How to Interpret an ELB Report

Knowing how to not only interpret the ELB report, but also how to relay its results to a Grower is essential as an Advisor.

The image on the top of the first page will give you and the Grower a visual of where the ELB was placed:

Next you will see data in regards to the Growing Environments:

• Precipitation
  • Is broken out in Emergence, Vegetative Growth, and Grain Fill Seasons
  • The months that are included in those seasons is listed.
  • The rainfall (in inches) is listed for the current year.
  • The 5 year average of rainfall is listed and is a great way to compare if the current year was a wet/dry year and if so, when it was.  
• Temperature
  • Is broken out in Emergence, Vegetative Growth, and Grain Fill Seasons
  • The months that are included in those seasons is listed.
  • The rainfall (in inches) is listed for the current year.
  • The 5 year average of rainfall is listed and is a great way to compare if the current year was a cool/hot year and if so, when it was.

Onto Page 2!

ELB trial and its treatments are now visible for the first time!

To the right of the trial image, you will see information that really gives you a more in depth look at the conditions in which the trial was conducted on and is not just an observation, such as with a traditional Learning Block.

• The Crop Details will tell you General information about the trial.
  • Cropping system
  • Dominant soil type
  • Dominant soil suborder
  • Seed Company & Hybrid/Variety
  • Planting Date
• The Soil Test Averages are the averages within the trial area, NOT the entire field.
• Experiment Details:
  
  • Type of trial (Nitrogen, Plant Population, Potassium, etc.)
  • Product that was used
  • Number of rates that were tested
  • Number of times that the rate was tested (replicates)
  • Total replicates in this trial
  • Area in which the trial took place on (in acres)
• Applied Nutrient Averages lists the average actual applied nutrients within the trial area for:
  • Nitrogen
  • Phosphorus
  • Potassium
  • Sulfur

Now, let's look at the Treatment Details:

• Labeled as A, B, C, D, & E used in “Comparisons of Average Yield
  Responses” section.
• Intended rate from the ELB prescription
• Average Applied Rate from the valid observations used in the
  statistical analysis (across all replicates)
• Coefficient of Variation is: Weighted standard deviation of all treatments about treatment mean ÷ Weighted treatment mean
• Average Yield from the valid observations used in the statistical
  analysis (across all replicates)
• Blue Ribbon Winner  icon (light blue at 85% confidence interval) makes it easier for those that don’t want to reference the “Comparisons of Average Yield Responses” section

The 1st thing you want to know is:

• Answer to the question: “Is there a difference in yield response
  between treatments?” “No” refers to region shaded gray (0-85%
  confidence). First “Yes” refers to 85% confidence region; second
  “Yes” to 95% confidence region.

Distribution of FE Chart is meant to make it easy for a Grower/Agronomist to quickly answer this question!

• In the above example, p = 0.093 means you can have ~90% confidence that the differences
  in yield are caused by differences from the treatment.  Another way to say this is, “You tested a management factor that impacted yield.”
• Individual ELB results (red star) that are “further to the right” on
  this test distribution increase confidence in results.
• Graph of the test statistic distribution is idealized (smoothed and
  narrowed to fit on report).
• Distribution of test statistic is derived by randomization.

• • Tests of contrasts between Treatments  A & B, B & C, and A & C.
  • Treatments A, B, and C are ordered by expected yield response (i.e., Treatment A is expected to produce less yield than Treatment B). Hence, contrasts are written as the arithmetic difference between the higher treatment and the lower (B-A, C-B, C-A).
  • If our expectations of B out-yielding A is correct, then the
    difference “B-A” should be positive (>0)
  • p calc values are based off differences in mean outcomes per treatment from the ELB experiment; compared to “B-A”, “C-B”, and “C-A” distributions generated by randomization.

In this example:

• • There is no difference between the mean yield of Treatment B and Treatment A at an 85% or greater confidence level.
  • There is a statistically meaningful difference between the average yields of Treatment C and Treatment B at the 85% confidence level. Hence C is the “Blue Ribbon Winner.”

Lastly, let's look at the Whisker Plots!

It displays the span from the smallest to largest means of the subplot
replicates

Frequently Asked Questions:

1. What happens if the Whisker Plots and CV values don't mirror each other? 
   1. Whisker plots can be misleading in that they do not illustrate how many yield points landed near the high end of yield, low end of yield, or near the mean.  A low C.V. value would reflect yield points more tightly clustered around the mean, where a high C.V. value could mean the yield points were less clustered.  You could still have a low C.V. value but the whisker plots could look to be a large spread.
2. What if what we tested wasn’t the yield limiting factor? How do we know?
   1. If there is no statistical difference in yield between treatments, the lowest rate will be the winner. The logic: why spend money on more of that input (seed, fertilizer, etc.) if there is no confidence that more of it gains you bushels?
   2. A lack of statistical difference will usually show big spans in yield within treatment(s) in the whisker plot that tend to overlap each other visually.
   3. Could have large C.V. values.
3. What does it mean if my ELB Report states that the trial has a "Type II Error"?
   1. A Type II Error is concluding that there is a significant difference when there is not.  The ONLY time you will see this message is if the confidence level is between 70 - 85%.
   2. You should still have a discussion with the Grower and would want to weigh their options.  In this example, 70% may be significant enough for the Grower and/or Advisor to be comfortable with the stated risk.
      
      
4. If the confidence level is below 70%, the lowest treatment will be declared the "Blue Ribbon Winner".