User Perception of CAPTCHAs: A Comparative Study between University and Internet Users
Arun Reddy, Yuan Cheng
TL;DR
The paper addresses the security–usability tension in CAPTCHAs by conducting a cross‑group online survey of university students and MTurk workers to gauge user preferences, difficulty, and perceptions of CAPTCHA effectiveness. It systematically analyzes multiple CAPTCHA types, device effects, non-English exposure, and abandonment tendencies using quantitative and qualitative data. Key findings reveal that distortion and readability drive difficulty, device size affects solving, and security perceptions differ notably between groups, with MTurk users more positive about usability and security than university participants. The study offers practical guidance toward more user-friendly CAPTCHA designs and motivates exploration of invisible or passive CAPTCHA approaches to reduce friction while maintaining bot-resistance.
Abstract
CAPTCHAs are commonly used to distinguish between human and bot users on the web. However, despite having various types of CAPTCHAs, there are still concerns about their security and usability. To address these concerns, we surveyed over 250 participants from a university campus and Amazon Mechanical Turk. Our goal was to gather user perceptions regarding the security and usability of current CAPTCHA implementations. After analyzing the data using statistical and thematic methods, we found that users struggle to navigate current CAPTCHA challenges due to increasing difficulty levels. As a result, they experience frustration, which negatively impacts their user experience. Additionally, participants expressed concerns about the reliability and security of these systems. Our findings can offer valuable insights for creating more secure and user-friendly CAPTCHA technologies.