The repeatability of a K+-selective glass electrode and a Cu2+-selective solid-state electrode was assessed by the determination of the variation of the slope and potentials in successive calibrations, and by analysis of covariance of the regression lines of potential vs. logarithm of ion activity. When assessed by any one of these three methods, both electrodes displayed high repeatability in most of the calibration sets. However, in several calibrations, the variation of the slope and potential was low, but the regression lines were statistically different. In no case was the variation of the slope and potential high and die regressions not statistically different. The results suggested that analysis of covariance can be used to test electrode repeatability, and provides a more restricted evaluation of this parameter than the variation of the slope and potential. The characteristics of analysis of covariance suggest that it provides an objective and rigorous test for electrode repeatability. The advantages of the use of analysis of covariance over the determination of the variation of the slope and potential are listed.
