The following fact is stated (but not proved) in Chapter 0 of Noll's Finite Dimensional Spaces (it's 07.21):
If $R$ is a commutative ring with unity and $r_1, ..., r_n$ elements of $R$ then $$ n!\text{ }r_1\cdots r_n=\sum_{k=1}^{n}(-1)^{n-k}\sum_{1\le i_1<\cdots<i_k\le n}(r_{i_1}+\cdots+r_{i_k})^n $$
Now as it happens I (think I) can see a proof of this.
Start by showing that it holds when setting $r_{1}=X_{1}, ..., r_n=X_{n}$ in $\mathbb{Z}[X_{1}, ...,X_{n}]$. To do this, first show that the RHS is zero at $X_{i}=0$ so is divided by $X_{i}$, so is divided by the product $X_{1}\cdots X_{n}$ (by UFD-ness). Thus by considering degrees the RHS is a constant times this product. Setting $X_{i}=1$ then reveals this constant and gives the polynomial identity.
Then use the free-ness of $\mathbb{Z}[X_{1}, ...,X_{n}]$ to map these elements onto any $r_i$ in any commutative ring with 1, to deduce that the identity holds there.
However seeing as how this fact is stated without proof in the introductory chapter to a not particularly advanced textbook (a textbook that does not have knowledge of abstract algebra as a prerequisite), and appears in the company of some trivial facts like the binomial theorem, I was wondering if there was a more elementary proof of the identity?