Hip position has been hypothesized to influence gravity effect torque (GET) at the knee during isokinetic testing; however, no data exist to support or refute this hypothesis. Therefore, the purposes of this study were 1) to determine if a significant difference exists between GET in seated and supine positions, 2) to determine the effect of the supine and seated GET on isokinetic peak torque values, and 3) to determine the relationship between hamstring flexibility and GET. Gravity effect torque was recorded in supine and seated positions. Peak torque values in flexion and extension were obtained on a isokinetic dynamometer at 1.047 and 5.235 rads.s−1 (60 and 300° s−1, respectively). Hamstring flexibility was assessed by the active knee extension test (AKET). The mean seated GET value was 5.64 Nm higher than the mean supine GET value (F(1,82) = 97.85, P = 0.0001). Significant correlations existed between hamstring flexibility and GET values measured in the seated and supine positions (r = 0.45, P = 0.0001, and r = 0.30, P = 0.0058, respectively). Significant differences in peak torque values occurred for three of the four isokinetic conditions when using different GET values (P-value 0.0002–0.0049). Although mean differences in peak torque values were only 2.43–4.23 Nm, these differences may translate to significant errors in the isokinetic measurement of the injured population undergoing rehabilitation. Furthermore, every attempt should be made to improve the validity of isokinetic testing. Therefore, we recommend the supine position for GET determination.