CLPEX Chat Board

A term that has come to be used in Friction Ridge Analysis is “Tolerance”. This area of the image is within tolerance, or this area is out of tolerance for inclusion. Those two events are just within the level of tolerance. So, just what exactly is tolerance. Here are a couple of definitions.

Permissible degree of variation from a pre-set standard.
www.neenahpaper.com/Glossary/index.asp

Allowable deviation from a nominal or specified dimension.
www.cattycorp.com/dictionary.htm

So tolerance is a variation/deviation from a standard or dimension. So where are the standards or dimensions? That is dependent upon exactly what you are observing. For example we have a standard definition of an “ending ridge”. So on one image we have what appears to be an ending ridge, while on the other image it appears to be a bifurcation. Is this out of tolerance?. Generally speaking I would have to say no (could be clarity issue). It is within tolerance. We also have a standard that the unit relationship between two images must have the same relative position. In other words the grouping of two ridge events having 4 ridges between them (unit relationship) then the second image must also have 4 ridges between comparative ridge events (same relative position). If one has 4 ridges and the second has 6 then we have the relative position out of tolerance.

For one to have an understanding of tolerance, one should also be familiar with the standards or definitions of the different features we may encounter.

The principal of tolerance states: That which is an allowable deviation from the friction ridge standard.

A bifurcation being observed as an ending ridge is within tolerance.

A bifurcation being observed as a short ridge is not within tolerance.

A short ridge being observed as a dot is within tolerance.

Under severe distortion a set of converging ridges appearing as a delta may be within tolerance.

The relative position between two events may be off by a single ridge and may be in tolerance. (dependent upon the accurate placement of the events).

I am looking for comments or thoughts on Tolerance.

I feel that tolerance is one of the most overlooked areas in our discipline right now. I think there is a lot of confusion about how much tolerance we should have for distortion. In part, this confusion is fueled by recent erroneous identifications and a general critical attack on the science of fingerprints. Examiners seem more cautious, and therefore are more likely to incorrectly state that something is out of tolerance, and not proclaim the match. This confusion is also fueled by recent discussion that erroneous exclusions are very different than erroneous inconclusive results. An examiner who reports in-conclusion for matching impressions is not viewed in the harsh light that is an examiner who reports an exclusion to those same impressions. This is a positive trend.

But some agency policy prevents examiners from reporting in-conclusion. They look to policies based on the one dissimilarity doctrine and if the examiner can't 'explain' the distortion, they are required to exclude, and there is no provision for in-conclusion. Perhaps a better doctrine would be the requirement not to identify, thereby leaving open the door for an inconclusive determination.

In the end, the decision of whether to tolerate a particular event that appears to be dissimilar is much like the evaluation phase in your 3-phased approach. Whether something is within tolerance or out of tolerance is based on the training, experience, talent, motivation, daily variables of the examiner him/her self. An incorrect conclusion could mean an erroneous identification or an erroneous exclusion.

An examiner who can't explain a dissimilar event should seriously consider whether they can 'tolerate' this event. If not, the examiner should separately consider whether this event is sufficient evidence upon which to base a conclusion of exclusion. SWGFAST guidelines require sufficient quality and quantity of difference to exclude, just as for individualization. I think we sometimes overlook the importance of that one decision. I have found on more than one occasion an erroneous exclusion where the examiner based the decision on a single poor quality feature that was distortion.

I would encourage examiners who are feeling the struggle between tolerance and conservatism to expand their inconclusive realm, not push more prints to exclusion. If agency policy prevents inconclusive reports, this is a separate issue that must be addressed or else the examiner will 'feel' the struggle even more as they try to resolve tolerance with making only black-and-white decisions.

I also feel in this environment that the profession needs to be more open to internal reviews of inconclusive results by other examiners, or even external reviews by other agencies or examiners when a conclusive result has not been reached by one or more examiners. In a recent project, I provided a mechanism for examiners to flag 'similar' prints and move on to the next comparison. These were subsequently provided blindly to a panel of senior examiners and this process saved what might have otherwise become an erroneous exclusion.

Blind review has been shrouded in a negative light by examiners using phrases of negative connotation such as 'shopping around' or 'fishing for a verification'. Instead we should herald the concept of blind review as one mitigation to the risk of letting the guilty go free in an era of relative conservatism in our discipline.

I have to agree with Stacy that tolerance is very much overlooked area in the discipline and from my POV right up there with suitability/sufficiency issues.

I am glad to see that the concept of inconclusive determinations are not considered under the same harsh light as erroneous exclusions.

I will go one further with Kasey and say that some agencies do have inconclusive as a conclusion, but LPE are only using it for conclusions where the exemplars are of poor quality. They will not use it when the latent print is marginal. The problem of suitable for identification vs suitable for comparison is still here.

As far as blind review, I find that concept more palatable than blind verification.

Back to tolerance, the following is from Quantitative-Qualitative Friction Ridge Analysis, by David Ashbaugh, p. 94. I want to think the person who contacted me and made me consider posting it here.

Second, clarity dictates our level of tolerance for discrepancy. In the case of a very clear print where not signs of pressure distortion are present, the tolerance for discrepancy in ridge shape is very low. In this case, major ridge path deviations must appear in the correct locations and have the same ridge path configurations and ridge shapes. However, in a very poor print that has been subjected to pressure distortion, small discrepancies can be tolerated, but all discrepancies must be consistent with being caused by a common force. At times more than one force can cause more than one type of distortion in the same print, but friction skin cannot stretch in several directions at the same time without rational cause.

The idea of allowing more tolerance with poor prints appears to be the reverse of what it should be. There is a cost for this increase in tolerance with poorly recorded prints. The cost is that during the individualization process there usually must be more ridge characteristics present to individualize than would be required in a very clear print. This is due to the fact that each area found in agreement in the unclear print carries less evaluative weight than if their configuration was clearly reproduced. The very clear print would likely have more detail available for the brain to compare in a smaller area due to the clear reproduction of the ridge details.

Let me see if I can change the Principle of Tolerance.

The Principle of Tolerance states: The allowable deviation from the friction ridge standards which is dictated by the level of clarity.

1. The clearer the ridge detail the lower the tolerance for deviation.
2. The less clear the ridge detail the higher the tolerance for deviation.
3. Distortion will allow small anomalies to be tolerated but must be consistent.

Anything to add?