Opinion

Check Tests, Recore Check Tests and Retests

AWTA Ltd uses internationally recognised standard test methods established by the International Wool Textile Organisation (IWTO) when issuing IWTO Certificates. Its laboratories are subject to audit by National Association of Testing Authorities, Australia (NATA) and licensed by IWTO to ensure that the procedures defined in these standards are rigorously adhered to.

Nevertheless, as a wool grower, wool broker or wool exporter you may have looked at an IWTO Certificate for a particular lot or consignment and thrown up your hands in horror and said: “These results are wrong!” So you grab a telephone and start calling to who ever you think can redress the error. If you are a grower you are likely to ring your broker or agent, who will then ring AWTA Ltd. If you are a broker then most probably you will ring or accost (if they are nearby) AWTA Ltd’s Sampling Controller or Sampling Manager. And if you are a buyer or a processor then the first call you make is most probably to one of AWTA Ltd’s Data Processing staff or to the most senior manager you can get hold of, even the Managing Director. You just want it fixed!

Once you make this call you set off a process that may or may not yield the outcome you want. Under the IWTO Regulations, which define procedures to address queries or challenges from clients about test results, AWTA Ltd must follow a very rigid protocol. This protocol is always followed. We think it is fair and impartial, because it is based on science and it is completely objective.

However, before getting too deeply involved in this discussion, we must define some terms, as these terms are a key part of the jargon surrounding the procedures that are followed when a test result is queried. So take some time to read through the definitions in the above table. They are important.

Understanding exactly what constitutes a testing error is particularly important. The definition in the table is highly technical and even dispassionate. But a client’s perception of a testing error is very much influenced by commercial as well as technical factors.

Case 1:

You are a wool grower producing superfine merino wool with an average clip diameter of 18 microns. You reasonably expect that the range of mean fibre diameter between your sheep will be 14-22 microns. But the range of mean fibre diameter between classed lots from your clip should be considerably less. So if you receive a Certificate on one of your lots, which states the mean fibre diameter is 23 microns, you will quite rightly conclude that an error has occurred. Despite every effort made by the Test House such errors do occur, albeit extremely infrequently, and are most often due to human factors causing a mix up of samples either at sampling or within the Laboratory. This is an example of a gross error.

Case 2:

Assume you are the same wool grower, and have prepared your clip using test data provided by an on-farm testing service provider, in the expectation that you will produce a lot of 15.8 microns. Instead you get a Certificate that says the lot is 16.1 microns. Perhaps your primary purpose for using on-farm testing as aid to classing is simply that you just want to produce at least one lot lower than 16 microns. Perhaps the economic justification for paying the extra cost of on-farm testing is strongly dependent on the extra income producing such a lot may yield. In either instance you are disappointed, and you will also consider that an error occurred, because your commercial expectations are not being met.

Case 3:

Spare a thought for the buyer or processor who has purchased a large number of lots at auction (based on Certificate data) at great expense, which average 18.8 microns, in the expectation that once the wool is scoured the mean fibre diameter will be in the range 18.8-19.0 microns, for a deliverable contract of 19.0 microns maximum. This expectation is usually developed by prior experience, and an understanding that processing often does change the greasy wool parameters slightly – hence the allowance between the greasy and the scoured wool. If the Certificate for the scoured wool says the consignment is 19.1 microns, 0.1 microns greater than the limit in the buyer’s contract, then he will certainly deem that this is an error, and reach for the telephone (you can also be certain that if the test result is less than 18.8 microns The test House will never hear about it). The cost of not meeting the contract specification can be very substantial, given the large amount of wool involved.

Quite clearly, there is a difference between gross errors such as the first example, and a failure to meet commercial expectations, as in the last two examples. Nevertheless from a client perspective they are all problems.

Regardless of the reason for the client query the procedures followed by Test Houses are the same. They have to be – to maintain the impartiality and integrity of the Certification system.

If residual sample material is still available then a Check Test will be conducted. If not a new sample will be obtained and a Recore Check Test conducted.

What happens next?

Firstly, the difference between the original result and the Check Test is calculated and compared with the Maximum Retest Range provided in the IWTO Regulations. If this is exceeded then the original result is discarded and a new Certificate issued, using the Check Test results only.

However, if the Maximum Retest Range is not exceeded, then the results from the original test and the Check Test are combined and a new Certificate is issued, using the combined results. Because of the power of averaging this actually improves the precision of the test result.

For presale and post-sale lots and for scoured wool this first round of Check Testing is provided by AWTA Ltd at no cost. If the client insists then a second, third or even a fourth round of Check Testing may occur, even if this sometimes requires additional sample material being obtained. However, fees are charged for this additional testing.

Gross errors, when they occur, are almost always corrected by the first Check Test, unless the residual sample material has been somehow misidentified or accidentally switched with that from another lot.

This process used to investigate possible testing errors has been developed by IWTO, with strong support from AWTA Ltd, and from the world wool industry at large, because it is fair, impartial and technically correct. It recognises that there is inherent variation in the testing system, which is difficult to remove without greatly increasing costs. The regulations, while they may appear to be bureaucratic, actually add to the integrity of the system. If followed they leave no room for the Test House to exhibit any bias for or against the client, and ultimately the client’s customers.

Of course, where a Check Test is requested for a commercial reason, this system means that the client is to some extent taking a chance. Provided no gross error has occurred the inherent variation in the measurements means that the new Certificate results have an even chance of going up or down. In Case 2 for example, a new Certificate may move the result to 16.2 instead of 16.0 microns. Indeed this does regularly occur, as many wool growers and brokers can attest.

However, buyers and processors are very good at their job and the rules they use for purchasing and assembling consignments are based on vast experience from past consignments. It costs them real money if they fail to meet the limits of their contracts so their buying decisions take into account this past experience. This knowledge means that when a test result exceeds their limits there is a better than even chance that it is at the high end of the range of variation that randomly occurs. Because of this there is also a better than even chance that the new result after check testing will be either the same as or slightly less than the original, and due to the power of averaging it will definitely be more precise.

The Maximum Retest Ranges that are applied in making these decisions are derived from very detailed statistical analyses conducted on a very large number of tests by independent laboratories. They therefore take into account variation arising from sampling, laboratories, instruments and people.

AWTA Ltd has available more detailed information covering these issues. This can be downloaded from the Company’s website: http://www.awta.com.au/Publications/Technical/
technical_limitations/Tech_Limitations.htm

The Maximum Retest Ranges for the key parameters Certified for Greasy and Scoured Wool and Wool Tops are detailed in the relevant IWTO regulations. Those applied to the Greasy Wool for Wool Base, Vegetable Matter Base, Mean Fibre Diameter, and Staple Length and Strength are replicated in the adjacent tables.

AWTA Ltd provides other testing services where results are not issued on Certificates e.g. Reports, Fleece Tests. When queries are raised on these documents the rules that are applied in determining whether or not an error occurred are the same, although in such cases sufficient residual sample material is frequently not available.

We are convinced that the procedures developed by IWTO, while not always yielding an outcome that a client wants, are critical to the overall integrity of the Certification system. However, their effective administration requires Test Houses that are independent and impartial.