PDiddie adds on to my post about the Harris County sample ballot, and includes a numerical tidbit of interest:

Harris County election officials project that 1.2 million votes will be cast here.

Let's assume this is an accurate estimate. How can we project statewide turnout from it? Well, Harris County represented 13.5% of the state total of 7.4 million votes in 2004. If the Harris/Texas ration holds this year - or, to put it another way, the state total increases by the same 20% as Harris - we're talking almost 8.9 million votes cast. That's a hell of a lot, probably too much more. I think eight million is well within range, but nine million feels like a stretch.

Another way to look at this is to consider turnout in Harris County. Statewide turnout in 2004 was 56.6%, which was slightly less than the 58.0% of Harris. I'm going to guess that 1.2 million votes in Harris County represents 60% turnout, which is to say 2 million registered voters. I think we should have 2.1 million registered voters, but I don't think we'll get it that high. Assuming 60% Harris turnout, that translates to 58.6% state turnout at that same ratio, so now we need to guess how many registered voters there will be. At 13.1 million, the same as 2004, we'd get 7.67 million votes. We know there's more registrations than that, so adjust accordingly. For eight million ballots statewide at this, we'd need 13.66 million registrations. To get to 8.9 million, we'd need 15.2 million registrations. I don't see that happening.

That's an awful lot of numbers and assumptions, any or all of which may be full of it. I'll take another crack at it after the voter registration deadline when we'll have some hard totals. I do think we'll get to 8 million votes - it's really not that big an increase from 2004 percentagewise - but I'll need some better data to feel confident about anything more specific than that.