With the rise in smartphone and other mobile device usage, survey organizations like the Census Bureau have begun rethinking their Web- survey screen designs. Studies in the market research sector are finding that, on average, 25% to 30% of respondents are choosing to do online surveys using their smartphones, depending on the population being studied. Hence, optimizing for mobile screens is now a priority. Research on the American Community Survey, a 40 minute mandatory survey sponsored by the Census Bureau, has shown that although the percentage of survey responses which come in via a smartphone or other smaller device is still relatively small, the characteristics of those who report via mobile devices is different – they are younger, more likely to be a minority and less educated (Horwitz, 2014). Because the American Community Survey has not been mobile optimized, there are more break-offs when completing the survey on a smartphones. Differential breakoffs by population groups could have measurement error consequences for the ACS. This presentation focuses on results of the smartphone usability testing of the current non -mobile optimized ACS. We conducted 30 interviews with participants in the Washington, DC metropolitan area. They all accessed and completed the current non-optimized ACS on their smartphone or tablet. During the testing, we observed possible errors made while completing the survey - both in terms of changes to the data reported and edits. We then debriefed participants after they complete the survey. This presentation focuses on the major challenges users encounter such as, screen size, the answer interface (aka the fat finger problem), and slow download speeds. Usability results focus on the characteristics of the screens which lead to the greatest potential measurement error during testing.