Working together for Australian excellence
This breeding value and $Index confidence range chart creator is under development. If you encounter problems using it, notice errors in its operation, or would like to suggest improvements, please submit details using the contact form.
Estimated Breeding Values (EBVs) are estimates of an animal's genetic worth, and their accuracy figures indicate likely margins of error (also known as confidence range or standard error) for each EBV.
However, the method used to determine an EBV's accuracy can exaggerate the true precision of an EBV by more than 40%. Furthermore, the apparently precise format of EBVs can further mislead users into believing they are more accurate than they really are. Even more misleadingly, $Indexes are published with no indication of accuracy even though they are frequently derived from very imprecise EBVs.
EBVs undergo significant change over short periods. For example, docility EBVs (even for published sires) can rise or fall by 30 points over two to three years. In a little over a year, the docility EBVs of seven Cycle 1 Limousin Information Nucleus sires fell by up to 23 points.
The challenge of determining likely margins of error for EBVs and $Indexes, and hence the degree of confidence stud breeders and their clients can place in them, was made more difficult by an ALBS board decision not to formally publish and distribute the annual Limousin Sire Summary to members. The annual Sire Summary was the only up-to-date source of information that identified the relationship between EBV accuracy figures and prediction error margins in Limousin cattle.
Fortunately, the time-consuming, confusing, and frequently overlooked, task of converting EBV accuracies into margins of error has been automated. Appropriately weighted EBV accuracy figures are also used to estimate likely margins of error for $Indexes, which have never been published before.
Results are displayed on a bar chart that enables EBVs, $Indexes, and likely margins of error, to be compared directly with the EBV and $Index rankings for 2011 born calves.
You can use either the option to retrieve data automatically from a database, or enter data manually.
For manual data entry enter an animal's name where indicated, and overwrite the EBV, accuracy and $Index default values. It is important that you delete and leave blank EBVs and accuracies not officially published for the animal you are analysing. EBV accuracy is used to estimate $Index margins of error, and where this information has not been published and is omitted on the data input form, a low EBV accuracy is assumed.
Displayed default values are breed average EBVs and $Indexes for 2011 born calves, and the default accuracy for all EBVs is 95%.
In order to become familiar with the chart creator's features you should initially use and submit the default values. You can also change accuracy figures to investigate the effect these have on the margins of error for EBVs and $Indexes. Click the "Submit Now" button at the bottom of the page to produce a chart displaying your data. Tick the box above the "Submit Now" button to display EBV and $Index confidence range values, which are the range of values within which true breeding values and $Indexes are expected to lie with a probability of about two in three.