This study evaluates the cost-effectiveness of alternative Integrated Pest Management (IPM) training methods and programs aimed at diffusing IPM innovations to farmers in Bangladesh. Various IPM innovations are categorized as being simple, intermediate or complex, while dissemination methods used to extend these innovations are classified as being less intense, moderately intense, or more intense. Examples of less intense diffusion methods include mass media and field day demonstrations, and moderately intense methods include visits from agents. A "farmer field school" (FFS) is considered a more intense type of training method. The study evaluates the effectiveness of these IPM diffusion methods based on a number of criteria such as a methods ability to reach the greatest number of farmers with a given budget, their capability of reaching farmers quickly, and their ability to influence adoption of IPM. Additional components of effectiveness include influencing appropriate use of IPM, influencing retention of IPM, providing a level knowledge that participants can adapt to other areas on the farm, and providing accessibility to limited resource farmers. Data used in the analysis come from a field survey conducted on 350 rice and vegetable farmers in Bangladesh during July and August of 2004. Three ordered probit models are used to measure adoption of simple, intermediate and complex technologies. Results from these models indicate that FFS participants are more likely than non-participants to adopt simple, intermediate and complex practices, while farmers visited by agents are more likely than non-participants to adopt simple and intermediate practices. Field day attendees are more likely to adopt intermediate and complex practices than non-participants. These results may be influenced by endogeneity of unobserved factors that influence participation in FFS but also affect adoption. The study uses a binary probit model to measure appropriate use, but it does not detect any significant differences among farmers trained through different methods, nor does it find any significant differences in retention rates among farmers who have been to different training programs. Another probit model identifies farmers who have been visited by an agent as being more likely to discover an IPM practice through their own experimentation, indicating that farmers who have been visited by an agent have received enough knowledge to adapt IPM to different problems on their farm. Results of the study indicate that limited resource farms are well represented by their participation in the various training methods, but that a higher percentage of large farmers than small farmers in the sample have been visited by an agent or been to FFS. Informal diffusion of IPM information is found to be occurring among people in FFS villages for simple practices. However, for intermediate and complex practices, farmers who participated in FFS are no more likely to spread information to other people in their villages than are less intense formal training methods such as field days and agent visits. This result indicates that while FFS graduates may be likely to share simple practices with others, they should not be relied upon as the primary means of informally diffusing more complex practices to other farmers. The cost-effectiveness estimation incorporates components of the effectiveness evaluation along with the cost of administering the methods. The study calculates that agent visits are the most cost-effective method for diffusing simple and complex technologies and practices, while field days are the most cost-effective methods for extending intermediate technologies and practices.