A research study was conducted during the 1991 Farm Costs and Returns Survey (FCRS) to identify and classify the reasons given to field interviewers by potential respondents who refused to participate in the survey. The reasons given by field interviewers for coding a sampled unit as inaccessible during the survey were also identified and classified. The research was conducted in all 48 surveyed states, and included 6 FCRS questionnaire versions. Upon receiving a refusal, interviewers were instructed to record the reason given for refusing on the face page of the questionnaire. If no reason was given, or in cases where more than one reason was given, the interviewers were instructed to discuss the concerns of the respondent in regards to completing an interview, and to identify the main reason for refusing. When a sampled unit was coded as inaccessible, interviewers were instructed to explain the reason. During the manual edit of the questionnaires, the reasons for refusal or being coded as inaccessible were reviewed and compared to a coded list of reasons for nonresponse compiled from previous research into this topic on the FCRS. statisticians could consider the comments from the interviewers as a match to a pre-coded response, or add additional codes for unique comments. The nonresponse rate on FCRS averages 30% per year. The reasons behind the nonresponse have been a source of speculation for many years, and previously only anecdotal evidence was available on which to base efforts to maximize response. This research shows the anecdotal evidence to have been sometimes on and other times off the mark.