[Online surveys and the false illusion of the large n. On a survey on euthanasia among medical practitioners]

The use of studies based on online surveys has expanded significantly. Despite having particularly small response rates, they allow a large sample size to be easily obtained. However, this strategy may entail a selection bias that significantly compromises the results. The results of two surveys on the regulation of euthanasia and assisted suicide are compared. One is an online survey with a self-selected sample and the other a survey with random sampling, conducted in 2018 among the members of the Medical Association of Bizkaia. The response rates were 10.4% (online survey) and 87.8% (random survey). No differences were found in sociodemographic characteristics, although there were differences in the opinion variables, so that the percentage of people who opposed euthanasia regulation was overestimated. The results of this study show that this sampling strategy generates biases in the results, some of which are difficult both to detect and to repair.