This is a homework question I've been stuck on for the better chunk of a day, and I feel like I'm overlooking something obvious.
So far I've used the fact that in a finite field with characteristic p, we have that
$(1 + x^p)^n \equiv ((1 + x)^p)^n \equiv (1 + x)^{np} \space \text{(mod p)}$
and that when the left and right ends of this equivalence are expanded via the binomial theorem, I get
$(1 + x^p)^n = {n \choose 0}x^{0} + {n \choose 1}x^{p} + {n \choose 2}x^{2p} + {...} + {n \choose n}x^{np}$
$(1 + x)^{np} = {np \choose 0}x^0 + {np \choose 1}x^1 + {np \choose 2}x^2 + {...} + {np \choose np}x^{np}$
But here is where I'm stuck. By looking at the terms for both of these expressions, I can see that in order for the problem statement to be true, then we require
${np \choose k} = 0$
whenever k isn't a multiple of p. And I've read and understand the proof for
${p \choose k} \equiv 0$ (mod p)
which relies on the facts that k < p and p is prime. But here k need not be less than p, and np is definitely not prime. So I'm not sure how to proceed with the proof at this point.
*Edit: My confusion is on showing that the terms ${np \choose k}$ are all zero whenever k isn't a multiple of p, or at least their sum is. The argument here is that they all must be 0 because we already know that the two sums are congruent mod p, but that seems kind of circular to me? If they're congruent, does each coefficient with a k which isn't a multiple of p have to be exactly 0, or would it be better to say their sum is 0 (i.e. all zero implies sum zero, but sum zero doesn't necessarily imply all zero). Is it because by definition the binomial coefficient is a non-negative integer and thus a sum of non-negative integers summing to zero implies each one must itself be zero?
Any help would be greatly appreciated :)