Abstract. The paper refers to the importance of test results uncertainty in the assessment of the construction product performance. Uncertainty understood as dispersion attributed to any value which is result of the tests, calculations or other evaluations, occurs at every level of assessment (material, product, construction). Authors presented the various approaches to uncertainty estimation, especially in situation of small number of tests results which is characteristic for testing of construction products. Effects of uncertainty on final assessment of construction product was analyzed using the example of bearing capacity of thin-walled structure obtained by numerical calculation. Different values of material tests uncertainty resulting from different approaches to its assessment was taken into account. It was demonstrated that the difference in the results of strength tests of a material, which falls within the limits of uncertainty, may result in a very significant difference in the evaluation of a structure.Key words: uncertainty, numerical calculation, uncertainty influence on product assessment. Each value resulting from tests may only be considered in terms of probability. Unfortunately, in a small number of cases it is possible to express this probability using a probability density function (PDF) which is characterised by the expectation value and standard deviation (SD). In all tests which yield qualitative results (ordinal scale with two categories e.g. "fulfilled", "not fulfilled") or semi-quantitative results (ordinal scale with more than two categories), which, in the authors' opinion, are the majority of tests to which construction products are subjected, the determination of a PDF is an issue requiring extensive inter-laboratory tests performed on repeatable items [5][6][7]. Usually this is not possible for tests on construction products.Linking a value being measured with a PDF that is assigned to it is much easier for quantitative methods, for which the estimation of the uncertainty of a test result, allowing one to provide a PDF, is subject to numerous guidelines and instructions. The basic principles have been developed by the Joint Committee for Guides in Metrology [8], and are commonly applied by virtually all laboratories. However, as will be demonstrated for destructive (e.g. mechanical) testing, the estimation of uncertainty may be associated with numerous ambiguities.Presented by Walker et al. [9] typology and terminology of harmonized uncertainty provide uncertainity identification by location, including inter alia parameter uncertainty (associated with the variability of input data) and model uncertainty (associated also, but not only with lack of sufficient understanding of the system). These two kind of uncertainty sources are presented in test results and in estimation of uncertainty of the test results (uncertainty of estimated uncertainty). Variability relating to laboratory testing results has many components. Some of them, e.g. metrological parameters of testing equipment, ...