Understanding Outlier Reports

What is an “Outlier”?

As part of the on-going verification of the performance information being added to your database, we check the variation in performance records between animals within a contemporary group. These groups should represent animals of similar type that have had equal opportunity to perform under similar conditions.

A certain degree of variation in performance is expected within each group, the amount depending upon the trait measured (for statistics enthusiasts, the limits are defined by standard deviations).

When the difference between a performance record for an animal and the average of all animals in the contemporary group is greater than we would normally expect for the trait measured, the record for this animal is flagged as an outlier and is excluded from the BREEDPLAN analysis.

We have set these outlier limits fairly wide so that the likelihood of valid records being flagged as outliers is slim. Outliers can occur for several reasons, such as coding/punching errors, an incorrect measurement being recorded,incorrectly defined management groups or data entry errors.

Why have we sent you this report?

The attached report lists contemporary groups in your herd in which outliers have been identified. The outlier animals are marked in bold with the symbol # in both the left hand and right hand margins and adjacent to the Raw Observation. Also reported are the other animals with which the outlier animals have been directly compared. Different contemporary groups are separated in this report by blank lines.This report gives you the opportunity to correct or verify the performance for the animals that have been flagged as outliers and also to confirm the management groups recorded for these animals.

What do you do with the outlier report?

Please check each group of records carefully and report back to BREEDPLAN staff at Angus Australia. Simply write a letter or make notes on the Outlier report relating to:

  • corrections to the performance information or management group of the outlier record and/or any of its contemporaries in the relevant analysis group, or
  • verification of the information that you have checked and that you believe to be correct.
  • send your letter or report to Angus BREEDPLAN staff for the corrections to be updated on your BREEDPLAN database.
  • there are no fees incurred to have this information corrected.

Please note. If you do not contact us to correct or verify these outlier records, the records will be excluded from future BREEDPLAN analyses.

 

Interpreting the report?

Most of the headings are self-explanatory. However, a few points to remember

  • Herd indicates the herd in which the animal was measured
  • Number in Birth indicators are 1 = single, 2 = twin etc.
  • By ET indicates if an animal was from an ET program (Yes/No)
  • Fostered indicates if the animal was reared on its bearer dam
  • Group is the breeder-defined management group code (blank = no code supplied)
  • Age is the age of the calf when this measurement was taken.
  • Raw Obsn is the raw (unadjusted) measurement of the animal on the nominated observation date
  • Adjusted Obsn is the Raw Obsn adjusted for such things as the age of the animal at measurement and the age of the dam when she gave birth to the animal. It is this value that is being considered in the outlier group.
  • Trait is a description of the reported trait in Raw Obsn eg WW – Weaning Weight, MCW – Mature Cow Weight, IMF% – Intra-muscular fat.

Note: the ADJUSTED performance data (Adjusted Obsn) will be different to the RAW performance data (Raw Obsn), and the outlier refers to the trait described in the Trait column!

 

Hints to find the problem?

Individual things to look for in the data are:

  • Is the birth date correct?
  • Was the calf reared on the cow that gave birth to it ?
  • Is the raw observation correct?
  • Is the sex correct? is the date of measurement correct ?
  • Is the twin flag correct?
  • Is the management group correct (not only for the outlier record but for other animals in the analysis
    group as well)?
 Management or handling differences might be?
  • Was the animal prepared for a show or sale team before its weight was taken, whereas others in the group were not?
  • Was the dam of this animal prepared for a show team whilst the calf was at foot, whereas the dams of other calves in the group were not?
  • Did this animal or its dam get sick (low adjusted observation)?

Remember, if you do nothing, the outlier record will remain excluded! If you do not check these records and notify BREEDPLAN of the corrections or verification of the data, then the Outlier records will remain excluded from all future BREEDPLAN analyses.

Advertisements