In car-paint analysis, sources of potential error can be divided into two groups. The first group is composed of those errors that are related to the quality of the analysis of material submitted for examination. The aim of car-paint analysis is generally to determine whether paint material on one object is the same paint as that on another object. Errors can occur due to variability of physicochemical features analyzed within one object, the variability of physicochemical features between objects in the same category (e.g., acrylic paints), and the frequency of occurrence in the general population of defined values for the studied feature. These errors can be accessed on the basis of results of population studies. Research into the possibility of differentiation of car paints that originated from different cars has been conducted in Canada and Australia. This research concluded that physicochemical analysis is useful in such an analysis. Incomplete data on the frequency of occurrence of a particular kind of car paint color in the general population are also available. The sources of errors that compose the second group are related to whether the object from which the evidence material originated was present at the accident scene. The sources of error are related to the possibility of primary and secondary transfer as well as environmental contamination. This paper discusses analytical procedures that influence the likelihood of errors occurring in these two groups, as well as statistical procedures for assigning a numerical value for the evidence regarding the priority it should be given in resolving the questions at issue. 38 references