We compare TEC measurements from the NASA Radar Altimeter and DORIS instrument on board TOPEX/POSEIDON with GPS TEC estimates, and evaluate different GPS data analysis strategies. We verify that global tomographic GPS analysis using a voxel grid is well suited for ionospheric calibration of altimeters. We show that a 1-day fit of 20-second-averaged NRA ionospheric correction data versus GPS tomographic TEC data has a bias of 3.4 TECU and a root mean square deviation of 3.2 TECU. Tomographic inversion using simulated data from the Parametrized Ionospheric Model highlights the strong correlation between GPS bias constants, electronic densities at the highest layer, and unmodeled protonospheric TEC. This suggests that GPS TEC estimates at the TOPEX/POSEIDON altitude are more accurate if the bias constants are estimated and if a layer above TOPEX/POSEIDON is added to the grid.