Submitted to: Proceedings of the Annual Precision Ag Conference
Publication Type: Proceedings
Publication Acceptance Date: 11/18/1998
Publication Date: N/A
Interpretive Summary: Accurate comparison of farming practices is important in the development of new farming methods. Effective tests can be conducted by farmers and agricultural support personnel if some simple but essential principles are followed. Without proper replication and randomization very little confidence can be had in making field measurements. The principles of replication and randomization are explained in simple terms for use in on- farm research.
Technical Abstract: On-farm research must be designed properly if it is to produce accurate data. Replicated side-by-side plots at least 700 ft in length have proven to give the most reliable results in the dryland Pacific Northwest. Four replications is considered a minimum for satisfactory results, and the treatment assignment must be randomized. The Least Significant Difference statistical test is the most commonly used statistic for deciding if average differences between treatments is large enough to be considered as having been caused by the treatments instead of being simply a result of natural field variability.