Know-How Exchange : Research/Metrics

I am a statistician, newly hired as a marketing data analyst. I have several huge databases which include weights for responses. Not just one weight, but about five weights for each respondent, depending on the subgroup in which it is being weighted.

The unweighted data is clearly giving us higher percentages of responses than we would see in the overall population. The data provider has included these weights but no documentation as to how to use them.

My question is: Suppose I want to calculate a weighted percent response to a yes/no question. Normally, I would just take the number of "yes" responses divided by the total N. But how do I now incorporate the weights? Do I multiply each response (coded as 0="no" and 1="yes" in the database) by the corresponding weight? In that case, I then no longer have a dichotomous outcome variable. It is now continuous. So, I was thinking that the correct weighted "percent" is actually now the mean of that weighted continuous variable. But when I do that calculation, it doesnt' match the summary tables provided with the dataset.

Perhaps I'm using the wrong weight of the five choices provided. But before I can determine that, I need to know if I am using the correct procedure.

Thank you for your insight. And yes, I have emailed the data provider but have not yet received a response.