Reliability of testing in kinanthropology/kinesiology as well as in all behavioral and social sciences is a crucial moment of empirical research in these sciences. In physics its counterpart is precision of measurement. Contemporary situation in organization and unification in physical measurement unfortunately does not have a similar international movement in the reliability of testing. The article presents some fundamental analogies as well as differences between the measurement models in the „soft“ sciences and makes some projections and comparisons with the physical measurement. It is presented that the main conceptual background for linking physical-measurement replications and their „soft“ science parallel forms of a test is the general principle of randomness. However, there is a substantial difference in other conceptualization of the physical-measurement Gauss model and behavioral-measurement parallel-forms model the concepts of „precision“ and „reliability“ are mutually transferable into each other. One of main specificities of classical test theory model is that best estimate of the true (precise) value is influenced by regression toward the mean and therefore is not unbiased as it is in case of physical measurement.

Charles University Prague Faculty of Physical Education and Sport Department of Kinanthropology and Humanities

Spolehlivost testování a přesnost měření: některé podobnosti a odlišnosti

Lit.: 41