ABSTRACT

The USDA, AMS, Cotton Program has offered module/trailer (module) averaging on a voluntary basis since 1991. When analyzing the historical results of module averaging for gins across the country, the Cotton Program has concluded that the reproducibility of quality measurements is significantly higher for cotton bales assigned the module average versus the bales’ individual readings. This increase in reproducibility indicates less variability in the fiber measurements and more stable classification results. The 2003 crop data continued to support the historical trends.

Module Averaging is advantageous to all users of the Cotton Program classification data. The number of gins participating in module averaging in 2003 rose 14% over 2002 to 210 gins, representing 3,808,979 bales (through December 26, 2003). Analysis of the classification results for these bales show that overall, module averaging proved more advantageous monetarily over individual bale values when using the government loan program as the basis for analysis. In 2002, the Cotton Program implemented a new policy of assigning the module average to all module middle-bale outliers without requiring a review classification (with some exceptions). The Cotton Program implemented this policy after historical data and studies proved that middle-bale outliers, when retested, consistently tend to approach the module average within acceptable testing parameters. This policy excludes the module’s first and last bales because they show to be true outliers a higher percentage of the time due to possible overlapping of modules having different fiber properties. This policy also excluded any outliers from a module when they accounted for more than 20% of the total number of bales in the module. These exclusions were implemented to detect rare exceptions to the norm. The Cotton Program also tightened outlier identification tolerances in 2003. Included in this paper are 2003 module averaging statistics, outlier statistics, and comparisons to previous crop years.