Some aspects of fourth generation evaluation procedures that have been advocated by E. G. Guba and Y. S. Lincoln were examined empirically, with emphasis on areas where there have been discrepancies between theory and field-based experience. In fourth generation evaluation, the product of an evaluation is not a set of conclusions, recommendations, or value judgments, but rather an agenda for negotiation of claims, concerns, and issues. This approach is distinguished from the earlier three generations of evaluation which dealt with measurement, description, and judgment respectively. Approximately 10 evaluations of educational programs conducted each year by the Curriculum Research and Development Group of the University of Hawaii College of Education over the past 15 years provided data for this analysis. An example involving the evaluation of federally-funded projects is provided. The data highlight dilemmas in the attempt to provide fourth generation evaluation, beginning with the expressed desires of stakeholders for the sort of information earlier generations of evaluation provided. An internal inconsistency appears to exist in the views of Guba and Lincoln about the impossibility of generalizing from one situation to another, even as they generalize about scientific theory. Guba and Lincoln have offered a theoretical model that promises to enhance evaluation if it can be applied in the real world, although the benefits of earlier models cannot be denied. (SLD)