The peer review process for manuscripts submitted for publication to scientific journals and for the evaluation of grant research proposals is unsatisfactory in several respects. We examine here some of the problems related with evaluation of scientific merit. Some criteria for rejection are proposed, i.e. a poor preparation of the manuscript, a lack of a distinct hypothesis, a disagreement between hypothesis and methodology, and a deficient methodology. Other important criteria causes of rejection would be lack of originality of the hypothesis, scarce relevance of the work, and inconsistency in the results. Conversely, interesting work are rejected for invalid objections such as "less than optimal design", "lack of experience of the group" and some conceptual objections which are controversial. In order to improve the peer review process, we propose a larger role of editorial committees in final editorial decisions, an improved mechanism for selection of reviewers, and more explicit criteria for causes of rejection for reviewers and authors.