Even nowadays, the use of computers can be diffcult for some people. The reason for that can be complex and multifactorial, such as lack of experience in using computers, lack of knowledge, problems related to usability and accessibility, fear, intimidation, or anxiety. Some people can experience multiple levels of anxiety and, for high ones, they manifest what is known as Computer Anxiety. People with Computer Anxiety (PwCA) face problems when using computers or technology at home, in the workplace, or for study purposes, which might result in multiple forms of barriers. This work reports a study that involved a pilot (4 elderly participants) and an experiment (39 elderly participants), contributing with a model that supports the identifcation of different levels of Computer Anxiety through the analysis of detailed interaction logs. Results show it is possible to automatically identify levels of Computer Anxiety in older adults while interacting with the web with 83.87% of accuracy. From the presented results, it is expected that specialists consider exploring the use of detailed interaction data and the presented model to identify Computer Anxiety levels, at scale, offering personalization, user interface simplifcation, removal of barriers, among other features to support this population.