Exploring the usability of the text-based CAPTCHA on tablet computers

This paper analyses and discusses the usability aspect of the text-based CAPTCHA in terms of response time and success in solving the CAPTCHA on tablet computers. The response time is the time spent by the user to find a solution to the CAPTCHA. The analysis is separately conducted on text-based CAPTCHA with only text and numbers. Then, the results are compared and the differences in response time and success in solving the two types of CAPTCHA are underlined. This is accomplished by asking 125 Internet users to solve the text-based CAPTCHA on the tablet computer. Their gender, age, education level, Internet experience, response time and success in solving two types of text-based CAPTCHA are collected in a dataset. Then, advanced statistical analysis by association rule mining is performed. It shows the dependence of the response time and success in solving the CAPTCHA on co-occurrence of gender, age, education level and Internet experience and the strength of this dependence by support, confidence and lift measures. This study provides relevant information for designing new CAPTCHAs which may be more accustomed to specific types of Internet users.