I have created a set of crosstabs that show the effect of re-weighting the data with renters represented in their proper proportion of households (just over half the sample).
DOWNLOAD CROSSTABS.
This reweighting by own vs. rent also brings down overall income levels, since Shorewood renters have a lower income profile than homeowners. I created these crosstabs relatively quickly (yes, I have a day job), and it is possible I made an error or two, so let me know if you see anything that does not seem right. I see this as a completely "open source" project (the data is a public
record), so let me know if you want to the SPSS files, or the Excel
version of the crosstabs. I am open to suggestions, revisions,
criticism, and personal insults, if you would like to participate in this.
Why, Steve? Is this some sort of methodological crusade you are on? Some nutty academic exercise?
Not at all. There is nothing wrong with using unweighted data, especially since part of the intent of the 2008 Community Survey was to track the changes in opinions over time. However, in this case, using raw responses has the additional effect of discounting the importance of renter households. I believe it is important to look at Village opinion with renters' opinions given its proper value, proportional to their share of the households in our Village.
There is a lot of interesting data in here, which could make for a great deal of interesting discussion.
As a side
note, I weighted to a count of 600, which means renters are weighted at about 1. Homeowners are weighted down to the proper
proportion within a weighted sample of 600. There is a lot of missing
data in individual surveys, so you will notice that the weighted counts bounce
around, and never quite reach 600.
