Some Extensions of the Nerlove-Press Model

摘要

In this paper we have reformulated the Nerlove-Press model in terms of an unobserved underlying variable with a continuous and three dummy indicators. In the example on agricultural practices of Phillippino farmers, the underlying variable is adoption of modern agricultural methods. Our basic problem is to study the determinants of this variable. In the case of the example on sexual attitudes of undergraduates at Northwestern University, the underlying variable is sexual permissiveness and our objective is to study the determinants of sexual permissiveness. Several models were estimated to study the determinants of the former latent variable using Mangahas' data that was used by Nerlove and Press. It was found that schooling, CO-OP and IRRIG were significant and positive and CROP was significant and negative and were the important determinants of 'adoption of modern agricultural methods.' The other explanatory variables were not significant. The results are encouraging since the schooling variable was not significant in the Nerlove-Press analysis. In their analysis only IRRIG, CO-OP and AREA were significant. It should, however, be admitted that the schooling variable, though statistically significant, had a very small coefficient and the experience variable was not significant. Thus, the results reported here are not substantially different from those of Nerlove and Press. The formulation adopted here should, however, prove useful in several other applications. The model we have estimated imposes a different and more readily interpretable structure on the data than that adopted by Nerlove and Press.