In attempting to locate respondents who have experienced a particular event, surveys often have two parts – a screening section that qualifies individuals and a follow-up section to gather more detailed information from those qualified. In some studies these two sections can be separated by weeks or months, often requiring rescreening to ensure that respondents meet the qualification criteria. How we re-screen has the potential to significantly affect the proportion of people requalifying for the study. This web-based survey project recontacted screened respondents and asked about their purchases of a number of products within the past 3 years using one of three randomly assigned response formats – a yes-no grid (YNG), a multiple response format (MRF - ‗select all‘), or a combination grid (CG) that allowed respondents to indicate self, other, self and other, or no one had purchased the product. For both self and household members, we found that the MRF had the lowest requalification rate while the CG showed the highest requalification. We will further present how response patterns (e.g. satisficing and nondifferentiation) and need for cognition affect the effects of response formats.