Self-report is a fundamental research tool for the social sciences. Despite quantitative surveys being the workhorses of the self-report stable, few researchers question their format--often blindly using some form of Labelled Categorical Scale (Likert-type). This study presents a brief review of the current literature examining the efficacy of survey formats, addressing longstanding paper-based concerns and more recent issues raised by computer- and mobile-based surveys. An experiment comparing four survey formats on touch-based devices was conducted. Differences in means, predictive validity, time to complete and centrality were compared. A range of preliminary findings emphasise the similarities and striking differences between these self-report formats. Key conclusions include: (A) that the two continuous interfaces (Slide & Swipe) yielded the most robust data for predictive modelling; (B) that future research with touch self-report interfaces can set aside the VAS format; (C) that researchers seeking to improve on Likert-type formats need to focus on user interfaces that are quick/simple to use. Implications and future directions for research in this area are discussed.