A note on the accuracy of the auroral electrojet indices