Multiple regression techniques were used to obtain prediction equations relating wheat response to applied nitrogen, soil moisture, rainfall, soil nitrogen content, variety, and seeding rate. Yield variability could be significantly related to these variables in both central and western Kansas. However, coefficients of determination were approximately 0.9 for western Kansas and 0.7 for central Kansas. All variables considered were important in accounting for variability in yields. Crop variables contributed most to prediction followed by moisture and soil nitrogen variables. Applied nitrogen was significantly related to yield, however, it contributed least to the prediction equations. (Author)