As respondents increasingly respond to Web surveys on their smartphones instead of their personal computers (PCs), one finds that respondents take longer to complete Web surveys when using phones than PCs, and longer completion times may increase the risk of breakoff, but little is known about what drives these differences. In the current study, we compare response times for 895 respondents who completed the same Web survey once using a phone and once using PC as part of a methodological experiment in the LISS panel (Longitudinal Internet Studies for the Social Sciences), a Dutch probability-based Web panel. We analyzed survey-level times, item-level times, self-reported information about where respondents completed the survey and whether they were multitasking or distracted at the time. Linear mixed models are used to account for paired observations and questions nested within respondents. We focus on three potential factors: item characteristics, respondent characteristics, and respondent context. We expect that longer and more complex items will have a larger effect on mobile Web completion times. We also expect that respondents with less familiarly or comfort with smartphones, and those who are older, will be most responsible for longer response times in mobile Web because of their potentially decreased visual-motor coordination and lower levels of familiarity with mobile browsers. Finally, we expect that distracting contexts should inflate response times, and responding in these contexts is more likely when using smartphones. The results will help shed light on the device differences in completion times and inform strategies to reduce such differences.