News, Pulse Beat Individual Articles

On-Farm Research – by Farmers, for Farmers

Megan Bourns, On-Farm Network Agronomist, MPSG

IN TODAY’S ERA of high input costs, low margins and considering the ever-increasing need to improve sustainability of the farm operation, validation of agronomic management decisions made on-farm are ever-more important. Agronomic recommendations are usually generated by small-plot research, which can efficiently and effectively compare numerous treatments in the same location, at the same time. But what happens when those treatments deemed most effective by small-plot research are used at a field scale? Do they behave the same? Are they just as efficacious? Are they economical? On-farm trials can help answer these questions.


Manitoba Pulse & Soybean Growers (MPSG) began funding on-farm research trials back in 2010. In 2014, MPSG made a commitment to expand its on-farm research program throughout the province, and in April of 2014 MPSG’s On-Farm Network (OFN) was officially launched. The OFN is a network of pulse and soybean research conducted on-farm, fully funded and directed by pulse and soybean producers in the province. The goal of the OFN is to test new products and practices for pulse and soybean production while empowering farmers to conduct straightforward, reliable research on their farms. Three principles guide the OFN:

  • Participatory – actively engages farmers in the research process
  • Precise – OFN trials produce robust and statistically sound data
  • Proactive – results from the OFN guide management decisions, aiming to improve productivity and profitability of the farm operation

Since its beginning, the OFN has facilitated approximately 335 trials (Figure 1, Table 1). In the 2019 growing season alone, the OFN conducted 51 soybean trials, six dry bean trials and eight pea trials, covering a range of agronomic questions from the efficacy of fungicide applications in soybeans, peas and dry beans to soybean seeding rate, and several in between.



Conducting on-farm research through the OFN is a collaboration between the producer and the MPSG on-farm research team. On-Farm Network research is producer-centric; trials are selected, designed and implemented to address the questions of MPSG farmers. This means if, as a farmer, you come to us with a question about agronomic management on your farm or are curious if a new-fangled biological product your chem guy has been trying to convince you to buy for months actually works, we will help you design a trial to   find those answers. Working with MPSG through the OFN, you don’t just get trial design — we will be involved in every step of the process from trial setup, establishment, midseason data collection and harvest. We want to facilitate the type of first-hand research that will help you as a farmer make management decisions going forward.

An important aspect of collaborating to conduct on-farm research through the OFN is the ability to build large datasets to answer production questions across space and time. Many farmers have the same questions — for example, is a fungicide application in soybeans necessary? Or, what N rate should I put on my dry beans? Do I need to double inoculate after multiple years of growing soybeans? By conducting trials to answer these questions over a number of farms and a number of years, trends can be analyzed. In this way, on-farm research through the OFN not only enables individual farmers to see first-hand benefits of research conducted on their own farms, but it also facilitates answering larger questions by pooling data from several farmers together.


On-farm trials are large, field-scale trials established with field-scale equipment by the farmer. Unlike small-plot trials which have a lot of treatments in a small area, on-farm trials compare just a couple of treatments but over a much larger area. The difference in scale between small-plot and on-farm trials results in differences in background variability across the trial area (Figure 2). Small- plot researchers generally want very uniform site areas. However, in on-farm research we want to capture as much representative field variability as possible to determine how a treatment behaves at the scale it would be applied if it were to become normal farm management practice.

There are a couple of key considerations to make an on-farm trial useful. A specific question will help guide treatment selection, trial location and design. Because on-farm trials are testing treatments to see if they should be adopted on the farm, the treatments selected should be realistic to normal farm operations. Physical layout of the trial is also important. Although an on-farm trial can encompass variability across a field, the variability within the trial area should not unnecessarily introduce bias to the trial results. For example, locating a trial where a field drain runs the length of an entire strip could skew the trial results. Replication of treatments across the field is important to determine whether yield differences are actually a result of the treatment, rather than simply a result of field variability within the trial area. Randomizing the order of treatments within the replicates helps reduce any bias the natural field variability could introduce into the data. Figure 3 shows an example of a randomized and replicated trial layout. Collecting midseason data, whether it be plant counts, nodulation ratings, disease ratings or otherwise, depending on trial type, can help validate yield results at harvest.

Harvest data from each individual strip should be collected either with a weigh wagon, grain cart with a scale or using a calibrated  yield monitor. Once all the data has been collected, statistical analysis is required to determine if any numerical differences in yield are actually a result of the treatment applied.


On-farm tests are not the same as in-field demonstration plots. A demonstration generally consists of individual strips of varieties or products, showcased side-by-side, that are not randomized and replicated. Demonstrations can exemplify visual differences between varieties or certain products but cannot be used to determine the efficacy of one over another. When it comes to a question of what treatment, variety or management practice results in more yield, a replicated, randomized on-farm trial is the way to go.


At the end of the day, on-farm research is done by the farmer, for the farmer. Well-conducted on-farm trials investigate questions  and outcomes on a case-by-case basis while evaluating the overall effects of management decisions through combining data across trial locations and years. Facilitating trials to generate meaningful results is a balance between our efforts and producer efforts. For producers, there is time involved in conducting the trials on-farm, particularly at seeding and harvest, two of the busiest times of the growing season. But, this investment of time generates valuable information on the agronomics and economics of different management practices and products. Results from on-farm trials can be used to shift management practices or validate current practices on individual farms, but they can also be pooled together across space and time to gain an overall, big-picture understanding of the impact of a treatment or decision.