AbstractIn the past, time‐temperature superposition has been used to extend the time scale of creep tests in polymers from the short times easily obtained in the laboratory to long times seen in actual use. A fundamental assumption of time‐temperature superposition is, however, that the polymer does not change in structure as a function of time. Because ductile amorphous thermoplastics physically age below Tg, the structure of the polymer changes on a time scale comparable to the time duration of the creep test. Thus, the time‐temperature superposition prediction greatly exaggerates the amount of creep in amorphous thermoplastics. For samples aged at the test temperature for one hour before testing, the difference between the time‐temperature superposition prediction and the actual creep data after 10 days is greater than a factor of ten in time.