This study was aimed at applying Data Envelopment Analysis (DEA) models to evaluate the efficiency of introducing and adapting different technologies in beef cattle production system in the Pantanal wetlands, Brazil. The economic data used in the analysis were recorded from 1995 to 2002. Two methods were used to select the DEA variables set. The first one is based on the multicriteria approach of variables selection, which combines the conflicting objectives of better ranking of decision making unit (DMU) and higher average efficiency. The second method makes use of statistic multivariate techniques (principal component analysis). The measures of efficiency with respect to the classic and to the inverted DEA frontiers, and the value of the composed score that includes both results were evaluated for the two resultant models (six inputs and one output and three inputs and one output, respectively). The linear trend analysis indicated as the best model the one selected by the multicriteria approach (including six inputs and one output), whose composed score resulted in significant estimates of the regression (0.0455) and determination coefficients (0.70).
efficiency; performance evaluation; production systems
