This paper examines major contributions to error when calibrating a microelectromechanical system (MEMS) gyroscope using a miniature vibratory micro-stage. Calibration via this micro-stage has been previously introduced by the authors to provide a reference stimulus, allowing gyroscope scale factor to be calibrated repeatedly over time. This requires that the reference stimulus can be measured with extremely high accuracy. Scale factor calibration to errors a few hundred parts per million has been demonstrated. However, determination of limiting factors for calibration performance remains of great importance for future device improvement. In this paper, the gyroscope model, stage dynamics, and calibration approach are briefly described. Then, sensitivity of calibration errors to underlying stochastic and parameter identification errors is analyzed. Analysis results indicate that even when signal-to-noise ratio is sufficient to for high accuracy (<;100 ppm) calibration with a perfect model, noise-induced limits on model identification inaccuracy have substantial impact on total velocity and scale factor accuracy.