I have been puzzled by the pollsters sometimes sampling Democrats and Republicans equally and then within months under sampling Republicans. I started looking for a reason. They are pretty sure what Democratic turn out will be. It is a trend line. But they have to be making educated guesses at Republican turn out.

From 1980 through 2004 the Democratic turn out in presidential elections has been. 35, 37, 41, 44, 47, 51 and 59 million voters. That is a consistent pattern of turn out. On the other hand Republican turn out over the same years has been 43, 54, 48, 39, 39, 50, and 62 million. Consider that the Perot vote fell from 19 million in 1992 to just 10 million in 1998, yet the Republican vote did not change.

It may not be enough to predict a trend, but the 2000 and 2004 Republican returns paralleled the Democratic returns. However in the preceding 20 years  Republican turn outs were up and down and all around.

I think I can explain it as the Karl Rove effect.

I grew up in a Democratic family heavily engaged in the grass roots efforts of that party. The Democrats had a superb if not entirely legal ground game. I was shocked when I first got involved in Republican politics. There was no ground game ... at least not in Ohio.

I remember when I first got involved in a Republican Mayoral race back in the 1970s. I laid out a grass routes plan for use in the campaign. A local Republican told me in no uncertain terms that I was not from HIS County. He added I certainly did not understand how things worked in HIS county. His candidate got the crap beat out of him largely due to the Democrat's ground game.  His attitude was sort of like REAL GENTLEMEN don't do ground games.

But then came Karl Rove. And volunteers were organized into something approaching the precinct captains order I had known in the Democratic party. Voter lists were kept. Serious efforts made to get out the vote. That continues. I don't think we will see the up and down turnouts of 20 the century in the Republican party from now on.  Karl Rove has seen to that.

This historically bumpy road of Republican turn outs  may be how the pollsters are justifying the under sampling of Republican voters. They can predict the Democratic turn out. They are just educated guessing at the Republican turn out.

But the times they are a changing. The Democrat grass roots efforts have always  depended on someone paying for the turn out labor. And for the most of the 20th Century Democratic grass roots get out the vote effort was provided by Union Labor.

 But as labor unions have decreased in size and influence the number of shop stewards used to get out the vote has greatly decreased. And the government worker unions are not nearly as astute at getting out the vote.

The 2004 election may have been the high water mark for Democrats  in mid-western industrial states. There seems to me to bee a good chance that the 21st century will portray rapidly growing Republican trend lines, while the Democratic trend line may become erratic.

Bush won Ohio in 2004 due largely to the Republican volunteers registering 80,000 new voters. Such things as the last 72 hour push, and well manned phone banks did not exist years ago.   If a good grass roots effort is made this year, Republicans can continue their winning ways.

The days  when 43 million Republican voters in 1980  rose to 54 million in 1984 only to fall to 48 million in 1988 and fall to 39 million in 1992 are over.

Democrats have a finite base growing at consistent rate while Republicans have been a wavy line all over the map. Those who fail to consider what Karl Rove has wrought do not understand that he has given Republicans a ground game.

Lack of  a ground game explains the wavy line of the Republicans voter turn out. The pollsters may give Rove lip service, but they still do not understand what he has wrought. On the other hand pollsters may not have clearly seen the reason for the Democrat trend line and perhaps are unaware of the chance it may soon collapse.

We need to welcome them to the 21st Century.