Evaluating as a profession seeks to improve governmetn by providing information on the impacts and value of public policies and programs to administrative and citizen stakeholders. In order to such efforts to contribute to meaningful governemtnal reform, it is important that the methodologies we employ yeld meaningful information. This is not a straightforward as one might think, as evidenced by the accumulating literature on "lessons leanerd" in evaluating governemtn policies and programs. If we ae to derive the most benefit from multinational conferences and associations, we need to encourage a dialogue on the lessons learned in each country or each professional community. This paper moves in that direction by presenting and, hopefully, generating feedback on three lessons learned in the evaluation community in North America, wich examples provided in the area of human services programs. In brief, one lesson learned involves recognizing that loyalty to specific evaluation methods often interferes with addressing the practical questions confronting decision-markers. Sometimes this take the form of using performance measurement techniques when a more sophisticated is required to identify program impacts. Other times the opposite mistake is made, using a complex design when simpler methods would have been adequate, and more timely, to address the real questions at hand. This pragmatic view of methods highlights a second lesson, the rejection of the traditional dichotomy between quatintative and qualitate analyses when attempting to assess the effectiveness of government programs. Recognizing that judgment of effectiveness involve causal inferences, pragmatic ecommendations for evaluation now routinely suggest a mixture of methods, particularly given the multiple audiences to be served in promoting accountability. Finally, a third lesson involves the subtle and not-so-subtle roles that values play in influencing the methods employed and conclusions reached in studies of government reform. Rather than presuming either value-neutrality (values-relativism) or basing conclusions on a single value foundation (e.g., Kaldor-Hicks efficiency for benefit-cost evaluation or Rawlsian equity for social critique), recent approaches have confronted the issues of "managing" the values that influence evaluation. In sum, government reform contains the value-based implications on "improvement". Our efforts at reform, therefore, depend on the adequacy of our methods for representing the efects of these efforts and for supporting judgements of their value. We have all learned lessons about the evaluation methods that we use to support reform; we gain by promoting a dialogue about these evaluation lessons