Library Categories

Why a single test plot isn’t enough

By TOM J. BECHMAN

The age-old divide between true research and on-farm trials still exists. Experts trying to prove beyond a shadow of a doubt that one hybrid, one variety or one practice is better than another want as many repeats of the same trial as they can get. It’s about probability and statistics.

They want to be sure that yield differences they might find are because of how they treated the plot, and not just due to chance, perhaps caused by differences in soil type or a host of other factors.

Balance that against farmers trying to get their crop in the ground. It’s what they get paid to do. And repeating trials, especially small plot trials, takes time.

Key Points

• Doing each treatment once is a demonstration, not a trial.

• Soil types or other factors can cause numbers to flip-flop.

• Ask questions about how a trial was done before drawing conclusions.

Still, they’re curious about what works on their farm. They hear it repeatedly — test it on your own land. Only trying say five populations in five passes of the planter isn’t proof. It may be an indication that there could be a difference, and it may just be a reflection that soil type changed.

“That’s why we like to have several replications of the same trial, preferably four, when we’re doing tests,” says Dave Nanda, director of genetics and technology for Seed Consultants Inc. “The more tests you do, the more confidence you can have in the results and conclusions.”

One example

Here’s an example from the Indiana Prairie Farmer/Precision Planting trial that compared seed depth, downforce placement and planting speed. Looking just at downforce, look at these two sets of numbers, and see what conclusions you would draw.

If you had only planted one rep and you happened to plant it on the same ground as replication 3, you would be ready to crank up the pressure on the planting units. By the same token, if you only planted rep 4, you would back off that pressure and look for a medium-pressure setting.

Note that if you only had reps 3 and 4 in your field, the edge seems to be to lightening up the pressure, although it’s not significant. But since all four reps were in the field, in the end all three pressures produced yields within 2 bushels of each other. None were significantly different. The moderate setting actually resulted in significantly better spacing than the light setting, but it didn’t translate into higher yield.

“Be very careful about drawing conclusions,” Nanda says. “It’s good to do on-farm trials, but you need to repeat them in different environments to draw conclusions and make changes based on them.”