We report on a program of original eye-tracking research designed to illuminate how people interact with forms and help designers create forms that both expert and non-expert (at-risk) users. Our studies are designed to answer the kinds of questions that practitioners answer, but that are incompletely addressed in the academic literature, such as Where should the instructions be placed? What type of feedback works best, really? Should long forms be one page or several? How should I gate them? Which icons do expert and non expert users recognize for help? When must we use words instead of icons? When do/which graphics increase comprehension? Critically, our findings provide strategies to help both expert users and at-risk populations (e.g., low literacy, older adults) complete complex forms correctly and confidently.