After much experimentation I have found that the most consistent way of measuring EW (for me at least) is to first rectify the spectrum to unit intensity for the continuum the use a function fit. I have found this simplest to do using SPLAT-VO. This has a modern interface and the best integrated set of tools with one exception. SPLAT allows polynomial and free hand fitting of the continuum, functional fitting of line with the standard functions for measuring the lines. It also has “flip and translate” of a spectrum for measuring line shifts. It is only for blended lines and emission lines where IRAF has better tools.
Here are four spectra of Rho Cas with 3 Fe I lines near 5572 A that I intend to monitor along with others. The spectra match to within a few percent in the continuum but are offset here for clarity. The lower two spectra are at half the exposure of the top two. The poor S/N on the first one is due to the very poor seeing at the time.