It is inherent in any acceptable definition of science that statements that cannot be checked by observation are not really about anything or at the very least, they are not science.
The Nonprevalence of Humanoids, Science, Volume 143, Number 3608, February 21, 1964 (p. 770)