The error rate for the T-Model to reliability define numbers of look-alikes for amounts of ridge formations in agreement based on relevant fingerprint population, was examined with regards to errors in measurements taken based on tools and methods used, error rate to misidentify known look-alikes and erroneous individualizations as sufficient to individualize, and the spread between the numbers of look-alikes found in precise fingerprint populations compared to numbers of look-alikes predicted by the model.

Errors in Measurement

Measurement is fundamental in science.  A good measurement should be both precise and accurate, whereby accuracy is determined by the care taken by the person making the measurement and the calibration of the instrument.   The T-Model applied the use of a standard metric ruler and protractor to measure friction ridge skin elasticity threshold.  These tools do not require calibration and therefore their condition were considered perfect.  However, such measurement involves entities that may be subdivided into smaller and smaller fractions and is thus always an estimate. Therefore although effort and care was taken to make measurements as accurate as possible and since all measurements are necessarily approximations, conservative rounded values were applied throughout.  

The T-Model applied the use of counting to define numbers of ridge formation types, intervening ridges, and diminishing pattern areas, in order to establish their relative frequency and subsequent quantitative weights.  Although counting is exact because it involves discrete entities that are not subdivided into fractions, the interpretation of ridge formation types, for example, and their frequencies, are subject to error.  In order to reduce the uncertainty for ridge formation misinterpretation and wrong assessment, objective observation and conservation interpretation and assessment of ridge formations applying lower bound values for ridge formation ambiguity was applied throughout.

With exception to frequency of occurrence for the trifurcating ridge unit, Osterburg’s interpretation and counting of ridge unit types was assumed correct.  The author considers the scientific knowledge, i.e. quantitative weights, reduction, expansion factors and so on, that was developed as a result of this study along with various other studies performed by the author, are at this time best objective, conservative estimates.  Since it is impossible to perform the most complete experiments and validation studies, definitive values for these variables will always remain relative. 

The error rate for 39 fingerprint samples used was previously defined as approximately ±1.1% and deemed small enough to be of little consequence especially since conservative and/or lower bound values for ridge formation quantitative weights were applied throughout the model.  However, pending peer review the author allows for the possibility that these values may not be sufficiently precise and therefore remains open to the possibility for a frequency of occurrence study with a larger fingerprint sample size and possible refinement of various values and formulae.

NEXT PAGE >>> 

No measurements are perfect.  All measurements are fixed yet uncertain.