Applied to evaluation design, Murphy's law states that in any given cost of manpower estimate, the resources needed to complete the evaluation will exceed the original projection by a factor of two; after an evaluation design has been completed and is believed to control for all relevant variables, others will be discovered, and rival hypotheses will multiple geometrically; and the necessity of making a major decision change increases as the evaluation project nears completion. Applied to evaluation management, the law states that the probability of a breakdown in cooperation between the project and an operational agency is directly proportional to the trouble it can cause, and that the total time needed will accumulate in the direction of becoming further behind schedule. When applied to data collection, the law states that the availability of a data element is inversely proportional to the need for it; historical baseline data will be recorded in units or by criteria other than present or future records; and none of the available self-report formats will work as well as expected. When Murphy's law governs data analysis and interpretation, it states that mathematical errors will accumulate in the direction of most damage to the results, that the figure most obviously correct will be the source of error, and that the line totals which should add up to a grand total, won't be correct. Finally, Murphy's law governing the presentation of evaluation findings states that the more extensive and thorough the evaluation, the less likely the findings will be used by decisionmakers.