Rzetelność i funkcjonowanie pozycji skali a ukryta niejednorodność populacji

The aim of paper is to evaluate the reliability and item bias in the measurement in cross-cultural comparative research and heterogenous populations. Well known reliability coefficients (Cronbach's, Armor's and McDonald's) are used in homogenous population with common frame of reference. We present the holistic approach to scale reliability assessment, when homogeneity assumption is violated in research practice. This approach enables full control of measurement bias in heterogenous population.