My research proposes a set of Web Accessibility/HCI Guidelines to support web developers and designers deliver accessible web applications for people with Autism, specially children. To evidence the effectiveness of the guidelines, I'm planning a study with children with autism using a functional prototype, however, I have several concerns about the right approach to be taken into account in this study: I considered using Concurrent Reinforcement Schedule (CRF), a behavioral analysis technique, to identify children's preferences, but I think it may not be the best choice; I also considered ethnographic study, cognitive walkthrough and correlational study, but literature has little clues about that. Most works focus on evaluate the interface, while I'm willing to evaluate the guidelines applied to the interface and their interference on children's interaction.