Generalized correlation diffusion ordered spectroscopy (GECO-DOSY) has been employed to increase the resolution of a more traditional DOSY experiment. We have compared purely linear polycaprolactone (PCL) with that of a sample that also contains impurities, probably small cyclic oligomers formed as a side product in the polymerization experiment. In low concentrations, cyclic PCL with more than two monomer units is not normally detectable by H-1 NMR, as there is little chemical shift contrast between the macrocyclic and linear environments. However, analysis with GECO-DOSY allowed mathematical deconvolution of the NMR spectrum and revealed that additional signals were present that were masked by the more intense linear PCL. This is the first use of GECO-DOSY for analysis of polymer experimental data and has proved that the technique has useful applications for polymer characterization. We also examine the effects of instrument errors commonly associated with DOSY in conjunction with the use of GECO-DOSY.