Characteristics of web-based and paper-based respondents in The Danish Capital Region Health Survey

Abstract Background and aim This study aims to investigate the potential impact of adopting web-only surveys on the representativeness of survey participants by examining the sociodemographic characteristics of participants in the Danish Capital Region Health Survey (DCRHS) who respond online versus those who respond using a paper-based questionnaire. Methods The study utilizes data from the DCRHS including participants invited by a mixed-mode approach and given the choice to respond to a web-based or paper-based questionnaire in 2021 (N = 53,874) and 2017 (N = 51,902). Information on age and sex was obtained from national registers. Results The proportion of paper respondents decreased significantly from 18% in 2017 to 16% in 2021. In both years, women had higher odds of responding on paper compared with men (OR = 1.20; 95% CI: 1.16-1.24). In 2017, the highest prevalence of paper respondents was observed among participants aged ≥80 (29%) and 16-24 years (21%). In 2021, the highest prevalence was likewise observed among participants aged ≥80 (28%) but with the second highest among the 65-79 years age group (17%). The decrease in paper responses from 2017 to 2021 varied across age groups, with a greater decrease among younger participants than among older participants. Conclusions This study highlights the potential risk of excluding a significant proportion of elderly citizens from large population surveys if the web-based response option is the only option provided. Elderly individuals are less likely than younger individuals to switch to digital responses, emphasizing the importance of considering alternative modes of data collection to ensure representativeness in survey samples. Key messages • The use of web-only surveys has increased; however, we risk excluding a significant proportion of elderly citizens from large population surveys if only the web-based response option is provided. • Elderly individuals are less likely to switch to digital responses, emphasizing the importance of considering alternative modes of data collection to ensure representativeness in survey samples.