Relevance & Research Question: Sample surveys routinely ask for respondent consent to link survey information with administrative databases, but not all survey units agree to the linkage. Next to respondent socio-demographic characteristics also design factors of the linkage consent question has been identified to influence consent propensity (e.g., wording, placement). Some of these findings run contrary to conventional wisdom. For instance, some studies have found that asking for linkage consent near the beginning of the survey interview is more effective than asking for it at the end. Surveys have long administered the linkage consent question at the end of the survey interview under the assumption that this is the point at which respondent-interviewer rapport is likely to be at its highest. However, these studies have only been performed in household surveys administered by an interviewer. Whether placement matters in self-administered modes (e.g., web) and for other types of target populations (e.g., establishments) remains an open question. We present results of a placement experiment in a web survey of establishments in Germany and offer insights and contribute to “best practice” guidelines for maximizing online linkage consent rates for establishments.

Methods & Data: A representative sample of establishments was invited to participate in an online survey with focus on establishment hiring practices. Over 4,200 responding establishments completed the online questionnaire, which included a consent request to access and link administrative information from Germany’s Federal Employment Agency to their survey records. The consent experiment included three placement manipulations. Establishments were randomized to receive the linkage consent question either at the beginning, middle, or end of the survey interview.

Results: We show that the beginning of a survey is significantly the best placement for achieving a high rate of consent, followed by the middle-, and end-placement. The result appears to be robust across different subgroups.

Added Value: To our knowledge, this is the only linkage consent experiment that has been conducted on a sample of establishments, and the only study that has manipulated more than two placement locations in a web survey.