Farmers like on-farm research because they can see test treatments applied across variable landscapes and they can use farm-sized equipment for management operations. However, many are unaware of the principles of experimental design and replication. They often conduct their own tests, but seldom replicate them.

Renewed interest in no-till following the 1996 Farm Bill spawned several on-farm projects in eastern Washington to explore transitions to direct seeding. These included the Wilke Project in Lincoln County, the Northwest Crops Project in Whitman and Garfield Counties, and the Spokane County Direct Seeding Project.

In the first two projects, participants compared 3-year and 4-year direct seed crop rotations. In the Wilke Project, 5 farmers each repeated one rotation for 4 years in 10- to 25-acre plots or as whole fields. In the Northwest Crops Project, 7 growers each maintained one replication of both rotations for 7 years in 30 ft by 600 ft plots.

Where crop rotations were in strips or small fields their management was often not optimal as the farmer's large fields were a higher priority. Conversely, when the farmer included whole fields in the test he sometimes changed the rotation due to weather or economics. Loss of replications, for a variety of reasons, reduced the ability of the statistical analysis to detect crop rotation differences. However, there was still much valuable experiential learning.

In the Spokane Project, 5 farmers each answered one question in their transition to direct seeding. They replicated trials in the field and for 3 years. These reductionist tests showed treatment differences, but did not provide a complete picture of direct seed systems. This farmer group received education on experimental design and demonstrated good understanding of the value of plot replication. Modest payments also provided them with an incentive to complete the trials as specified.