Nineteen states didn't have exit polling in 2012. My own state was one of these, and this obviously frustrated me a great deal, considering that detailed exit polling in Georgia would have been vital in determining any changes in trends (after all, GA was the second-closest state that Romney won).
I have access to three congressional districts in Georgia that allow me to take a snapshot of the electorate in 2012. What is great about this system (at least, in GA's case) is that deceased voters/others are filtered out automatically
but not removed from the system for 4-6 years.
There are also scoring models that help organizers determine the likelihood of an individual to do any number of certain things (turning out, voting Democratic/Republican, voting Obama/Romney, supporting gun control and climate change, etc). While these scoring models can be somewhat inaccurate at the individual level, they are
very accurate when assessing a larger group of voters.
When you take the entire voting bloc of an area and analyze it via these models, you can more or less conduct your own exit polling. Since...- deceased and otherwise now-inactive voters can still be analyzed
- you can filter results to show everyone who voted in 2012
- scoring models when assessing hundreds of thousands of voters at a time are very accurate
...I have been able to produce some "exit polling" of my own for these three congressional districts (GA-9, GA-11 & GA-14) that is perhaps even more accurate than actual exit polling. The results I got seem spot-on, although the Asian numbers are odd (we have mostly Vietnamese populations here, so that could be a factor). At any rate, here are the results I was able to fetch for the aforementioned areas. I'm hoping some of you out there with access to the same systems can provide similar insight to your own areas, or even if you don't, give us some approximate insight into what you expect the results were in your own districts.