HOW ACCURATE ARE HAIR COMPARISONS?
Historically, the goal of most forensic hair examinations is the microscopic comparison of a questioned hair or hairs from a crime scene to a known hair sample using a comparison microscope (Figures 1 and 2) based upon a list of characteristics that cover the root, the microanatomy of the shaft, and the tip . The evaluation of microscopic traits within and between samples is key to the comparison process; Figure 2 shows hairs that are positively and negatively associated.
FIGURE 1 A comparison microscope is used for the examination. A comparison microscope is composed of two transmitted light microscopes joined by an optical bridge to produce a split image. The sample on the right appears in the right-hand field of view, and the sample on the left appears in the left-hand field of view. This side-by-side, point-by-point comparison is central to the effectiveness and accuracy of a forensic hair comparison. Hairs cannot be compared properly without one.
FIGURE 2 A hair comparison is a good method of demonstrating possible association between questioned hairs and individuals. A suitable known sample of hairs from the same body area is necessary to conduct a comparison. Hairs are not a means of positive identification, however; statistics or frequency estimates cannot currently be applied to microscopic hair comparisons.
But how to interpret the results of a forensic hair comparison? This process is not as simple as might be imagined. Wording of microscopic hair examination results usually takes the following forms:
• …a hair sample from the crime scene and another taken from (the defendant) were “similar” and “consistent.”
•…the pubic hairs from the crime and (the defendant’s) pubic hairs exhibited “similar microscopic characteristics.”
•…that hairs from the crime scene were “similar” to (the defendant’s).
•…hairs from the crime scene exhibited “the same microscopic characteristics” as (the defendant’s) hair.
A fundamental misunderstanding about what can and cannot be said with hair evidence lies at the heart of this issue; none of the preceding statements are beyond the realm of what is considered reasonable testimony. But how specific are these statements and what notions do they convey to laypersons? Given the list of traits in Chart 1, it seems that hairs could be coded, entered into a database, and eventually frequency information could be derived. A hair’s traits could be entered as a query, and at the push of a button, a frequency of occurrence for a population could be calculated. But it’s not that easy. Exploratory research has been conducted to evaluate hair results statistically (Gaudette and Keeping, 1974; Wickenheiser and Hepworth, 1990) but, to date, no universal approach for calculating significance has been published. And probably none will be (Gaudette, 1978, 1982). Hairs are a very complicated composite biological material and the expression of hair traits across the population is highly variable. Being three-dimensional makes quantifying the traits that much more difficult; characteristics would need to be recorded in X, Y, and Z dimensions, none of which are easily measured under a microscope. And now that DNA analysis is more accessible, a statistical approach based on microscopic examination is hardly justified.
