The problem of publication bias and ways to minimize it is addressed based on a non-systematic literature search and review. Publication bias is the tendency to publish original research with the achieved statistical significance of the results (p<0.05). Consequently, in the aggregate of accumulated data, there is a shift towards positive results, while the results of studies with negative results are not published for several reasons. This bias poses a significant risk for both patients and the medical community, as well as for scientific departments and the entire healthcare system. The current status quo, which shapes the socially expected result from researchers, can contribute to data manipulation to obtain statistical significance in scientific papers. The problem of publication bias is particularly relevant in theses, where the requirements for scientific novelty often lead to a situation where authors of theses focus on finding any differences between groups. Even if they have no meaningful practical value, these differences are often highlighted as the study’s main finding. In this context, there may be a temptation to present results that are not true, which further exacerbates the bias problem and reduces confidence in the results of scientific research. In order to reduce the impact of publication bias, the decision to publish the results of a study should be separated from the statistical significance of the results obtained and be focused on the purpose of the study and the correctness of its methodology. For theses, the requirements should also shift from the search for any differences to the study’s methodological quality, emphasizing the publication of those results that the authors initially planned. This approach will avoid the artificial focus on insignificant differences and reduce the risk of publication bias, ensuring greater transparency and reliability of scientific data.