Severe weather outbreaks are frequent phenomena across much of the central and eastern United States. No study has attempted to determine of the extent to which synoptic-scale features control the difference between tornado outbreaks and non-tornadic severe weather outbreaks, information that would be of great value to a forecaster. This study uses synoptic scale NCEP/NCAR reanalysis data to perform a principal component analysis on five diagnostic variables, u and v wind, temperature, relative humidity, and height, at all 17 of the reanalysis pressure levels. The goal of this analysis was to determine features in each outbreak type that were unique, providing the ability to discriminate between the types. Owing to artificial bias introduced by northward converging longitude lines, the reanalysis data were interpolated onto an equally-spaced grid in the horizontal. Three-dimensional composite fields were created for the two outbreak types to determine a composite depiction of each. 50 case studies of tornado outbreaks and 49 studies of severe weather outbreaks were used in the creation of the composites. Features of both tornado and non-tornadic severe weather outbreaks were outlined in the study, emphasizing those features that aid in distinguishing the outbreak type.