The effects of position, context, and order on consent to data linkage in a survey, by @jburton123 Mick P. Couper & @AnnetteJackle - uses @usociety COVID-19 study. https://www.understandingsociety.ac.uk/research/publications/547726
We find that order of the requests does not make much difference to whether or not respondents consent, but making the requests early in the survey significantly increases consent rates compared to asking them after a series of content-related questions or later in the survey.
The most frequent reason chosen for not consenting was “too personal, shared enough information” – especially if they were asked for consent at the end of the survey. This finding points to some form of fatigue, as does the effect of the placement of the consent request we find.
Our finding that trust is selected more often as a reason for consenting by those who got the consent questions early suggests that the prevailing belief that trust is built up over the course of a survey is not empirically supported in our study.