Geodetic observations in the oceans are important for understanding plate tectonics, earthquake cycles and volcanic processes. One approach to seafloor geodesy is the use of seafloor pressure gauges to sense vertical changes in the elevation of the seafloor after correcting for variations in the weight of the overlying oceans and atmosphere. A challenge of using pressure gauges is the tendency for the sensors to drift. The A-0-A method is a new approach for correcting drift. A valve is used to periodically switch, for a short time, the measured pressure from the external ocean to the inside of the instrument housing at atmospheric pressure. The internal pressure reading is compared to an accurate barometer to measure the drift which is assumed to be the same at low and high pressures. We describe a 30-months test of the A-0-A method at 900 m depth on the MARS cabled observatory in Monterey Bay using an instrument that includes two A-0-A calibrated pressure gauges and a three-component accelerometer. Prior to the calibrations, the two pressure sensors drift by 6 and 2 hPa, respectively. After the calibrations, the offsets of the corrected pressure sensors are consistent with each other to within 0.2 hPa. The drift corrected detided external pressure measurements show a 0.5 hPa/yr trend of increasing pressures during the experiment. The measurements are corrected for instrument subsidence based on the changes in tilt measured by the accelerometer, but the trend may include a component of subsidence that did not affect tilt. However, the observed trend of increasing pressure, closely matches that calculated from satellite altimetry and repeat conductivity, temperature and depth casts at a nearby location, and increasing pressures are consistent with the trend expected for the El Niño Southern Oscillation. We infer that the A-0-A drift corrections are accurate to better than one part in 10⁵ per year. Additional long-term tests and comparisons with oceanographic observations and other methods for drift correction will be required to understand if the accuracy the A-0-A drift corrections matches the observed one part in 10⁶ per year consistency between the two pressure sensors.