Fun With Turnout Projections: 2nd Attempt

Last week with one day early voting turnout totals I boldly, and in retrospect, foolhardily attempted to make a voter turnout prediction based on a comparison to the 2008 primary turnout compared to the turnout on day one of early voting.

Remember the prediction? Based on that one day and setting up a simple proportion I came up with the remarkable number of 13,573 Democratic voters, or roughly 19% of the turnout when it was still undecided who would be the Democratic nominee for President.

Then I compared that number to the 2006 mid-term primary election turnout, which saw 6,550 Democratic voters go to the polls. And I predicted that the actual turnout would be somewhere between those numbers.

Well, I have a slight revision based on another tweak of the data.

Now with 7 full days of early voting results in I have run the same proportion calculation for each day’s cumulative totals. Each day, except for yesterday, the predicted turnout decreased. One day by as much as 13%. Here is the daily analysis showing how the prediction changed from one day to the next.

It varies a bit, and Monday was the only day to have an actual gain over the previous day, a no-brainer because that was a figure for Sunday and none of the polls in churches were open. So with today’s numbers the predicted turnout is 9,037 Democratic voters.

But we have 4 more days in the early vote, so I went ahead and calculated the average daily decrease over all 7 days (-7.3%) and applied that to the next 4 days.

Know what I got as a final vote turnout prediction for the 2010 Democratic primary?

6,671 voters.

Spooky, huh?

No the really spooky thing is that in the contested races, if this turnout holds, 3,336 voters will decide the fate of the Fort Bend County Democratic Party for the next 4 years, at a time when Democratic voters will finally attain majority status in this county.

That is 4.7% of the voters who called themselves Democrats in 2008.

Yikes.