Effects of Video Enhancement in a Stated-Choice Experiment on Medical Decision Making

Background. The internet can be useful in administering stated-choice experiments to understand medical decision making and refine the content of patient decision aids. In internet-based stated-choice experiments, video and audio files can be used to provide information to respondents. Quality of data may or may not be affected.Objectives. In a methodological experiment concerned with administration of a stated-choice experiment on the internet concerned with knee-replacement surgery, we compared the data quality obtained with video-enhanced and conventional text formats.Methods. Members of the RAND Corporation’s American Life Panel and 50 years of age or older (n=1616) were randomly assigned to one of two survey modes: video (80%) or text (20%). Quality of data was measured by both the frequency of respondent speeding and plausibility of respondent choices.Results. Typical respondents completed the research task much more rapidly in the text condition than the video condition (10.6 versus 16.4 minutes). The choices were 5 to 10% more likely to be plausible for those in the video-enhanced condition than for those in the text version. The mode of administration did not affect respondents in the extent of their reported interest in the survey or in the degree to which they reported identifying with patients in vignettes.Conclusions. The advantages of the video enhancement in engaging respondents appeared to be offset by the greater time demands on respondents. Video-enhancements in stated-choice experiments may be most useful when used selectively to deliver information that cannot be delivered adequately by text.