Real-time embedded systems are becoming ever more complex. We are reaching the stage where even if static Response-Time Analysis (RTA) was feasible from a cost and technical perspective, the results of such an analysis are overly pessimistic. This makes them less useful to the practitioner. In addition, the temporal validation and verification of such systems in some applications, e.g., aeronautics, requires the probability of obtaining a worst-case response time larger than a given value in order to support dependable system functions. All these facts advocate moving toward statistical RTA, which instead of calculating absolute worst-case timing guarantees, computes a probabilistic worst-case response time estimate. The contribution of this paper is to present and evaluate such a statistical RTA technique which uses a black box view of the systems under analysis, by not requiring estimates of parameters such as worst-case execution times of tasks. Furthermore, our analysis is applicable to real systems that are complex, e.g., from a task dependencies perspective.