With the aim of studying the statistical procedures used in the postgraduate cour-se in Tropical Agriculture (PAT) of the Universidad de Oriente, period 1998-2016,the statistical management applied in 53 master work presented was analyzed.Variables related to the experimental design, treatment arrangement, number ofrepetitions, characteristics of the experimental unit, types of variables and scalesused, type of statistical analysis, verification of assumptions required by the testsand interpretation of interactions, among others, were identified. The information was systematized in a database and analyzed through descriptive statistics, Binomial test and correlation analysis, according to the variable analyzed. It was determined that the use of statistical techniques differs between mentions of the PAT. The tendency is towards the development of critical experiments with the use of robust experimental designs and treatment arrangements. More than 60 percent of the experiments employ less than four repetitions and there are deficiencies in the description of the experimental units. The assumptions of the ANAVA are not verified and there were deficiencies in the interpretation of significant interactions. The most frequent comparison tests of media were Duncan and Minimum Difference Significant (42.42 and 30.30 percent, respectively), which induce type I error.