Crop yields of cereals, carrots, beetroots and potatoes from 28 different field plot and pot experiments (on a site near Marburg/Germany) were compared to determine the influence of the biodynamic preparations 500 and 501 on yields. Under generally low yields the preparations tended to increase the yields. When the yields reach a medium level this positive effect was smaller. At higher yield levels preparations tended to lower yields. For the yield effect of the preparations 500 and 501 (= Y) and the yield levels in the untreated control (= X) a significant linear regression could be calculated: Y = 4.497—0.181 X (r = −0.615; α < 0.01%). In another experiment with spring wheat conducted for 9 years in Darmstadt/Germany the application of all eight biodynamic preparations modified yields similarly, but only at a high fertilization level. In this treatment the effects followed a significant linear regression Y = 28.930—0.87 X (r = −0.767; α = 1.59%). Yields in the untreated control varied from 1.6 to 5 t/ha. These effects have previously been discussed as a normalization of yields or as a compensation of a non-optimal nutrient supply. A new model is suggested here based on the regressions of preparation effects and yield levels.