Partial testing of a population of items is routine in manufacturing where statistical quality control procedures include sampling plans that identify how many units of a production run are to be tested. One approach to quality control is acceptance sampling, based on a sampling plan that minimizes total costs of the quality control program. The sampling plan consists of the number of items to be tested and a decision rule that specifies under what sampling outcomes the entire population should be tested. A field evaluation of the acceptance sampling approach to drug testing was conducted in intensive drug supervision programs of six Illinois county probation offices. Sampling plans were identified for drug testing programs that required less than 100 percent testing of collected urine specimens and that yielded total costs lower than the current approach of 100 percent testing. The amount of feedback on drug tests provided to probation officers and thus to probationers was controlled during the evaluation. Counties were assigned to one of three feedback conditions: no feedback, random feedback based on an acceptance sampling plan, or 100 percent feedback. Results showed that using acceptance sampling could have reduced testing without increasing the proportion of those testing positive. The percent testing positive increased throughout the study in counties with no feedback. The authors conclude that an acceptance sampling approach to drug testing may offer improvements over traditional approaches. An appendix tabulates the results of time utilization data collection. 8 references, 23 tables, and 9 figures