Title: Detection of Data Fabrication from Reported Results Author: Ivan Ropovik Affiliation: Faculty of Education, Charles University Abstract: Due to cumulative character of science, it is inevitable that every researcher builds their research program upon the results of others. At the same time, the current replication crisis underlines the naivety of unbounded reliance on any single research result. Although there are other more likely causes of a finding being not replicable, many recent uncovered cases of scientific misconduct, quickly rising number of paper retractions due to the unavailability of original data as well as some empirical work suggest that one of the culprits, data fabrication, might not be as rare as previously thought. This presentation sets out to demonstrate several examples of forensic procedures aiming to detect outright fabrication and fraudulent manipulation of data or results. All these techniques use just the reported results as input, applying the probability theory analytically and using simulations. Suchlike critical examination of the credibility of cornerstone published results increases the chance that the researcher's or lab's research program will build upon robust empirical foundations, thus having a progressive character.