Farmers conduct a wide range of experiments on their farms every year. Most are done for the farmers’ own information and don’t require the “blessings of science.” However, a few simple techniques from scientific methodology can give an experiment’s results more credibility, making it easier to share with scientists, academics and skeptical neighbours.
The question
The first step in a scientific investigation is to clearly form the research question. For instance, a broad question such as “would compost tea help my crop” needs to be focused before the research project is set. A more specific question is often easier to answer.
Read Also
Agriculture needs to prepare for government spending cuts
As government makes necessary cuts to spending, what can be reduced or restructured in the budgets for agriculture?
Usually, it is a matter of narrowing down what you think you might find. Scientists call this the hypothesis. A specific question might be, “would five gallons per acre of a specific compost tea applied at the two-leaf stage of my chickpea crop reduce disease?” The question determines what needs to be measured to decide if the treatment worked.
It is always tempting to ask complex questions. Perhaps, in the compost tea example, the researcher may hope to see effects on weed numbers, numbers of insect pests, disease severity, sugar concentration, yield and seed size.
The researcher might want to compare one tea application to two or three applications, at different times and at different rates. If time and energy are limitless, these might all be useful ideas, but in most cases they can quickly overwhelm the experiment.
Science is an ongoing process. No experiment can include everything, but a simple one-comparison test can be extremely useful.
Controls
A control is pivotal in being able to make comparisons. If a field produced a bumper crop when the farmer harrowed, this does not necessarily mean that harrowing was beneficial. Perhaps it was just that sort of season.
To scientifically test the effect of harrowing, the farmer would need to have a portion of the field that was harrowed and a portion that was not as the control. The effect of harrowing is seen by comparing the harrowed area to the control area.
A control might also be the “usual” way of management – the old treatment – compared to the new. For instance, in a comparison of seeding depth, the meaningful comparison might be between seeding at a one-inch depth compared to seeding at a three-inch depth.
A meaningful control has to be as similar to the treatment as possible in every way except for the treatment. If the treatment is applied to the centre of the field and the control to the headlands, or the treatment on the knolls and the control in the valleys, the difference might be due to position as well as to the treatment. Landscape and field history are important to consider when deciding where to place a test.
Replication
A single occurrence may not mean much. If something happens in the same way several times, it is more likely to be meaningful.
This is the idea behind replication. A minimum of three replications is needed to do a statistical analysis. The more replications, the easier it is to find statistical differences between treatment and control.
The more replications there are, the more work is involved. Many scientists use four replications in field trials, which means the entire set of treatments is repeated four times in the field. Some recommend six replications for on-farm trials.
Strip trials
A strip trial is a simple way to make comparisons. Let’s consider the comparison between seeding at one bushel per acre or two bushels per acre.
If the seeder width is 14 feet, the field could be staked out in 14 foot (one seeding pass) or 28 foot (there and back) strips. Using four replications and two treatments would require eight strips across the field. Each pair of strips would be one comparison, or one replication.
The farmer could seed the one bushel per acre strip of replication 1, then the one bu. per acre strip of replication 2 and so on before recalibrating the seeder and seeding all the two bu. per acre strips.
Of course, a good map of the field, with the trial area well marked on it, and clear and obvious stakes, will make this pattern easier to follow.
Randomization
For each replication, it is important that treatments are in random order: they don’t just follow a pattern of one, two, one, two, one, two, one, two. This is to help avoid unintentionally favouring one treatment.
If the one bu. per acre treatment is always downwind, uphill or nearer the shelterbelt than the two bu. per acre treatment, this might unfairly influence the comparison.
Collecting data
Collecting data can be time consuming, but must be done carefully to be of value. Measures or assessments are generally taken for each treatment in each replication.
Sometimes several measures are taken in each strip. This can be especially useful if there is field variability. Samples should be taken in a pattern that is determined without looking at the crop. Taking samples at spots that “look good” can bias the sample.
Good field records can help make sense of confusing data. It is helpful to keep a record of all operations and any observations, such as weather, flooding, insects and deer.
These observations will be especially useful if results vary from year to year or if one replication seems to behave differently from the others. Careful planning and willing helpers make data collection easier.
Analysis
Sometimes the results of an experiment are clear cut. If each harrowed treatment has fewer weeds than its paired control, there is probably no need for statistics. More likely, though, the data will vary from replication to replication and even from sample to sample within a strip. In this case, statistical analysis can help. Statistics show how big a difference there is and how likely that difference is to occur by chance alone.
Repetition
An experiment may give different results in different years. Doubling the seeding rate may increase yield in a good year, but decrease yields in a drought. Repeating an experiment in a different year helps to show if the result is generally true or true only under specific circumstances.
Conclusions
Scientific experiments on the farm can be a way of testing new ideas or demonstrating interesting innovations. They can be the focus of farm field days. By doing these experiments themselves, farmers can make sure that the research they want is done in their environment and with their constraints.
By making sure that each test follows a scientific methodology, with at least a control and some replications, these experiments can form the basis of a credible network of farm-based information.
Frick is the prairie co-ordinator for the Organic Agriculture Centre of Canada located at the University of Saskatchewan. She can be reached at 306-966-4975, at brenda.frick@usask.ca, or www.organicagcentre.ca.
