EL ARBITRAJE ESTADISTICO Y SU INFLUENCIA EN LA CALIDAD DE LAS PUBLICACIONES CIENTIFICAS

the application of adequate statistical methods at the moment of presenting and publishing the scientific results plays an important role in the research process because they guarantee the accuracy and precision of the estimations and the conclusions. this paper is a continuity of that presented in the iv meeting of editors of scientific journals held in havana in 2002 and its purpose is to continue revealing the difficulties in the experimental methodology described in many scientific journals of latin america and the caribbean that were edited between 2000 and 2004. this paper summarizes the aspects related to the classical experimental designs of greatest negative influence. the deficiencies that are observed more frequently are, among others, the lack of specification of the statistical package used for data processing (62 %), not mentioning the standard error in the tables (51 %), the incorrect presentation of the statistical elements in the tables (33 %) and the inappropriate description of the experimental design (30 %). at present, the fulfilment of the reader？s expectancies related to the statistical tools applied is a challenge for all the scientific editions and it is indispensable to make an effort and use all the possibilities at hand to produce indicators of quality in the publications, mainly in the scientific journals of the region