Files
Abstract
Straw man model 1 is one of the simplest regression models which can be used to predict crop yields. A one-line regression of yield over time, it represents the increases in yield which have occurred through the adoption of improved varieties, and the increased use of fertilizer and other cultural practices. The performance of this model in predicting soybean yields in Iowa, Illinois, and Indiana is evaluated. Eight model characteristics are discussed. Indicators of yield reliability obtained from bootstrap testing show that the bias is generally small for this model. However, the model is unable to predict the low and high yields accurately. The model is objective, adequate, timely, simple, and not costly. However it does not consider known scientific relationships and does not provide a good current measure of modeled yield reliability.