Abstract

The task of biometrics in anthropology is to describe and analyze the random variability inherent in a person as an object of study. At the modern time, multivariate biometric models are widely used for this, in which it is assumed that the observed quantitative organism traits are expressed as weighted sums of latent random factors (for example, principal component analysis, factor and discriminant analyses, canonical variables, etc.). The given assumption is consistent with the existence of general and specific biological factors-causes that determine the organism traits. Since the main task of science is to study the laws of cause-and-effect in nature, it is important to analyze the relationship between statistical and causal dependences. Results and discussion. Statistical factors obtained as a result of data processing (such as the general size, shape, etc.) in the general case not only disagree with the causal factors, but may even contradict them. It is shown that the statistical coefficients of correlation, regression, and covariance, even in the simplest case of two traits randomly varying under the influence of one common cause and several specific factors, can take on infinitely many values under the same causal relationship, and with the same statistical pattern of variability, there are infinitely many causal explanations it. Conclusions. Interpretation in relation to the cause-and-effect patterns of trait determination and other causal conclusions should not follow from the results of the application of one-dimensional and multidimensional biometric models that use the indicated statistical coefficients as input data.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call