Evaluation of the average comparison tests in the "Pesquisa Agropecuária Brasileira" Journal from 1980 to 1994

The choice of average comparison tests (ACT) is very important in the data analysis of experimental designs. Due to this fact, it was carried out a survey of the papers published in the "Pesquisa Agropecuária Brasileira - PAB" journal from 1980 to 1994 in which some type of ACT was used. The objective of this paper was to point up the growing problems of the indiscriminate use of the ACT applied to the statistical analysis in the papers published in this journal. The study was carried out using 628 papers, and the ACT were classified in the following categories: appropriate, not appropriate and parcially appropriate. As a consequence of this study, it was found that 57.01%, 31.53% and 11.46% of the surveyed papers applied the appropriate, not appropriate and parcially appropriate ACT, respectively.