Using survey responses of Illinois corn farmers to differently framed yield questions, we examine their subjective information by relating stated yields and risk to the corresponding objective county measures. The results show that farm-level yields can be best characterized by soliciting probabilistic information, which provides more accurate yield assessments than an open-ended frame and consistent estimates of producers- subjective risk. Moreover, we find that overconfidence can be confused with differences in relevant information and that using recent data may be more appropriate in examining subjective risk statements. Our results are important for agricultural policy-makers and researchers, particularly those who work with surveys that include questions about producers' yields.