I recently tried to estimate relative levels of support for Iraqi terrorists and for the constitutional process.
My estimate was that participators in the constitutional process outnumbered supporters of terrorism by at least four to one. This was calculated from the lowest end estimate of turnout in the election, around 50%. If the turnout was nearer 60%, as seems likely, then the margin is more like eight to one.
The national turnout in yesterday's referendum is not yet known. It is clear, however, that in Sunni areas, where turnout in the election was worryingly low, turnout in the referendum was considerably higher. By contrast, in some Shia areas turnout seems to have dropped.
If overall turnout is around the same, but more evenly spread, this will be a good sign indeed.
Since we don't yet have the full turnout figures, it is not yet clear why turnout has dropped in some areas. Certainly in Shia areas there has been more violence lately, even as the terrorists are retreating in the centre of the country. But it could also reflect rational choice about voting as well as combination of dissillusion, boycott and intimidation. There is little doubt that the overall vote will be in favour of the constitution. What remains in clear doubt is whether some Sunni areas will clearly defeat the constitution and thus cause a renegotiation. In areas where the constitution is sure to pass there is less incentive to vote than in areas where the result is in doubt, so campaigning has been harder in the latter areas.
On exactly the same reasoning, turnout in US elections tends to be higher in swing states than in states that are clearly red or blue.