Four research principles for reliable field trials on your farm this season

When a farmer hitches up a planter or climbs into a combine cab, a flood of adrenaline fuels his Mission From On High. Full speed ahead! No time for flagging test strips or recording weigh-wagon samples. (I'm just acknowledging reality!)

February 26, 2021  By Jerry Carlson — Based on 30 years of encouraging practical field research with biological products and ideas, I've noted that the largest-acreage growers are usually the least willing to pause the planter to change products for test strips. The most diligent researchers in Midwest row crops are typically those farming 500 to 2,000 acres. Big enough to run GPS systems on planters, sprayers and combines, and keen on widening profit margins with innovative management.

Most of the biological innovations flowing into farming offer modest yield gains: 4 to 10 bushels of corn, and only roughly 80% of the time because field conditions and weather dominate results. Thus simply finding which bug-in-a-jug really works requires replicated field strips and standard research protocols. Too often we hear, "I sprayed that stuff on half a field, and didn't see a difference." No matter how uniform the soil map looks, variations are often so wide across several acres that a single application can easily mislead.

Thus we've long urged growers to use multiple replications and follow four basic field research principles. We've used those standards in field research and greenhouse tests for a decade. Our grandson Lane Carlson, 12, wrote one of the simplest summaries of those rules of the game. He did this after sharing in several of our greenhouse experiments. Here's his summary:

Lane Carlson

Principles of biological research 

By Lane Carlson

1. Test only one variable in each experiment. Keep it simple. In our experiments, we tested how corn plants responded under drought stress with — or without — a treatment with the endophyte combination BioTango/BioEnsure (BE/BT). When you start adding more variables to an experiment, interactions make the test very complex.

2. Use identical test subjects to eliminate as much random variation as you can. In our experiments, we planted seed of the same hybrid at the same time, and watered each potted corn seedling with a precise 60 ml. of water at the same intervals. We planted into normal soil from our farm, so results would be relevant to field conditions. Soil conditions across a field are a powerful source of variability, so our soil was blended uniformly. 

3. Replicate the test subjects as many times as you economically can in each experiment. Minimum: Four replications. Eight is better, to find statistical significance. If you had two plants, one was treated with beneficial bacteria while the other was not — and one died ­— how much information would that give you? Not much! In our drought “torture” tests with four corn seedlings, one untreated plant out of four sometimes survived while three died. That’s a meaningful signal. 

4. Keep accurate records.  Make notes so complete that anyone can read your experimental protocol and repeat the same experiment. This information is very important because (a) it’s easy to forget what you did and when; (b) somebody is sure to ask “How did you get that result?”



We've adapted a field trial protocol using replicated strips, based on Practical Farmers of Iowa recommendations. Here's a background story on PFI's system

The image below is an example.  You can also download a higher-resolution PDF version at this link and print it out as a "keeper" guide to follow when setting up experiments in the field.