Making changes on the farm can be costly and risky. How do you get the maximum return for minimum risk? Simply put: You have to evaluate it to see if it works. At the Central Peace Conservation Society, we’ve been conducting unbiased trials since 1988. I often tell farmers that good, bad or otherwise, the results of our trials are simply what they are.
Companies frequently set up trials and insist on having total control over how the trial is conducted. If a company is expecting you to buy their products and cannot provide evaluation trials conducted by a third party, then run and run fast.
Having conducted over 600 field-scale replicated trials over 20 years, I have come to the conclusion there are no shortcuts. There is a required protocol that needs to be followed for trials to be of value.
Single strip trials
I’m amazed how many farmers think you can do a trial by seeding these 80 acres this way, those 80 acres that way and just weighing a single strip at the border. While this is a whole lot easier than doing things the proper way with replicated trials, can you have any confidence in the results of a strip trial? No.
I suspect most farmers don’t understand how soil variability in a field affects results. Until yield mapping and yield monitors came along, most would have never believed how much yield could vary within a field.
I only first began to understand when one of my favourite mentors, Garry Coy, a regional soils specialist with Alberta Agriculture and avid pilot, used to take me flying with him. Flying over fields, especially my test plots, was an eye-opening experience. If have such an opportunity, you should definitely take it. It will be immediately clear how plant growth and consequently soils vary across a field.
Variability is the single biggest issue in field trials. If soils within the field, how do you know at harvest time if the differences in yield are due to treatment differences or are the result of inherent soil variability?
You can’t answer that question if all you’ve have done is a strip trial.
Field trials — protocol
When farmers tell me they’re doing field trials, I like to ask a couple of questions.
“Did you calibrate your yield monitor?”
“Well no,” they’ll answer. “But we’re just looking for relative differences so if the yield monitor’s not calibrated, it’s not very important”. That might be okay.
My next question is, “Did you take a sub-sample?”
The farmer often looks confused at this point.
“Did you collect a representative sample from the combine or cart taken when unloading?”
The answer is typically, “Why would I do that?”
When we use a yield monitor or the scale on a cart we have the gross weight. This needs to be adjusted for moisture and dockage to give us the true net yield.
Many farmers would be surprised at how the moisture or dockage content from a set of samples can vary.
I also like to get the per cent protein and per cent green seed (depending on the crop) along with grade and bushel weight. These factors are often important when performing an economic analysis on the trial.
I collect sub-samples from my plots and place them into plastic zip-lock bags, evacuating the air and sealing them to preserve the moisture content. During the plot harvest season I periodically take the samples, marked with only a number, to my local elevator. While it seems minor, this is actually the most important part of the process.
Later I input the date into an Excel spreadsheet and further analyze it by conducting a statistical analysis of the results. (Many programs are available. I use one called Costat.)
There’s not a lot of value in looking at trial results if there’s only been one strip trial, rather than several replications. Each individual replication can have results that are very different that what you might see from one strip. Moisture and dockage can also vary — this only becomes apparent when you’re able to look at several replications.
It comes down to a simple point: Do you want to be making management changes on your farm based on the results of a single strip trial? Or do you want to have confidence that when you make a management choice, it’s the right one?
For me, it comes down to this: If you just want to make farmers aware of the name of a new crop variety or new herbicide, a strip trial is fine. But if you want to provide empirical data — things such as differences in yield, protein, green seed, maturity and so forth — the trial has to be replicated.
When you’re attending an industry meeting and they start to flash numerical data of any type on the screen, put your hand up and ask, “Is this data from a strip trial or a replicated trial?” If it’s not from a replicated trial, then it would be just as accurate to roll a dice and rank the treatments that way. If you go to this kind of meeting, I hope the meal is good, the door prizes expensive, and it doesn’t cost you anything to attend.
But don’t despair. Help is available. In Alberta we are lucky to have a number of agricultural research groups that can help you set up properly designed trials that will give you solid, trustworthy results. If you don’t know who a group in your area, contact the provincial umbrella organization, ARECA (Agricultural Research and Extension Council of Alberta) at 780-416-6046 and someone will steer you in the right direction. †