Having fitted an nth order polynomial to one data set, how can I now fit that exact curve to a second data set and obtain goodness-of-fit data?
The resulting curve will be the same one that you obtained from the first analysis--since you have fixed all the parameters, Prism actually has no fitting to do. Nonetheless, it still reports sum-of-squares and Sy.x to quantify goodness-of-fit to the second data set.