Corrosion tests, fretting tests and other environmental endurance tests are performed to provide information on performance in use or to demonstrate compliance with standards or specifications. In this paper, pitfalls concerning the use of tests to predict the durability of electrical connectors in the field are discussed. The discussion covers corrosion testing resulting in unrealistic chemistries, fretting tests concentrating on incorrect failure mechanisms, self-healing effects during testing resulting in misleading results, as well as lack of evaluation of effects of testing, such as contact resistance measurements under conditions not revealing surface films formed. Testing under unrealistic conditions or without proper evaluation of results may not only risk not detecting a poor design but also to reject a good one.