We consider interstellar scintillations as a cause of intermittency in radio signals from extraterrestrial intelligence (ETI). We demonstrate that scintillations are very likely to allow initial detections of narrowband signals from distant sources (> 100 pc), while making redetections improbable. We consider three models in order to assess the non-repeating, narrowband events found in recent SETI and to analyze large surveys in general: (I) Radiometer noise; (II) A population of constant Galactic sources undergoing interstellar scintillation,; and (III) Real, transient signals (or hardware errors) of either terrestrial or ET origin. We apply likelihood and Bayesian tests of the models to The Planetary Society/Harvard META data. We find that Models II and III are both highly preferred to Model I, but that Models II and III are about equally likely. Ruling out Model II in favor of Model III requires many more reobservations than were conducted in META *or* the reobservation threshold must be much lower than was used in META. *We cannot rule out the possibility that META events are real, intrinsically steady ETI signals.* We recommend that future surveys use thresholds far below the typical false-alarm threshold to lessen the effects of intermittency. The threshold level is best defined in terms of the recording and computational technology that is available at a cost commensurate with other survey costs.