of saccades. As expected, these users have more difficulties in understanding the text.
metadata.dc.type: conferenceObject
Appears in Collections:CSIG - Indexed Articles in Conferences

Files in This Item:
File Description SizeFormat 
P-00K-S7M.pdf844.7 kBAdobe PDFThumbnail

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.