Over 60% of U.S. adults now own smartphones (Pew Research Center, 2015), and smartphones are increasingly being used to respond to web surveys (Dillman 2015). Even with surveys that have been optimized for smartphone use, visual presentation of survey items on smart phones range from similar to very different than the presentation of the same items on a personal computer (PC), which has the potential to lead to mode effects. In this presentation, we examine visual layout symmetry and mode effects between smartphone and PC-based versions of a web survey, drawing on two web successive web implementations of the Residential Energy Consumption Survey (RECS). In both studies, the visual layout of the web instrument was designed to be similar across web and paper modes. Due to software constraints, however, visual differences between smartphone and PC presentations were more pronounced for question items (e.g. where a visual aid such as a photo was necessary) than others (e.g. standard radio button questions). Respondents were encouraged to respond via web in both studies, but free to response via an electronic device of their choosing. To understand any impact of device on response quality, we analyze differences in key survey estimates and item nonresponse rates between smartphone and PC modes, controlling for demographic differences related to mode selection. We classify formatting differences between smartphone and PC, and we report how these are associated with differences in survey response.