Although virtual reality exposure has been reported as a method to induce paranoid thought, little is known about mechanisms to control specific virtual stressors. This paper reports on a study that examines the effect of controlling the stream of potential paranoia evoking events in a virtual restaurant world. A 2-by-2 experiment with a non-clinical group (n = 24) was conducted with as two within-subject factors: (1) the cycle time (short/long) for when the computer considers activation of a paranoia evoking event and (2) the probability that a paranoia-evoking event (low/high) would be triggered at the completion of a cycle. The results showed a significant main effect for the probability factor and two-way interaction effect with the cycle time factor on the number of paranoid comments participants made and their self-reported anxiety.