In my algebra class, we went over the theorem:
Suppose that $N$ is a normal subgroup. Then $G/N$ is abelian iff $[G:G]\subseteq N$
The proof we used seemed to imply that the normal subgroup is the commutator. I know that this observation is actually meaningless, an artifact of the exact proof we used, but my professor couldn't think up an example to the contrary, neither could wikipedia, and I would like one.
Just to be clear, by proper, I mean also non-trivial, rather than simply of a smaller cardinality.
Often one can find examples by looking at groups of small order. The quaternion group $Q_8=\{\pm 1,\pm i,\pm j,\pm k\}$ has commutator subgroup $[Q_8,Q_8]=\{\pm 1\}$. All subgroups of order $4$ contain the commutator subgroup as a proper subgroup; namely $\{\pm 1,\pm i\}$, $\{\pm 1,\pm j\}$, and $\{\pm 1,\pm k\}$. Moreover all subgroups of order $4$ are of index $2$, and hence normal subgroups.