8127/frame/fm

 
A.
Optimization
 
Optimization
might be defined as relating to the test system itself. Optimization
establishes that the test system is working properly so that any variation seen is
due to the samples themselves. For example, for latent fingerprints, optimization
might include testing visualization reagents with known prints, or surveying
different inks, ink pads, and paper for taking reference prints to see which gives
the clearest, most reproducible results. Optimization for shoe impressions might
involve testing various casting materials to see which is the most durable and
retains the most detail. Optimization of microscopic techniques would center
8127/frame/ch09 Page 222 Friday, July 21, 2000 11:44 AM

---

Good Laboratory Practice — Establishing Validity and Reliability
223
on testing the quality of various lenses alone and in combination, and in
determining which lens, filter, aperture, and light combination might work
best for a specific application.
For strictly visual comparisons, optimization of the data comparison
process itself is a bit hard to conceive because the instrument used is the
human brain. Perhaps it is best analogized to sensitizing and calibrating one’s
brain to discern the characteristics upon which various comparisons depend.
To extend the analogy, directed training, practice, and experience allow the
skilled examiner to determine useful characteristics reliably and to differen-
tiate between optically encoded signal and noise. The way in which we can
determine if an individual brain is optimized is by proficiency testing.
For instrumental systems, such as Fourier Transform Infrared Spectros-
copy, Gas Chromotography, or Capillary Electrophoresis, optimization might
include establishing calibration tolerances for the instrument, deciding what
settings give the best balance between signal and noise, and verifying that
known samples reproducibly give correct results. Any sample preparation tech-
niques, such as DNA amplification or chemical derivitization, must also be
optimized so that the cleanest results are obtained and sample differences
cannot be attributed to a preparation process needed before instrumental anal-
ysis can be employed. Interestingly, although an instrument is used to translate
chemical data into a visual form that humans can understand, and the visuals
are usually accompanied by quantitative data, in the end, we are left comparing
visual patterns. Each analyst’s individual experience at discerning the significant
aspects of those patterns contributes to the quality of the analysis as a whole.
The foregoing are only a few examples of the type of work involved in
establishing that a particular test system is optimized for a particular appli-
cation. The reader might like to think through various forensic test systems
and consider what might constitute optimization. Note that, in the process
of optimizing the system to produce the highest quality and most reliable
results, a body of work is accumulated that begins to address both the capa-
bilities and limitations of the system.