Program impact methodology--usually referred to as evaluation research--is described as it is developing in the United States. Several problems face the field of evaluation research. First, those issues grouped as "meta-scientific" include: (1) the distinction between qualitative and quantitative studies; (2) the separation of implementation and evaluation; (3) maximizing replication and criticism; and (4) evaluation research as normal rather than extraordinary science. Other problems are those of statistical issues. These include: (1) all problems related to the interrupted time-series design; (2) regression adjustments as substitutes for randomization; (3) problems with randomized experiments; (4) attrition and differential attrition; and (5) regression-discontinuity design. Political-methodological problems form a third major group of evaluation research problems, which include resistance to evaluation. It is contended that many new methodological problems will be apparent as the field moves from the laboratory into social problem evaluation. Three figures, seven graphs, and a 107-item list of references are provided. (SLD)