Total amount: € 0,00
Online ISSN 1827-1677
Stephan C. N.
Joint POW/MIA Accounting Command Central Identification Laboratory Hawaii, USA
The capabilities of facial reconstruction (FR) methods are ideally assessed through empirical measurement, thus reducing subjectivity and increasing repeatability. Whilst many prior accuracy studies on FR aspire to this goal, all possess significant limitations. In the absence of crucial tests, conclusions concerning the accuracy of the methods must be drawn from the collective evidence. However the differential allocation of interpretive weight to single investigations depending on their advantages/disadvantages is required. This paper reviews assessment protocols and published data to arrive at an overarching assessment of method capabilities. Frequently it is observed that results implying method accuracy are produced by less attractive assessment techniques and small samples of skulls/practitioners/assessors (e.g., resemblance rating studies). Also, the rare production of convincing facial approximation successes by more rigorous tests (i.e., relatively larger sampled face arrays) suggests that the accuracy of current FR techniques is not outstanding. Recently identified flaws in specific facial feature prediction rules underscore methodological problems as a major cause of poor method performance. These observations hold importance for discussions on the validity of: the term “facial reconstruction”; labelling of FR as a “science” — either in part or in whole; and claims that poor method performances are underpinned by practitioner experience/skill. In the future it is clear that more stringent empirical tests are required, including more tightly controlled face array assessments, if overall method performance is to be robustly validated.