Using information about the respondent to tailor data collection to the respondent in responsive and adaptive survey designs has been shown to reduce survey error. Allowing interviewers (or web surveys) to tailor question wordings, instructions, or definitions to respondents in conversational interviewing has also been shown to improve data quality. However, tailoring has not been explored in depth in the presentation of confidentiality language prior to beginning the data collection. In this paper, we describe a conceptual model, drawn from the literature and suggested by qualitative evidence. This model illustrates how respondents arrive at their level of assurance about confidentiality and specifies how tailoring confidentiality pledge information may increase the number of respondents who feel assured. The model illustrates that a simple and direct confidentiality statement does assure some respondents, but others need additional information to be assured. However, providing respondents with additional information can also be counterproductive, leading to concerns where there were none or heightening already present concerns. The implication is that to achieve assurance and ultimately gain a completed interview, survey practitioners need to tailor their responses to individual concerns about confidentiality.