I would suggest reviewing your chemical shift tables....
1. C6H5 is a phenyl... Yes, the aromatic hydrogens will be deshielded but not because they are furthest away from an electronegative atom, the aromatic pi system is deshielding in itself. I think you may have "shielding" and "deshielding" backwards....
None the less, there's definitely a hydrogen that would be even more deshielded than the aromatic hydrogens. If it helps, you can think back to acids and bases from gen chem. The most acidic hydrogen is the one that has electron density being pulled away from it (lowest pKa). It works the same with deshielding. If a hydrogen has all of its electrons being pulled away from it towards an electronegative atom like oxygen, it's going to show up way "downfield" (aka to the left of the spectra). This is due to the electron not being able to "protect" the proton from the microwave pulse.
2. A singlet occurs when a hydrogen has no "neighboring" hydrogens on adjacent atoms. A hydrogen on an -OH will always be a singlet even if the carbon the oxygen is connected to has hydrogens on it.
3. Yes, and there are times where you wont even get a peak for the H on an alcohol (frustrating!).
4. The hydrogen on the -CH only has two neighboring hydrogens (remember, hydrogens on an oxygen don't split others and are not split themselves). It's likely that with the 5 hydrogens on the benzene ring you could see either one singlet integrating for 5H, or you might see some stranger splitting patterns (doublet of quartets, doublet of triplets or something along those lines). It's hard to tell without a spectra, but no other hydrogens would give you a quintet signal....
I would suggest reviewing splitting patterns (recall the n+1 rule), the patterns of shielding/deshielding, aromatic hydrogen patterns, and of course keep your tables at hand!
Good luck, hope this helps. Let me know if you have more questions.