If the two sets of protons (ortho and meta to the para methyl group) are at similar resonance they will have produce smaller coupling constants. Your apparent singlets may actually be broadened doublets. This broadening is common for molecules with hindered rotation on the nmr time scale. Perhaps this is happening with your tosyl amide.
This is a particular situation (not rare) where you have a magnetic equivalence of ortho & meta protons. To be sure that it is your case, you can do a simple heteronuclear 2D experiment such as 1H-13C HSQC. You will brake the symmetry as your 13C are surely inequivalent. If you find again an equivalence in carbon, it will be strange.