Evaluation of reliability of description of dermatoglyphic patterns

Abstract

Numerous papers dealing with dermatoglyphics and yielding not quite uniform results allow one to raise a question: to what extent descriptive classifications of dermatoglyphic patterns are objective, what is their reliability? The aim of the present paper is to evaluate reliability of finger prints description and classification with use of the Lestrange-Bocheńska method [Bocheńska 1964] as modified by Rogucka [1968] — see fig. I. Finger prints were collected five times from 36 students of A. Mickiewicz University in Poznań. In this way the author obtained 180 dermatogrammes grouped into 36 units representing copies of the same hand each. After appropriate coding of material all 180 dermatogrammes were randomly mixed and evaluated by the above mentioned method as if each of them was a print of a unique set of fingers. After classification 36 units were restored on grounds of coding and their descriptions submitted to statistical analysis. Two methods of statistical evaluation were applied. Uniformity of description of repeated prints of the same finger was measured in two ways in both methods. The first way (K I) was to estimate uniformity of pattern and direction description, the second way (K II) was concerned only with uniformity of pattern description. Comparison of these two ways allows to conclude upon reliability of direction classification (table 10). With the first method reliability was estimated from the formula I. In each of the 36 units for each finger a number of identical descriptions (pattern and its direction) — n — was established. It has been arbitrarily assumed that the actual pattern is most frequently described for a given finger. In this way for each finger in each unit a matrix pattern was established and identity of its four replicates evaluated. Appropriate examples are shown in table 1. As a measure of replicability of description of all fingers of all hands, taking into account asymmetry as well (table 5, figure 5), is taken an arithmetic mean p of a number of identical replicates (p) — see formula 2 where: k — number of units. From the formula 3 an arithmetic mean p of identity of basic patterns classification was calculated (table 4, figure 4) as well as of all types of patterns established by Lestrange-Bocheńska method (table 3, figure 3). In this formula N denotes sample size. The second method — variational — does not require an arbitrary assumption on a „true” pattern to be made. Values of variance (R) express a degree of discrepancy between five replicates of the same finger in its description. Formula 5 was used for appropriate calculations (in this formula: 7 — number of identical variants, i — index of a pattern type described in a unit, k — number of the same finger replicates). Data presented in tables 6 and 7 are averaged values of R for all units. The analysis of data allows to conclude that for all characters replicability of results is incomplete. Generally, reliability of pattern description decreases with pattern complexity. The found inaccuracy of dermatoglyphic description points toward necessity either to introduce major changes into systems of pattern classification or to consider values of reliability during elaboration of dermatoglyphic materials. It is of great importance since many of the hitherto obtained results due to a varying degree are influenced by subjective judgement.