Relevance & Research Question: Slider scales – made from a horizontal line and a handle being moved to give an answer – are frequently used rating scales in Web surveys. In contrast to simple radio buttons – where answers are given by (1) move the mouse to the answer option and (2) click – the use of sliders is more demanding: respondents have to (1) move the mouse to the slider, (2) click the handle, (3) hold the mouse button, (4) drag the handle to the answer option, and (5) release the mouse button. The research presented focuses on how slider scales influence the process of data collection.
Methods & Data: After a quite extensive study (median response time 14 minutes) respondents (_N_ = 1067) were asked three items about the study. In a 2 X 3 design, the rating scale (radio button vs. slider) was experimentally modified between respondents as well as the number of response options (3 vs. 5 vs. 7).
Results: Break-off was about three times higher with slider scales, especially for respondents with lower cognitive abilities. Furthermore, response times were considerably higher with slider scales. The most serious finding is that the distribution of values considerably differed between both scales. Regardless of the number of response options, there is evidence that a large share of respondents does not see the initial position of the slider being a valid response option. Thus, n-point slider scales are actually used as n-1-point scales.
Added Value: Slider scales are part of most Web survey software. This study provides evidence that this type of rating scale may not only increase respondent burden but also may seriously harm data quality as well as bias sample composition . The present results strongly advice against using this type of rating scale. Standard, low-tech radio buttons should be preferred. Further research on the impact of different implementations of sliders (e.g., initial position of slider, number of response options, and design) will be discussed.