I guess you have ?

But if you have the data you can do a t test, but there you need to calculate .

If you don't have normality you can do an approximate test via the central limit theorem if n is large.

If you don't have normality and n is small I guess you can do a nonparametric test.

You can do a lame approximation of S via range divided by four, but that is based on normality.

SO, I don't exactly know what your instructor is getting at.