Using Visual Analogue Scales in eHealth: Non-Response Effects in a Lifestyle Intervention

Background Visual analogue scales (VASs) have been shown to be valid measurement instruments and a better alternative to Likert-type scales in Internet-based research, both empirically and theoretically [1,2]. Upsides include more differentiated responses, better measurement level, and less error. Their feasibility and properties in the context of eHealth, however, have not been examined so far. Objective The present study examined VASs in the context of a lifestyle study conducted online, measuring the impact of VASs on distributional properties and non-response. Method A sample of 446 participants with a mean age of 52.4 years (standard deviation (SD) = 12.1) took part in the study. The study was carried out as a randomized controlled trial, aimed at supporting participants over 8 weeks with an additional follow-up measurement. In addition to the randomized questionnaire, participants were further randomly assigned to either a Likert-type or VAS response scale version of the measures. Results Results showed that SDs were lower for items answered via VASs, 2P (Y ≥ 47 | n=55, P=.5) < .001. Means did not differ across versions. Participants in the VAS version showed lower dropout rates than participants in the Likert version, odds ratio = 0.75, 90% CI (0.58-0.98), P=.04. Number of missing values did not differ between questionnaire versions. Conclusions The VAS is shown to be a valid instrument in the eHealth context, offering advantages over Likert-type scales. The results of the study provide further support for the use of VASs in Internet-based research, extending the scope to senior samples in the health context. Trial Registration Clinicaltrials.gov NCT01909349; https://clinicaltrials.gov/ct2/show/NCT01909349 (Archived by WebCite at http://www.webcitation.org/6h88sLw2Y)