Our goal is to provide growers in New York State with reliable information on vegetable variety performance and new cultural practices. This often means comparing many varieties or treatments within one or a few locations. To keep the trials manageable, we use small field sites, but we often calculate the yield per acre to make the results more meaningful to you. The results reported in the following tables are most useful to compare among the treatments or cultivars within the trial. You should test any promising cultivars or practices on a small acreage at first, to determine the performance on your own farm.
To help identify yield if differences among the varieties or the practices are real, we use statistically sound experimental layouts, such as a Randomized Complete Block (RCB). Growing conditions may vary from place to place within a field, due to many factors, such as soil type, slope or surrounding landscape. To prevent making the wrong conclusions about a treatment, we replicate, or repeat the experimental layout across a particular field. In addition, the order of the treatments is randomized to avoid giving any one an unfair advantage (i.e. always in a highly productive part of a field) (see diagram below).
We use these types of trials to help distinguish between real cultivar differences and chance differences (due to other factors such as field variation). If we do not have some way of identifying the cause of these differences (cultivar vs. other), we might be misled by the conclusions from the trial. To help compare among the different cultivars, we have provided a measure called the Least Significant Difference (LSD). A treatment difference (i.e. difference in yield between two cultivars) must be greater than the LSD before we conclude that it is real (or significant).
In this report, we tolerate only a 5% chance of being wrong if we say that a difference between cultivars is real (LSD.05). In other words, we maintain a 95% level of confidence that our measurements of cultivar performance are due to the plant material and not random variation.
Please be aware that the LSD value is unique and specific to the trial. If the same was repeated in a different year or place, this value would change, based on the analysis of the field and treatment variation.
An Randomized complete Block Design
Four Cultivars (A, B, C, D), randomized within four replicates
| Replicate 1 | Replicate 2 | ||||||||
|---|---|---|---|---|---|---|---|---|---|
|
|
|
| Replicate 3 | Replicate 4 | ||||||||
|---|---|---|---|---|---|---|---|---|---|
|
|
Return to 1999 Results
Return to Vegetable Research & Extension Program
©
Copyright by the Department of Horticulture Website at Cornell University.
Please direct comments/corrections to the Horticulture Department Web Team at
Hort_Web@cornell.edu.
.
Home Page URL: http://www.hort.cornell.edu