Selection bias in online surveys in veterinary medicine—review of reporting practices and recommendations
Online surveys are popular, also in veterinary medicine. However, it is challenging to obtain results that are valid not only for the participants (study population) but also for the target population due to selection bias. This work aims to (I) summarize guidelines for survey reporting on recruitment, representativeness, and selection bias, (II) review current practices, and (III) discuss the gap between recommendations and practices. Four guidelines have been identified. Even though these do not specifically refer to veterinary medicine, they provide helpful standards. They indicate that, inter alia, a detailed description of the recruitment process, a definition of the target population, and calculations of the response rate (RR) are required. In the methodological review on the reporting practices, 62 studies published in Frontiers in Veterinary Science from 2023 to 2025 have been included. They displayed a broad range of topics, target populations, recruitment strategies, and participant numbers. While nine studies (15%) included responses from fewer than 50 individuals, nine other studies included responses from more than 1,000 individuals. Nearly 70% of surveys were open, meaning that researchers had no list of (potential) participants. In 19 studies (31%), an RR (defined as the percentage of participants/invited persons) was reported, and it was possible to calculate it for an additional five studies. The median RR of those 24 studies was 28% (median; IQR: 10–62%). In 21% of all reviewed studies, comparisons between the study and target populations were made, and in 66% of the studies, selection bias was discussed. However, it should be kept in mind that, due to the diversity, not every criterion was relevant to every study. Results indicate a gap between practices and guidelines regarding selection bias. Recommendations of the guidelines are difficult to meet when little is known about the target population and the persons invited, which was often the case, especially in open surveys. Nevertheless, one has to wonder how valid the results are if a study is based on a selection of participants. Therefore, future studies should develop recruitment strategies to reduce selection bias, conduct sensitivity analyses, and discuss how selection bias might have affected their results.
Authors
- Katharina Charlotte Jensen (ORCID: https://orcid.org/0000-0002-4656-4464)
- Roswitha Merle (ORCID: https://orcid.org/0000-0002-8688-2926)
Institutions
- Freie Universität Berlin (DE)
Publication Details
- Journal
- Frontiers in Veterinary Science
- Published
- 2026-11-18
- DOI
- https://doi.org/10.3389/fvets.2026.1993195
- Primary Topic
- Survey Methodology and Nonresponse
- Type
- article
- Field-Weighted Citation Impact
- 0.00