You are here
Conduct Your Own Field Trial
By Kacey Birchmier
Like most farmers, you’re inundated with new products and practices to try on your farm. With commodity prices tightening, it’s imperative to know which products increase yields and which ones are economically viable.
On-farm trials allow you to gather data on your own acres and management practices. Trials add another layer of information to aid in the decision-making process. While the data you collect can be very valuable, trials require intense planning and management.
Bryce Caple, a farmer from Maxwell, Iowa, believes in conducting trials on his own farm. Caple has used trials to evaluate fungicides, in-line ripping, and N rates. He conducts field trials to have firsthand experience of new practices in his operation, instead of committing to a new practice or product before knowing how it will work on his farm.
“Conducting on-farm trials allows me to evaluate what works best on my soil types and field locations,” says Caple.
For example, he wants to set up a plant population trial this spring to see if it’s economical to increase population.
How to conduct a trial
If you want to evaluate a new practice, there are several steps you need to follow to ensure that you get valid results.
“The first step is to decide what you are trying to test,” says Richard Davis, a USDA-ARS research plant pathologist. “If you want to find out if some new pesticide works as well as one that you’ve been using for years, you need to compare the two.”
To set up your own trial, Davis recommends you follow these three steps.
- Choose a goal. The objective you choose will help decide how to run your trial, says Davis. All other factors hinge on this choice. For Caple, his goal is to see if it’s economical to increase his current plant population.
- Decide on a treatment. This is the practice or product you want to test. You can only decide on a treatment after you have a clear goal.
It’s best to evaluate one type of practice or produce per trial, rather than to combine multiple types into a single trial, says Davis.
“You can test things together, but it gets to be more complicated,” says Davis. “It’s probably better to test things separately.”
- Include controls in the trial. This part of the trial will allow you to make side-by-side comparisons. At the end of the trial, compare all of the data to decide if the treatment is justified and economically viable. With certain trials, there will be visual changes in the field. Visual assessments can be added to the yield data at the end of the trial.
“Usually with research trials, it’s important to have a positive and a negative control,” says Davis.
For example, to compare a new herbicide with the one you currently use, you would have three components:
- The treatment. This is the new herbicide.
- The positive control. This would be your old herbicide. This allows you to compare the two.
- The negative control. You leave a check with no herbicide. This negative control will let you know if the treatment is better than doing nothing, says Davis.
Next, you need to decide on the trial type – small plots or replicated strips. This comes down to personal preference, says Davis. Caple uses replicated strips because the design works best with the size of his equipment.
“Generally speaking, the test area you are going to collect your data from should be fairly uniform,” says Davis.
If you want to manage weeds or a disease, you want the pest pressure to be uniform throughout the trial area. If the pressure isn’t uniform, results may be skewed.
Researchers use small plots because they are easier to control and monitor. Farmers often prefer field-long strips because it’s easier to apply treatments. Small plots and strips usually work well for testing tillage, nematicides, and herbicides. Mobile insect or disease tests may require larger plots or wider strips, says Davis.
Randomly place treatments
Randomization means placing treatments in a different order throughout a field. “There could be an effect from one edge of the field to the other,” says Davis. “Whatever that effect may be, it may unfairly make one treatment look better or worse than it should. If you randomly assign them, it removes that possibility.”
Replicate the treatments
Replication or repeating the treatments is critical in this process. You need multiple replications due to potential variation within the trial area. Davis recommends using at least four replications in your trial, but the more replications you use, the more useful the information you glean will be.
“You can’t have one strip with a treatment and one that is untreated,” explains Davis. “You need multiple sets to compare. With anything you measure, there’s going to be variability.”
At the end of the trial, you are going to look at the overall effect of the treatments. One replication won’t provide enough information to make a decision, says Davis. Multiple replications allow for the chance that one of the replications will need to be discarded due to error.
Another important part of running a field trial is collecting data. You need to determine what data will be collected and how it will be collected during the planning process.
“Technology makes it easier to track the treatments,” says Caple. “It makes it easier to get the data, and I know where everything was planted in my trials.
“I’m constantly receiving data from companies on their trials promoting products,” says Caple. “Until I’ve tried it on my farm, under my management practices, I can’t be convinced it’s profitable for me.”
Collect data in an unbiased manner
It’s important to collect data in an unbiased fashion. Still, you need to consistently collect data. With most trials, you can predict the data type you will collect, says Richard Davis, a USDA-ARS research plant pathologist.
“If you are collecting data on a leaf spot disease, you know ahead of time you want to collect leaves and then count the leaf spots as part of the data,” says Davis. “You also know that you are going to want to collect yield data.”
Good documentation can help you identify why you saw reduction in disease pressure or tell you the cause of yield reduction. In order to be confident in your results, it’s best to not know where the different treatments are assigned throughout the field. Try to avoid the method of writing the treatment on a stake and placing it in the field. You may accidentally let that influence your data collection. Instead, try using a numbering system with a key.
“You need a system that helps prevent unintentional bias,” Davis points out.