Goal:
The yield potential of different varieties of the same species varies greatly. Our goal is to, using the information that is readily available to producers and differentiating by specific environments, to lend objective analysis to the process of determining the genetics with the greatest yield potential.

Description of the problem:
One of the most significant factors affecting the profitability of a crop enterprise is the selection of the variety to plant. When growing corn, these yield differences can be as much as 20 or more bushels per acre. Under current economic conditions, this difference of \$35/acre (corn at \$1.75/bu) of revenue may determine if the field will show profit or loss. Unfortunately, many farmers spend less time or effort determining which varieties they will plant than is economically justified.

Practical problem:
Extensive data bases of variety yield information are compiled by seed companies as they accomplish for advertising and information side by side comparisons. Seed companies frequently encourage their seed dealers to establish at least one local side by side variety trial. Results from these trials are published on the web and in company reports. Further analysis of this data by producers and their consultants can result in improved management. This paper presents one method of analysis of this data. From a professional statistician¡¯s perspective, there are significant theoretical problems associated with how these side by side trials are designed and thus how they can be analyzed. The method described here fails to overcome these problems. However, a question that could be fairly asked is which is the greater error? 1). Our failure to attempt to further glean more knowledge from this trove of information. 2). Our attempt to further analysis data with known significant statistical problems resulting from a less than ideal experimental design. Obviously our conclusion is that we should attempt to further analyze the data despite the data¡¯s statistical problems.

We have looked at the standard deviation across years at single sites from the 30 years of precipitation data used to generate the map below. We have found the standard deviation from the mean to fall between 4.6 and 4.9 inches. This means that if the precipitation data is normally distributed, about 33% of the time we can expect the annual precipitation to be greater than or less than the mean ¡À ¡Ö4.75 inches of precipitation (4.6 to 4.9). We can look at the map below and crudely translate the 4.75 inches to a east ¨C west distance of 100 to 150 miles. This leads us to the conclusion that data taken 100 to 150 miles in both east and west of our location of interest is pertinent data to our decision making process. A north ¨C south distance is more difficult to quantify but given climatic variability, a distance of 50 miles in either north or south direction is justifiable. This defines the block of interest to us for analysis.

In this area of influence we will collect all side by side comparisons with greater than ¡Ö5 or more varieties being compared. We will compile all of this data into a single spread sheet in the following step wise manner

Continue to steps 1 through 4 >

Step 1.
From the discussion above, side by side comparisons that are within 100-150 miles of radius from our farming area will be considered. Compile as much of this data as is available and load the data into a spreadsheet as follows.

 Company Variety yield Farmer pioneer 37m81 176.4 Dedrich,D pioneer 3751 174.2 Dedrich,D pioneer 37h24 173.4 Dedrich,D

Step 2.
Number each row.
Insert column A. In A1 put 1. In A2 put = A1+1. Fill down to bottom of yield data. Select column A. Copy. Paste Special, values. (these last three key stokes appear to do nothing but are critical to this analysis. They remove the equations. Not doing this will create a mess later on)

 A B C D E 1 Company Variety Yield Farmer 2 pioneer 37m81 176.4 Dedrich,D 3 pioneer 3751 174.2 Dedrich,D 4 pioneer 37h24 173.4 Dedrich,D

Step 3.
Calculate an index for each variety within each farmer test site (hereafter all plots at one location will be referred to as a site).
In column F in the last row of each farmer’s site, calculate the average of each site. Select column F. Copy. Paste special, values. Copy up for each farmer’s site to fill all of column F.

In column G for each variety within each plot calculate an index as follows.

In cell G2, Index = (yield-farmer’s plot average)/ farmer’s plot average. Select the column from G2 to the bottom of the spreadsheet and copy down.

Select column G. Copy. Paste Special, values.

 A B C D E F G 1 Company Variety yield Farmer average Index 2 pioneer 37m81 176.4 Dedrich,D 169.7 0.039481 3 pioneer 3751 174.2 Dedrich,D 169.7 0.026517 4 pioneer 37h24 173.4 Dedrich,D 169.7 0.021803

Step 4.
Select entire spreadsheet. Sort by Company, Variety.

