Statistical significance, agribusiness and tobacco grower profitability
Dr. Bill Easterwood, territory manager for YARA North America, offers this analtsis on using research data responsibily.
by David Williams
When outstanding research scientists from renowned universities make grower recommendations regarding fertilizers, their message is based on extensive field research and statistical data analysis. Unless a specific fertilizer treatment performs much better than others 95 percent of the time, the researcher will indicate that there is no “statistical difference” between fertilizers tested. Researchers follow proper scientific methods to evaluate products.
However, does statistical significance sometimes contradict known agronomic principles? How does this recommendation influence tobacco grower agribusiness? Are researcher statistical methods too sensitive for farm-based fertilizer decisions in terms of leaf yield and profitability?
Tobacco plants require rapid nitrogen uptake that terminates just before topping. Nitrate is the fertilizer of choice. “It has been clearly shown from numerous greenhouse and field experiments that tobacco will absorb both ammonium and nitrate nitrogen; however, when substantial quantities of ammonium nitrogen are absorbed, the growth of the plant is adversely affected—both in reduced yield and quality … it is suggested that at least 50 percent of the nitrogen in preplant fertilizer be in the nitrate form; in some cases a favorable response can be expected for up to 100 percent nitrate nitrogen” (Principles of Flue-Cured Tobacco Production, NCSU).
Recent nitrogen recommendations for flue-cured fertilization suggest using UAN or 24S solutions that have just 25 and 22 percent nitrate nitrogen respectively, with ammonium/urea as the predominant nitrogen forms. Researchers indicate that during the past few years of tests, the weather has been favorable for nitrification (microbiological process of converting ammonium nitrogen to nitrate nitrogen) and there was no statistical difference between fertilizer sources. Also, since ammonium fertilizers are cheaper than nitrate fertilizers, growers could save money.
Why should we disregard the important principle of nitrate nitrogen fertilization of tobacco based on the results of these “non-significant” tests? It is high-stakes gambling of leaf yield and quality in return for slightly lower fertilizer costs. Nitrifying bacteria populations are influenced by soil moisture, temperature, pH and fumigation. Unfavorable conditions inhibiting nitrification result in ammonium nitrogen as the principle nitrogen form in the soil, which reduces grower profitability. Also, ammonium can promote leaf potassium deficiency. Recently, growers in Wilson County, N.C., have observed potassium deficiency when they started using UAN solution.
Everyone likes a bargain. But saving money for fertilizer that doesn’t meet crop needs and can reduce your return on investment is no bargain. Do you put regular gasoline in your high-performance car?
During 2003-2005, six trials were conducted by NCSU comparing an all-nitrate liquid to UAN 30. The all-nitrate treatment increased yield by 112 pounds leaf per acre and was valued at $295 per acre more than UAN 30. Fertilizer costs per acre at today’s prices were $41 for UAN 30 and $109 for CN-9 or a $68 investment per acre for the nitrate form. There were no statistically significant differences between fertilizer treatments as determined by research scientists, but losing a net $227 per acre ($295 - $68) is highly economically significant to the grower. C.B. McCants and W.G. Woltz reported similar results in 1963 (Growth and Mineral Nutrition of Tobacco, Advances in Agronomy) where nitrate produced 12 percent more value per acre on fumigated soil than an all-ammonium fertilizer.
Should agribusiness decisions be based on a 95 percent statistical significance? Probably not. Successful profitable businesses today and in the past do not base their decisions on such a narrow protocol.