HERD CHARACTERISTICS

• Breed of Herd: Primary breed reported by the DRPC

• Number of Cows: Number of cows enrolled in herd at time Herd Profile was generated.

• Average Turnover: Percentage of animals culled during the previous 12 months.

• Records Completed: Percentage of animals completing a lactation record. (Cow has freshened and dried.)

• Complete ID: Percentage of animals within a herd that are completely identified. This is found in USDA Format 14 and defined as “Total number of cows with usable identification divided by number of cows, both based on last available sample day. Usable animal identification means having all the following: valid cow and sire identification (registration or eartag number) and birthdate.”

• Number of ID Changes: Percentage of animals within a herd with an ID change. This is found in USDA Format 14 and defined as “Total number of cows with an identification change divided by number of cows, both based on last available sample day.”

• Milking Frequency: Frequency herd is milked: 2X, 3X, 4X or Other. The terms 2X, 3X, 4X indicate the entire herd is milked 2, 3 or 4 times in a 24 hour period. “Other” indicates the entire herd or part of the herd is milked differently in a 24 hour period (i.e. cows are milked every 10 hours, or high producing string is milked 4X, low producing string milked 2X).

• Weighing Device QC: Indicates whether the meters used to collect milk weights and (if applicable) component samples meet QC standards.

• DHI DRPC: Indicates name of computing processing center where herd data is calculated and stored.

• DHI Affiliate: Indicates name of DHIA Service Affiliate where producer is a member.

• DHI Laboratory: Indicates name of DHIA laboratory where component samples are processed and analyzed.

• DHI Field Service: Indicates name of Field Service responsible for DHIA data collection.

• Message: Indicates a brief text message (if applicable).

• RHA Graph: RHA plots by month the DHI calculated rolling herd average for milk and, when available, components (butterfat (F), protein (P), or solids not fat (SNF)).

• Bulk Tank Graph: Bulk tank comparison is also referred to as a Percentage of Milk Shipped. This statistic acts as a guide to ensure the accuracy of DHI milk0 weight data collection. AIPL research concludes that an acceptable bulk tank comparison range is 96-110 percent. (Current NCDHIP Rules, Appendix 2, states that the bulk tank weight shall be estimated by bulk tank pickup weights for three shipments immediately prior to the test date).

BOX PLOTS

• Distribution of PTA Milk: The purpose of the PTA ( Predicted Transmitting Ability) box plot is to display the estimated genetic potential of the herd.

• Distribution of ME Milk: The purpose of the ME (Mature-Equivalent production) is to display the production profile of the herd. It is computed using ME values calculated between 50 and 305 days in milk.

• Distribution of Deviated Standardized Test Day Milk: The original committee requested an average lactation curve line graph to provide a general view of lactation persistency. This was to be presented as averages of completed lactation curves in groups by lactation number. However, upon further investigation of the recommendation it was deemed difficult, if not impossible, to present meaningful information in a standardized curve presentation format. This was due to the fact that no one has developed multiplicative factors to produce a single “standard” that accurately accounted for breed, age, season of calving, regions, etc. Therefore, an alternative approach was developed by a subcommittee comprised of Dr. Ivan Mao, Dr. Larry Schaffer, and John Clay, who recommended utilizing a standardized test day milk approach, until newer research could be implemented.

The Distribution of Deviated Standardized Test Day Milk box plot calculates a 150 day Standardized Test Day Milk values from all actual test day milk weights between 50 and 305 days in milk. Each 150 day Standard Value is compared to the average of all the calculated values. A typical lactation curve is expected to have little variation between the two values. The Distribution of Deviated Standardized Test Day Milk could be used in conjunction with the M.E. Milk box plot. If cow(s) appearing as outliers on the M.E. box plot do not have several outlier values on the Distribution of Deviated Standardized Test Day Milk box plot, there is greater confidence that the animal had a typical lactation curve. We anticipate being able t6o update the presentation of lactation curves within a herd at a later date when current research methods are fully adopted and implemented.

INTERPRETING BOX PLOTS (by Dr. Todd Meinert, edited by Jill McGregor)

Box plots are a graphical way to display data that shows outlying data points as well as the average and spread of the data. To illustrate how a box plot works, consider the example herd of 100 cows with mature-equivalent milk yields. If the cows are ranked from lowest to highest for yield, the yield of the 25th cow (percentile 25) is where one end of the box starts, and the yield of the 75th cow (percentile 75) is the other end of the box.

The difference between the percentiles 25 and 75 is the interquartile range (IQR), and is used to measure the spread of milk yields in the herd. The ME milk yield of the 50th cow (percentile 50) is the median and is shown as a line that splits the box into two parts. The average yield of the herd is marked with a plus sign.

The lines extending out from the ends of the box are its “whiskers.” The end of the left whisker marks the cow with the lowest yield within 1.5 IQR (1.5 times the spread of herd yield) from percentile 25. Similarly, the end of the right whisker designates the cow with the highest yield within 1.5 IQR from percentile 75.

Outliers are observations that fall outside the values expected, given the spread of the herd outside t6he whiskers, and suggest that further scrutiny of these observations may be needed. Negative outliers have values less than expected, and positive outliers have values higher than expected. Outliers are further classified as either “possible” or “probable” outliers.

Possible outliers are cows with records that are “possibly” unusual. Their records are between 1.5 and 3 IQR units from the end of the box (percentiles 25 and 75). Herds with 50 to 147 cows are expected to have one negative and one positive possible outlier. Herds with 1,000 cows are expected to have 11 negative and 11 possible outliers.

Probable outliers are cows with records that are more extreme and unusual than cows designated as possible outliers. Their records are beyond 3 IQR units from the ends of the box (percentiles 25 and 75). Probably outliers are extremely rare occurrences, and such cow records should be monitored carefully. A probable outlier is not expected in a herd unless the herd exceeds 15,000 cows!

Research presented at the 1995 meeting of the American Dairy Science Association at Cornell University in Ithaca, NY, indicated that box plots could be useful in determining accuracy of DHI data.

Return to home page.