You can look at the problem in terms of coefficients in the expansion of (x+y)^n, as Matt suggests. Another technique I like to use - and this requires thinking about how and what you are counting - is to find an experiment that involves selecting (and ordering) objects which can be looked at in two different ways.
In this case, you ask yourself how many ways there are of picking 'any' number of objects out of the given set of n objects. One way, is to find the no. of ways of picking zero objects + the number of ways of picking 1 object + ...+ the no. of ways of picking all n objects. So that covers all cases, and is exactly the LHS of the equation you want to prove. The other way of do this same thing is to look at each object and decide whether to pick it or not. This way, you are assigning 2 values to each of the n objects. The total number of ways of doing such an assignment is your RHS.
This kind of approach is probably just what Matt was talking about. The above example should give you somewhere to start from.