A web-based diabetes “risk calculator” is being developed and evaluated to determine the impact of personalized risk estimates and interactive feedback on user attention and systematic information processing. Preliminary experiments that randomized participants to two different health websites suggested that a risk calculator with personalized risk estimates did not increase (and may have decreased) systematic processing, focused immersion and information seeking. We describe a series of think aloud user studies which were conducted to provide a qualitative evaluation of the experimental protocol and explore alternate explanations for these unexpected findings. User study results suggested that the prior findings may have been driven by a lack of perceived novelty of the risk information, selective attention, and an expectation of personalization in both experimental conditions. Findings are consistent with satisficing in information search and have implications for the design of health information and future experiments that evaluate these types of interventions.