Statistical Methods Used for Identification of Art Prices Determinants

Abstract
The paper deals with the problem of determinants of the auction prices of the works of art. A specialattention is given to the attributes of works of art i.e.: the name of artist, time of creation, format (size), theme, technique, signature and their influence on the level of auction prices. The author of the paper uses statistical methods to verify empirically the relationship between variables. The dependences have been explained by implementation of three methods ie.: analysis of variance ANOVA, multiple regression models and classification trees, with the use of the computer program STATISTICA. The analysis is based on the cross-section data on art auction results of severalauction sessions in Poland. The empirical studies have shown that name of artist, year of creation and technique determine the auction prices of works of art. The influence of technique and signature is not important statistically.