The installation and maintenance of traditional traffic data collection technologies such as inductive loops can disrupt the smooth flow of traffic, undermine pavement quality, and be a significant safety concern to both motorists and construction personnel. Consequently, there has been a strong emphasis on the development and deployment of non-intrusive traffic detection systems. These devices serve as alternatives to conventional systems by allowing agencies to collect data at high traffic volume locations without compromising operations or safety. The main objective of this research is to investigate the accuracy, reliability, cost, and user-friendliness of various non-intrusive detectors. The study is based on data collected at a test bed in Nebraska where four types of non-intrusive detectors (Autoscope Solo Pro II video image detection system, ISS RTMS G4 microwave radar system, Wavetronix SmartSensor microwave radar system, and GTT Canoga Microloop magnetic induction system) have been deployed. A paired or yoked statistical analysis is used to compare the technologies "side-by-side" so that any environmental and traffic effects are accounted for. The study identifies the relative accuracy of each detection system for each traffic variable (volume, speed, and classification) as a function of environmental conditions, time of day conditions, and traffic conditions. In addition, the variances of the traffic parameter estimates are also provided so that transportation agencies will have reasonable operating bounds for the different technologies which are key input to their ATMS algorithm designs.