AmericanFarm.com

Nielsen urges growers to do their own research on products

By NANCY L. SMITH
AFP Correspondent

OCEAN CITY, Md. (Dec. 2, 2014) — Dr. Bob Nielsen of Purdue University wants you to quit believing everything you hear.
He says, “Frankly, today’s agricultural world is teeming with misinformation, half-truths, pseudo-data and sincere but incompetent researchers.”
Nielsen, an agronomy professor, advised farmers to decide for themselves the credibility of claims that are made for products or treatments.
Speaking at the 20th annual Crop Management School at the Princess Royale Hotel and Conference Center on Nov. 18, Nielsen advised farmers to ask what the presenters have to gain by you adopting their recommendation or buying their product. 
“Determine for yourself if the product claims include enough details to help you evaluate the validity of the research that supports their recommendation,” he said.
He told farmers to consider factors including whether the source of the information is public or private, whether the research was conducted in growing conditions similar to yours, what other treatments were considered and whether the data was statistically analyzed.
He cautioned, “There is hardly an agronomic decision you make that is black and white. It’s very challenging.”
He says farmers should critique the validity of trial design.
Look for details such as what was compared to what, how many trials were run, and whether there was a control scenario against which results could be compared, he said.
“Research answers questions,” he said. “The purpose of research is to predict whether documented responses to some treatment can be repeated.”
He also urged farmers to critique the validity of the statistical analysis used to support the claims made for the product or treatment.
“The challenge is to filter out the treatment effects from the ‘background noise.,”’ Nielsen said. “We use statistical analysis to mathematically identify and isolate background noise. It also calculates the probability of the conclusions being true.”
He advised against falling into the trap of relying on data that is not statistically significant. “If the statistical analysis shows the effects are due to background noise, you must conclude the true difference between treatments is actually zero.”
He also urged a critique of the interpretation of the results.  “If it says a product is ‘10 bushels better,’ ask ‘better than what?’”
Nielsen said these steps will help farmers learn how to distinguish between factual agronomic information and “crap-tual” agronomic information.