Julian,
It is well known that the ELSD is not a direct linear response detector however linear calibration plots can be obtaned for several orders of magnitude by using double logrithnmic coordinates such as:
LogA=bLogC +loga
where C is your analyte concentration expressed in mgL-1 (for example) and A is the peak area response obtained after integration of the chromatographic peak of each analyte.
In other words quantitation and method validation is possible. For more information look at:
1) Mourey TH and Oppenheimer LE Anal Chem 1984 56, 2427-2434
2) Stolyhwo A et al J Chromatogr. 1984, 288, 253-275
3) Dreux M and Lafosse M, LC GC Int. 1996, 9, 148-156
4) Petritis et al. Chromatographia, 2004, 60, 293-298.