Knowledge of an explicit, falsifiable psychological theory should not provide the (unaided) knower with the means to falsify it at will in every empirical context. Put another way, psychological theories should not turn out to to be nonfulfilling propheties any more than they should be fulfilling. In practice, this means that the scientist should be confident that an experimental or field design exists that allows the theory to be tested despite the subject’s knowledge of the theory. I call this the non-oxymoron criterion. (pp. 1-2)
Luce, R. D. (1995). Four tensions concerning mathematical modeling in psychology. Annual Review of Psychology, 46, 1-26.