This thesis presents a look into citation counts as a measure for scientific impact which in turn is used to determine the replication value (RV). first, by comparing citation sources (WoS, Crossref, Scopus and Scite) from which citation counts can be retrieved. Secondly, by removing contradicting citations from the citation count, and comparing this new citation count without contradicting citations with the original total citation count. In both cases, based on the citation count, rank order lists are formed which are compared with the use of two tests. First, Kendall’s tau is calculated to see how well the compared pairs of lists correlate. Second, the rank biased overlap (RBO) is calculated to see how well pairs of lists overlap. The RBO is different than Kendall’s tau because it is able to give more weight to citation counts at the top of the list emphasizing the importance of high ranked articles as opposed to low ranked articles. Both measures indicate a significant correlation and overlap between ranked lists originating from Scopus and Crossref and WoS, and a lower correlation and overlap between Scite and all other sources. Based on the difference between Scite and all other sources, Scite is not yet the best choice as a citation source for determining scientific impact. Both measures also indicate a strong correlation and overlap between the ranked list formed from the total citation counts and the ranked list formed from the total citation count minus the contradicting citations. Based on this high correlation and overlap, taking out contradicting citations is not needed when determining scientific impact.