Many of the evaluations summarized in state and local Elementary and Secondary Education Act (ESEA) Title I reports have been negatively affected by methodological errors or the use of different, incompatible formats. Although new, detailed evaluation and reporting procedures have been developed for Title I research projects, two problems remain: bias caused by the tendency to report only favorable results, and inadequacies in the instruments used to measure achievement gain. Some suppression of negative evaluation findings will probably always exist, motivated either by fear of the repercussions, face saving, or the sincere belief that something is wrong with the data. Therefore, it is likely that all aggregations of evaluation data will produce systematically inflated results. It may, however, be possible to estimate the extent of the bias. The testing methods also need improvement. Standardized tests can be chosen to be as relevant as possible to what the students are being taught; and criterion- or objectives-referenced tests may offer a solution to the insensitivity problem--but at the expense of norm-referencing. In short, there is no effective way to interpret large, norm-referenced gains in isolated skill areas. (Author/MV)