April 1999 - The sensitivity of empirical results to potential data errors and model misspecification can yield misleading policy implications and investment signals. A widely disseminated study of the impact of the training and visit (T&V) system of management for extension services in Kenya is a striking example of how innocuous data errors and alternative specifications lead to strikingly different results. Gautam and Anderson revisit the widely disseminated results of a study (Bindlish and Evenson 1993, 1997) of the impact of the training and visit (T&V) system of management for public extension services in Kenya. T&V was introduced in Kenya by the World Bank and has since been supported through two successive projects. The impact of the projects continues to be the subject of much debate. Gautam and Anderson's paper suggests the need for greater vigilance in empirical analysis, especially about the quality of data used to support Bank policy and the need to validate potentially influential findings. Using household data from 1990, Bindlish and Evenson found the returns from extension to be very high. But Gautam and Anderson find that the returns estimated by Bindlish and Evenson suffer from data errors, and limitations imposed by cross-sectional data. After correcting for several data processing and measurement errors, the authors show the results to be less robust than reported by Bindlish and Evenson and highly sensitive to regional effects. When region-specific effects are included, a positive return to extension cannot be established, using Bindlish and Evenson's data set and cross-sectional model specifications. After testing the robustness of results using a number of tests, Gautam and Anderson could not definitively establish the factors underlying strong regional effects, largely because of the limitations imposed by the cross-sectional framework. Household panel data methods would have allowed greater control for regional effects and would have yielded better insight into the impact of extension. The impact on agricultural productivity in Kenya expected from T&V extension services is not discernible from the available data, and the impact may vary across districts. The hypothesis that T&V had no impact in Kenya between 1982 and 1990 cannot be rejected. The sample data fail to support a positive rate of return on the investment in T&V. This paper-a product of the Sector and Thematic Evaluation Division, Operations Evaluation Department-is part of a larger exploration by the department of the effects of the investment in agricultural extension in Kenya. The authors may be contacted at mgautam@worldbank.org or janderson@worldbank.org.