Abstract: With certified reference material (CRM) and reference material (RM), 11 kinds of test methods (including 6 kinds of test methods used in TENSTAND report) were used to measure the yield and proof strength. It was confirmed that without considering stiffness, crosshead control and strain control could get nearly the same results with the same nominal speed. Analyzing the reason of recommending strain control method for yield and proof strength and considering stiffness in the tensile test method standards, it was found that there were mistakes in Fig. 38 and Fig. 39 of TENSTAN WP4 report which misleading the conclusion of TENSTAN WP4 report, and then the report misled ISO 6892-1-2009 as well as GB/T 228.1-2010. Through simulated tests, it was found that Fig. 38 had been controlled with real time stress rate, which had been switched to crosshead control when it could not be maintained, but not with crosshead control mode as reported. Analyzing the reason why a higher proof strength got with crosshead control mode than that got with strain control mode, it was found that the error was not because of crosshead control but considering stiffness. The theoretic source inducing the mistake of considering stiffness was also reported as misusing Hooke’s law in plastic range to measure yield and proof strength.