Investigating recruitment and completion mode biases in online and door to door electronic surveys

Abstract Electronically assisted survey techniques offer several advantages over traditional survey techniques. However, they can also potentially introduce biases, such as coverage biases and measurement error. The current study compares the relative merits of two survey distribution and completion modes: email recruitment with internet completion; and door to door recruitment with either tablet or internet completion. Presentation mode is held constant so that we are able to separate the impacts of recruitment mode and completion mode on responses. Recruitment mode appeared to influence both response rates and which socio-demographic groups were represented. However, the difference between the two recruitment modes was relatively small. Completion mode appeared to have little or no impact on responses although it did influence completion times. The email distributed survey performed better with regard to time costs and the number of respondents obtained. Differences between the two survey modes appeared to be largely due to recruitment method rather than completion mode.