Accurately weighting political survey response data by age group, region, gender, party (if applicable) and race (if applicable) is the ”special sauce” of a reliable survey. Researchers use a myriad of data sources to determine the proper weighting formula including prior election exit polls or voter files. I personally use voter file data to determine my demographic quotas across several past election cycles, and rarely do age group demographics change by more than 4 or 5 points for a general election.
The 2008 general election was the exception, with the 18 to 34 age group increasingby 10 points from prior election cycles. These voters were overwhelmingly Democrat, Independent and first time voters. Theirincrease in overall contribution to the total vote decreased the 65 and older contribution by 5 to 8 points. Politically the result was a strong year for Democrat candidates. The increase by 10 points among this age demographics occurred in states across the country including Kentucky, Pennsylvania, Colorado, and Illinois. Looking ahead to the 2010 election the big question is if these younger voters will be motivated to show up. If their vote contribution reverts to their traditional 8% to 15% overall contribution it could make for a very bad day for Democrat candidates up and down the ticket.


