Several related methods for determining the calibration formula of a continuously scanning bridge circuit using on‐line data are analyzed and discussed. High‐resolution resistance thermometry requires a sensitive, accurate conversion of off‐balance voltage measurements to resistance measurements. The objective is to properly determine the relationR=f(V,RB,Ci), whereVis the bridge output‐voltage,RBis a balance resistance, and theCiare unknown coefficients that depend onRBand other fixed circuit parameters. Linearization to the formR = RB + C1(RB)Vcan yield a fairly accurate conversion fromV(t) toR(t). IfR˙(t) or higher derivatives are required, linearization leads to systematic errors and a more sophisticated method is needed. From the analysis of three techniques for determining corrective terms, we conclude that parameterization of a model for the bridge circuit is the most satisfactory procedure. An empirical, self‐consistent method is useful if the circuit model is unknown or extremely complicated. Calibration data from a high‐resolution specific‐heat experiment were used to test the various methods.