SUMMARY
Introduction: The transparent evaluation of projects proposals for scientific research with financial support through calls made by governmental institutions or private organizations becomes extremely important, along with increasing confidence in the researchers who participate in the calls for the allocation of financial resources. Therefore, this research presents an alternative to the process of evaluating projects with financial resources. The current method of evaluating funded projects is to obtain an average of the scores assigned to each project by a group of reviewers. However, this method does not discount the effect of the reviewer and produces a bias in the final evaluation of each project, since projects are likely to be favored when reviewers are relaxed and affected when reviewers are demanding and very strict. Method: This research proposes the use of two experimental design statistical methods for project evaluation: (a) incomplete balanced block design and (b) Alpha block design. In addition, we provide codes in the statistical software R, for both statistical designs, which allows us to implement project evaluation quickly and accurately. Results: Based on the proposed methods, it was found that the traditional method presents a very important bias due to the heterogeneity among the reviewers. Discussion and Conclusion: The use of experimental design statistical methods (balanced incomplete block design or Alpha design) can reduce this bias considerably, obtaining a higher accuracy for evaluations.