This document summarizes a research paper that uses association rule mining to predict software defect associations and the effort required to correct defects. The paper analyzes defect data from over 200 NASA software projects spanning 15 years. Association rules are discovered from the data to predict what other defects may co-occur with a given defect and the estimated effort to correct defects. The predictions are evaluated and found to have higher accuracy than alternative methods like decision trees. The paper also examines the impact of varying the minimum support and confidence levels for rules on the prediction performance and number of rules discovered.