On Users' Preference on Localized vs. Latin-Based CAPTCHA Challenges

A Completely Automated Public Turing test to tell Computers and Humans Apart (CAPTCHA) is a widely used security mechanism for constructing a high-confidence proof that the entity interacting with a remote service is actually a human being. Stimulated by the facts that: a) nowadays CAPTCHA challenges are solely based on the Latin alphabet, b) currently Internet population consists in its majority of non-native-English speakers and c) numerous web sites consist of exclusively localized content, we conducted an empirical study aiming to examine the effect of various factors on users' preference in solving localized vs. Latin-based text CAPTCHA challenges. The study embraced a between-subject design using a self-developed localized CAPTCHA mechanism, capable of producing text challenges based on the participants' native alphabet. A total of 384 non-native English speakers participated in the frame of the reported study which followed an ecological valid experimental design. Analysis of interaction results provides interesting insights which can be taken into consideration for designing more usable CAPTCHA mechanisms.