Which Impedance-Related Function to Fit

The next problem is which function Z, Y, M, or e to fit. The answer is that it is most sensible from a statistical point of view to fit the data in measured rather than transformed form. Suppose that Z is measured in rectangular form. When both Z and the associated Y = Z data are separately fitted with P weighting, it is found that there are often significant differences between the parameter estimates obtained from the two fits. This is not unexpected the operation of taking an inverse (complex or not) on data with errors generally introduces a bias in the fitted results it is for this reason that the directly measured results should be fitted directly. 

In most antomated measurements the rectangular components are measured directly, but sometimes the modulus and phase angle are directly obtained. These are the two quantities,/ = / , and/ = 0 = which would then appear in Eq. (2) 

---