Response rates to the Federal Waterfowl Hunter Questionnaire Survey (WHQS) have declined since the 1950's, suggesting that harvest estimates may be biased. Consequently, we investigated reasons for WHQS nonresponse using surveys of waterfowl hunters in Arkansas, California, Louisiana, Minnesota, New Jersey, and Texas [USA]. Sampling frames were constructed using lists of buyers of state hunting licenses or state duck stamps. We mailed questionnaires to 16,452 randomly selected hunters, with 2 follow-up mailings at 3-week intervals. Questionnaires were completed by 8,812 respondents, and a further 587 interviews were conducted by telephone. Post offices accounted for between 53.7% (Minn.) and 92.8% (N.J.) of federal waterfowl duck stamp sales, and stores accounted for most other sales. Of hunters who bought a federal waterfowl stamp from sample post offices, between 16.7% (Minn.) and 40.0% (Ark.) reported receiving a WHQS contact card. Of those receiving contact cards, between 30.0% (N.J.) and 64.3% (La. and Tex.) reported returning them. Because survey coverage of the target population is poor, we recommend that a new sampling frame be developed for the WHQS.