The purpose of this project is to develop crash reduction factors for overhead flashing beacons at rural two-way stop sign-controlled intersections in North Carolina. Overhead flashing beacons are a common countermeasure used in North Carolina to help alleviate crash problems at intersections where drivers have difficulty recognizing the stop-control condition. The goal of this analysis is to develop crash reduction factors that reflect North Carolina conditions and decision making. Thirty-four treatment sites were chosen for analysis. Each treatment site was a rural four-leg intersection with no turn lanes and two-way stop control. Each treatment site had at least 3 years of after-period crash data available. Several methodologies were used to calculate the crash reduction factors. The biggest threats to the validity of the analysis that must be accounted for at the 34 treatment sites in this study were regression to the mean and the increase in traffic volumes. Regression to the mean is a significant threat as each treatment site was chosen because of its crash history. The increase in traffic volumes was also a concern because of the long duration of before-and-after periods at each site. Empirical Bayes before-and-after techniques were used to overcome the threat of regression to the mean. One hundred and seventy reference sites were chosen and the method of sample moments was carried out to calculate the necessary parameters. A linear assumption was used to account for the increase in traffic flow. On average, all categories of crashes studied decreased in the after period.