Trials take time - even strip trials where you plant and harvest the whole length of the field. If the plot is set up right and the treatments are assigned at random, you find yourself driving back and forth, looking for where the next plot goes, or else changing certain settings back, and forth.
Yet Bob Nielsen, Purdue University corn specialist, tells farmers that without replication, what they really have is a demonstration plot. There's no way to run statistics if each treatment, whether it's a different hybrid or a different planting population, are only entered once in the same test.
To get answers, you need replication, and the more the better, says Dave Smith, Johnson County Extension ag educator. Smith, who earned a pHD degree, is trying to get farmers across the state, working with their local educators, to help conduct tests of twin row vs. 30-inchj row corn. By repeating the trial at different locations, it has the same effect of doing it multiple years, since weather is different at each location.
Smith says he ran into the same problem highly-skilled researchers sometimes run into. One farm took a big storm hit that interfered with some of the plots. Another cooperator misunderstood the setup and used different populations. So it's definitely a learning experience in trying to set up on-farm trials.
Without replication, you can't apply statistics. And without statistics all you have are numbers that make you scratch your head and wonder whether any trend you see is real or not. Perhaps it was just caused by experimental error, also known as chance.
The Indiana Prairie Farmer /Precision Planting trial at the Throckmorton Purdue Ag Center proved the value of replication this year, noted Jeff Phillips, Tippecanoe County Extension educator who helped oversee the plot. If the only rep was rep one, about 8 of the 27 treatments would have been wiped out.
"We had pounding and flooding and it turned corn in low spots yellow, he said. Due to all the rainfall, those spots leached out nitrogen. The lowest yielding one was 99 bushels per acre. Others in that rep were over 200 bushels per acre.
'We know that the treatment didn't cause all of what we saw," he says. "But if we had only done it once and didn't have the same treatments in three other places, we wouldn't have been able to tell how that treatment staked up."