Full-scale fatigue testing of wind turbine blades is conducted using resonance test techniques where the blade plus additional masses is excited at its first resonance frequency to achieve the target loading amplitude. Because there is not a direct relationship between the force applied by an actuator and the bending moment, the blade is instrumented with strain gauges that are calibrated under static loading conditions to determine the sensitivity or relationship between strain and applied moment. Then, during dynamic loading the applied moment is calculated using the strain response of the structure. A similar procedure is also used in the field to measure in-service loads on turbine blades. Because wind turbine blades are complex twisted structures and the deflections are large, there is often significant cross-talk coupling in the sensitivity of strain gauges placed on the structure. Recent work has shown that a sensitivity matrix with nonzero cross terms must be employed to find constant results when a blade is subjected to both flap and lead-lag loading. However, even under controlled laboratory conditions, potential for errors of 3 percent or more in the measured moment exist when using the typical cross-talk matrix approach due to neglecting the influence of large deformations and torsion. This is particularly critical when considering a biaxial load as would be applied on the turbine or during a biaxial fatigue test. This presentation describes these results demonstrating errors made when performing current loads measurement practices on wind turbine blades in the lab and evaluating potential improvements using enhanced cross-talk matrix approaches and calibration procedures.