A methodology for estimating the relative probabilities of different compromise paths for protected information by insider and visitor intelligence collectors has been developed based on an event-tree analysis of the intelligence collection operation. The analyst identifies target information and ultimate users who might attempt to gain that information. The analyst then uses an event tree to develop a set of compromise paths. Probability models are developed for each of the compromise paths that use parameters based on expert judgment or historical data on security violations. The resulting probability estimates indicate the relative likelihood of different compromise paths and provide an input for security resource allocation. Industrial facilities face insider information theft as well as compromise by visitors. The method should be adaptable to most industrial situations with modifications to fit the specific situation and is direct and simple to use. When historical data are not available, expert judgment data can be used as an input. Even in the absence of quantitative data, considerable insight into espionage risk can be gained by developing the compromise paths and their attendant probability models. These qualitative insights may be the greatest benefit gained when applying this methodology.