Unweighted and Weighted Data (Chi Square)

Here's my situation:

I have a dataset with unweighted data taken from a small, annual survey. I also have a large dataset with weighted and unweighted data from a larger survey taken every five years. I am trying to compare the groups on a number of variables to determine if the small, annual survey is representative of the estimates gleaned from the larger survey. I have frequency distributions and was planning on using chi-square to test for statistical signficance between groups. First, is this the right approach?

Second, what is the procedure for chi-square with weighted and unweighted frequencies? Can I use it to compare weighted frequencies with unweighted frequencies? If not, should I only use unweighted frequencies, and if so will this be in any way accurate with a data set that was designed for weighting?

Thanks for your help!