 A B C D E F G 1 Company Variety yield Farmer average Index 3 pioneer 3751 174.2 Dedrich,D 169.7 0.026517 27 pioneer 3751 142.7 kasperson 154.7659 -0.07796 56 pioneer 3751 154.7 sanderson 163.175 -0.05194 67 pioneer 3751 157.7 Vanderwal 177.5429 -0.11176 84 pioneer 3751 152.3 volkers 147.1067 0.035303 90 pioneer 3751 160 hecla 157.61 0.015164 99 pioneer 3751 139.9 harry 150.4143 -0.0699

Continue to steps 5 through 8 >

Step 5.
Calculate for each variety it’s index average.
In Column H, at the bottom of each variety, calculate the variety index average for each variety. Select column H. Copy. Paste special, values. Select column H, copy, and paste into column I. Fill I up within each variety, so there is an index average in every row.

 A B C D E F G H I 1 Company Variety yield Farmer average Index var ave ind var ave ind 3 pioneer 3751 174.2 Dedrich,D 169.7 0.026517 -0.03351 27 pioneer 3751 142.7 kasperson 154.7659 -0.07796 -0.03351 56 pioneer 3751 154.7 sanderson 163.175 -0.05194 -0.03351 67 pioneer 3751 157.7 Vanderwal 177.5429 -0.11176 -0.03351 84 pioneer 3751 152.3 volkers 147.1067 0.035303 -0.03351 90 pioneer 3751 160 hecla 157.61 0.015164 -0.03351 99 pioneer 3751 139.9 harry 150.4143 -0.0699 -0.03351

Step 6.
Calculate for each variety it’s index standard deviation.
In Column J, at the bottom of each variety, calculate the variety index standard deviation. Select column J. Copy. Paste special, values. The standard deviation will give you an estimate of how consistent a variety is.

<><
 A B C D E F G H I J 1 Company Variety yield Farmer average index var av ind var av ind std dev 3 pioneer 3751 174.2 Dedrich,D 169.7 0.026517 -0.03351 27 pioneer 3751 142.7 kasperson 154.7659 -0.07796 -0.03351 56 pioneer 3751 154.7 sanderson 163.175 -0.05194 -0.03351 67 pioneer 3751 157.7 Vanderwal 177.5429 -0.11176 -0.03351 84 pioneer 3751 152.3 volkers 147.1067 0.035303 -0.03351 90 pioneer 3751 160 hecla 157.61 0.015164 -0.03351 99 pioneer 3751 139.9 harry 150.4143 -0.0699 -0.03351 -0.03351 0.058411

Step 7.
Select all data. Sort by column H.
This will rank varieties by their relative competitiveness within this years data.

 A B C D E F G H I J 1 Company Variety yield Farmer average index var av ind var av ind std dev 104 pioneer 37r71 169.4 harry 150.4143 0.126223 0.116369 0.116369 0.066562 74 pioneer 35n05 149.9 volkers 147.1067 0.018988 0.080594 0.080594 0.096225 103 pioneer 36f30 159.7 harry 150.4143 0.061734 0.062866 0.062866 0.08217 102 pioneer 38p06 152.2 harry 150.4143 0.011872 0.029312 0.029312 0.083897 96 pioneer 38w36 150.8 hecla 157.61 -0.04321 0.026706 0.026706 0.054416 100 pioneer 37h24 151.9 harry 150.4143 0.009877 0.014986 0.014986 0.006144 76 pioneer 3559 158.9 volkers 147.1067 0.080169 0.014339 0.014339 0.070776 88 pioneer 3730 161.7 hecla 157.61 0.02595 -0.00433 -0.00433 0.056348 79 pioneer 35r57 147.9 volkers 147.1067 0.005393 -0.01567 -0.01567 0.020111 101 pioneer 38p05 141.1 harry 150.4143 -0.06192 -0.02366 -0.02366 0.037435 99 pioneer 3751 139.9 harry 150.4143 -0.0699 -0.03351 -0.03351 0.058411 80 pioneer 36a43 141.4 volkers 147.1067 -0.03879 -0.04748 -0.04748 0.007579 98 pioneer 37m81 138.7 harry 150.4143 -0.07788 -0.055 -0.055 0.06356 95 pioneer 38r21 148.8 hecla 157.61 -0.0559 -0.09036 -0.09036 0.03156 97 pioneer 3893 143 hecla 157.61 -0.0927 -0.10405 -0.10405 0.024995 78 pioneer 3733 133.6 volkers 147.1067 -0.09182 -0.10838 -0.10838 0.020154

Step 8.
Calculate how good each farmer did at selecting top varities.
Sort by column A to get the data back to it’s original order. Use column K to calculate the average of the indexes from column I. This is an interesting Statistic for it evaluates a farmers ability to pick the top varieties to include in his on farm research trial..