A new methodology assessment for the laboratory in engineering courses to reduce academic violations is proposed. It is based on invigilated face-to-face practical exams, and it is suggested as an alternative to the conventional laboratory assessment through student reports. This traditional method is sensitive to commit academic violations by students, like cheating, however, with the new proposal, it is drastically reduced. In the new assessment, the laboratory sessions are divided into theoretical-practical and tutorial sessions, when students work by themselves; and test sessions, when invigilated face-to-face practical exams are carried out to assess the laboratory part of the course. In these exams, the students must prove their knowledge in real-time with 100% practical exercises. Before implementing this new methodology, the results in the course under analysis showed that cheating was becoming widespread with a clear increasing tendency. To test the proposed assessment, a comparison between the traditional and the new methodology laboratory assessment is accomplished in a Computer Architecture undergraduate engineering course. Descriptive and inferential statistics are used to analyze the influence of the new assessment on the learning results. The results show that with the new proposal the number of detected copies disappeared, and the withdrawal rate is reduced without having a significant influence on the final mark or the theory mark. Therefore, the new methodology assessment has removed the academic violations without interfering with the learning process of the students.