Gini's coefficient of mean difference as a measure of adoption speed: theoretical issues and empirical evidence from India
提出基尼平均差作为衡量技术采用速度的无分布方法，用印度高产小麦品种数据证明其优于逻辑和威布尔系数，适合事后分析单调递增的采用过程。
ABSTRACT This paper represents a variant on the empirical analysis of technological adoption by Griliches. It follows up on previous research regarding the potential of the Gini coefficient of mean difference as a distribution‐free measure of adoption speed. A formal proof is supplied for the derivation of an empirically workable definition of the Gini. Data regarding the uptake of high‐yielding wheat varieties in India are used to compare the performance of the Gini relative to that of the logistic and Weibull adoption speed coefficients. The results suggest that, for the purpose of ex‐post analyses of monotonically increasing adoption processes, use of the Gini in measuring adoption speed can result in a better explanation of aggregate adoption behavior.