A novel control model for technical and economic analysis of fertilizer recommendations is based on Liebig's Law of the Minimum, Mitscherlich's relative yield theory, and the notions of yield plateau and soil fertility carry-over. The empirical model consists of a multistage separable programming specification which maximizes the discounted stream of net revenues subject to crop response and fertility carry-over functions. The model is applied to the wheat-soybean cropping system in Southern Brazil. It is found that, while the optimal fertility target levels are within a small range of those determined by Brazilian agronomists, their maintenance recommendations could be substantially improved.

This paper address the issue of performance evaluation -productivity and quality -of academic departments at an University. A DEA model was used to simulate a process of cross-evaluation between departments. The results of DEA in the dimensions of teaching, research, service and quality were modeled as fuzzy numbers and then aggregated through a weighted ordered aggregator. A single index of performance for each department was generated. The proposal is to identify departments with low performance in one or more dimensions that should receive additional evaluation from an external auditing committee. A by-product of the model is to enlarge the possibility of working with more variables than a conventional DEA model. The model applied to a set of fifty-eight departments of a Brazilian University showed fifteen with low performance. Zero correlation between department teaching, research and service were observed. Weak correlation was detected between research productivity and quality. Weak scale effects were detected.Keywords: data envelopment analysis, productivity analysis, fuzzy logic. ResumoEste artigo trata da avaliação do desempenho -produtividade e qualidade -de departamentos acadêmicos de uma Universidade. Um modelo DEA foi usado para simular um processo de avaliação cruzada entre departamentos. Os resultados de DEA nas dimensões de ensino, pesquisa, extensão e qualidade foram modelados como números difusos e então agregados através de um agregador ponderado. Um único índice de desempenho para cada departamento foi gerado. A proposta é identificar departamentos com fraco desempenho em uma ou mais dimensões que deveriam receber priorização para avaliação por um comitê externo. Um resultado adicional do modelo proposto é a possibilidade de inclusão de um número maior de variáveis do que no DEA convencional. O modelo aplicado a cinqüenta e oito departamentos de uma Universidade Brasileira apontou quinze com baixo desempenho. Nenhuma correlação entre produtividade em ensino, pesquisa e extensão foi observada. Uma correlação fraca foi encontrada entre produtividade em pesquisa e qualidade. Fracos efeitos de escala foram detectados.

Starting in the 1970's, the Brazilian agricultural sector has experienced an important process of modernization, whose principal effects include advances in technological progress and gains in productivity. The primary objective of this paper is to analyze technological progress and total productivity growth in the Brazilian agricultural sector during the period from 1970 to 1996. The methodology used here is based on the Malmquist productivity index and techniques in mathematical programming called Data Envelopment Analysis. The results show that significant progress was made in this sector of the economy but concentrated in only some regions of the country.

Resumo: Este artigo relata um estudo sobre as propriedades de confiabilidade e validade de instrumento de avaliação da docência, utilizando técnicas de análise fatorial exploratória e confirmatória. A hipótese inicial é que o modelo de avaliação da docência da Universidade do Sul de Santa Catarina é formado por oito dimensões correlacionadas: exigência, avaliação, conteúdo, método, relacionamento professor aluno, produtividade, ética e relacionamento professor curso. Estudos confirmatórios revelaram, no entanto, que o modelo apresentava alguns indicadores estatísticos não adequados. Reavaliando a estrutura do instrumento por meio da técnica de análise fatorial exploratória, evidenciou-se uma estrutura com apenas quatro fatores correlacionados, quais sejam: competência, relacionamento e ética, motivação e avaliação. Submetida a testes através de técnicas de análise confirmatória, essa estrutura revelou índices de ajuste global, incremental e de parcimônia adequados sob o ponto de vista estatístico.Palavras-chave: Instrumento Avaliação. Confiabilidade. Validade. VALIDITY AND RELIABILITY OF THE FACULTY EVALUATION TOOLS FROM THE STRUTURAL EQUATION MODEL PERSPECTIVEAbstract: This article describes a study on the features concerned with the validity and reliability of the faculty evaluation tool by using corroboratory and exploratory factorial analyses tecniques. The first hypothesis put forward is that the faculty evaluation model of the Universidade do Sul de Santa Catarina consists of eight correlated dimensions: requirement, evaluation, content, method, professor-student rapport, productivity, ethics and professor-course rapport. Corroboratory studies, however, have shown that the model displayed some inadequate statistical indexes. By re-evaluating the structure of the tool through an exploratory factorial analysis technique, a structure with only four correlated factors; competence, rapport and ethics, motivation and evaluation, became evident. This structure, put to the test through corroboratory analysis techniques, showed indexes of global, incremental and parsimony adjustments as being ideal from the statistical standpoint.

No abstract

The rationale of this study was to jointly estimate response functions for N, P, and K for wheat and soybean cropping in southern Brazil and carryover functions for residual fertility. The two sets of functions were then combined in a dynamic model having general applicability for optimizing fertilizer recommendations. The model is composed of (a) a yield‐nutrient response function incorporating Liebig's Law of the Minimum and Mitscherlich's relative yield theory, (b) fertility carryover functions for P and K, and (c) a multiperiod economic programming model which combines crop yield response and carryover functions with crop and fertilizer prices. Important implications are that (a) the Law of the Minimum assume no nutrient substitution, (b) the relative yield theory allows pooling response data from different experiments, and (c) the fertilization problem is treated as a dynamic process that utilizes revised measures of soil fertility as they become available through soil tests. The crop response was specified as a piecewise linear and plateau function while the fertility carryover was postulated as a first order difference equation, incorporating an exponentially declining process of soil fertility. The solution of the overall dynamic model took the form of a multiperiod linear programming problem. The N, P, and K response functions for wheat (Triticum aestivum L.) and soybean (Glycine max L.) yields were generated from 38 fertilizer field experiments conducted at Rio Grande do sul and Santa Catarina states in southern Brazil between 1968 and 1976. The seven experiments that were repeated for 4 years provided the information for estimating the carryover effects of P and K fertilization. The results indicate that the optimal target levels of soil fertility are in accordance with those recommended locally. The maintenance applications recommended, however, are substantially different than the optimal quantities determined by the dynamic model presented here. The model calls for a reduction of the current maintenance application for both N and P of 10 kg/ha and an increase for K of 12 kg/ha. This would result in a U.S. dollars saving of $27/ha 1 year.

No abstract

No abstract