On-Farm Research
Designing On-Farm Research Trials
It is often difficult to find production information for specialty crops that is relevant to the growing conditions found in Ontario. One way to overcome this lack of information is to conduct well planned on-farm trials, which begin by determining the goal or objective of the experiment. For example, are you interested in how a new crop performs when given different levels of nitrogen or water, or perhaps how planting density affects overall yield? Understanding what you want to achieve will help you design the appropriate experiment. The individual items you want to test are called treatments, and the sites you put them on in the field are called plots. The design of the experiment sets the treatment and plot locations in such a way as to allow you to make unbiased comparisons of the data collected from the various treatments, thereby allowing you to make informed decisions on your production practices.
Steps in Designing your on Farm Trials
- Determine the objective or reason for doing the trial
- e.g. Does planting density affect final yield?
- Determine how many treatments you need to achieve your objective
- e.g. Are three or four different planting densities needed?
- You may want to include both positive and negative controls. Positive controls are those that should always work, whereas negative controls should not.
- Decide on what data is needed, and when you need to collect it.
- Select the number of replications you need. The larger the number of replicates, the more trust you will have in the final results. Try to use at least 4 replicates
- Choose an appropriate plot size. It should be large enough to adequately mimic the growth you would see in a normal field. The size will depend on the spacing of the plants being grown and the equipment you have on hand. On the research farm we typically use plots 1 to 2 m wide and 5 to 8 m long, each containing 3 to 4 rows of plants.
- Select an appropriate experimental design. There are many different ways to lay out your plots within the field, but some designs are better at reducing the normal variations you encounter such as changing soil types or field drainage. OMAFRA specialists can provide more information on this.
- Determine the best way to lay out the experimental blocks within the field. Each block contains all of the plots that represent 1 replication of the experiment. If you have 4 replications, you will have 4 blocks.
- Randomize the plots within each block to reduce variations in the experimental data.
- Collect all of the necessary data. OMAF and MRA extension specialists may be able to help you with the statistical analysis of the data.
Below is shown an example of one type of experimental design that could be used to test how the growth of a new crop is affected by nitrogen fertilizer rate. In this experiment 4 levels of nitrogen (treatments) were applied, and the treatments were replicated 4 times for a total of 16 plots. Furthermore, each set of replicates was arranged in a block, for a total of 4 blocks, and the plots were randomly arranged within each block.
Simple visual observation of these plots over the growing season will provide some indication of the nitrogen requirements of the crop. An analysis of the final yield data will help determine where yield is not significantly improved with further additions of fertilizer. Since this type of information is missing for many specialty crops, the value of conducting these types of experiments comes from being able to obtain good crop yields without spending money on unneeded fertilizer.
