Abstract

The aim of the article is to present the nodal issues associated with the use of statistical measures in transfer pricing analysis. Obstacles to standardising the process of determining the transaction value range include small sample sizes, high data dispersion or the difficulty in identifying outliers. The article also presents recommendations related to the use of selected statistical approaches in the context of their strengths and weaknesses for data analysis and maintaining comparability in the TP area. Taxpayers do not have to choose the median or the mean, but can adopt a value that is more favourable to them but still within the interquartile range. Taxpayers may use any measure, but the choice of a particular measure should be reasonably justified.