Projected New Senate:     50 Democrats     50 Republicans    

Polling data in Excel Battleground states Previous report Next report News Strong Dem (43) Weak Dem (5) Barely Dem (2) Exactly tied (0) Barely GOP (2) Weak GOP (2) Strong GOP (46) No Senate race   Map algorithm explained

Oct. 20	New polls:	AZ CT FL MA MD MI MN MO MT NJ NM NV NY OH PA RI TN TX VA WA WI
	Pickups:	Missouri Montana Ohio Pennsylvania Rhode Island

Welcome	General FAQ	Polling FAQ	Votemaster FAQ
Donations	All Senate races	Hot House races	Poll Graphs
More election data	Senate info	Political websites	Articles and books
2004 Election	EVP track record	Political humor	For bloggers
Animated map	Contribute a page

News from the Votemaster

Senate Polls

Zogby has a new batch of Internet polls today. In Maryland, Rep. Ben Cardin (D) is comfortably ahead 51% to 43%, as he has been all year, in contradiction to yesterday's SurveyUSA poll showing the two tied. This poll is much more plausible than yesterday's.

In Missouri, one of the four key Senate races, Zogby has incumbent Sen. Jim Talent (R-MO) ahead of state auditor Claire McCaskill (D) 50% to 47%, but this one keeps switching back and forth. It will probably be the closest Senate contest.

Surprisingly, Zogby has state senator Tom Kean, Jr. ahead of incumbent Sen. Bob Menendez (D-NJ) in New Jersey. Other polls have shown Menendez in front recently. New Jersey is the only state the Republicans might pick up.

Zogby has former Chattanooga mayor Bob Corker (R) ahead of Rep. Harold Ford, Jr. 49% to 42%. Other polls have shown Ford with a lead in Tennessee. Zogby has consistently shown Corker ahead.

Finally, Zogby also has incumbent Sen. George Allen (R-VA) slightly ahead of former Navy Secretary Jim Webb (D), 50% to 47%, in line with previous polls in Virginia.

House Polls

Over in the House, we have a couple of surprising results. In IN-07, Eric Dickerson (R) leads incumbent Rep. Julia Carson (D-IN) by 45% to 42%. While this is within the margin of error, In the other tight Indiana races (IN-02, IN-08, and IN-09), Democrats are likely to pick off Republican seats.

In KY-03, the race between incumbent Rep. Anne Northup (R-KY) and publisher John Yarmuth (D) has tightened, with Yarmuth now leading by an insignificant 48% to 47%.

In NY-26, Tom Reynolds (R-NY), one of the key players in the Foley drama (he is head of the NRCC and although he knew about Foley a year ago, he didn't want to rock the boat by ringing the bell too hard), is now ahead of Democrat Jim Davis again. However, the dynamics of this race depend so much on the Foley case and potential similar cases that anything could happen here.

But the weirdest result of all is in NY-20, where a Siena College poll puts John Sweeney (R-NY) ahead of challenger Kirsten Gillibrand (D) by 14 percentage points, 53% to 39%. Yesterday we had a poll from RT Strategies showing Gillibrand ahead 13%. That is a 27% swing in a day. I don't buy that. Something is radically wrong here.

As it turns out, I know NY-20, which covers the historic Hudson valley and a few additional counties. I have friends there and have been there a number of times. It is reasonably Republican, so RT Strategies poll showing Gillibrand way ahead seemed odd to me. RT Strategies is run by Thomas Riehle, a long-time Democratic pollster and V. Lance Tarrance, Jr., a long-time Republican pollster, so they balance each other when working together. I don't think RT Strategies is cooking the books. They do use robopolling, however, but so do Rasmussen and SurveyUSA. Now it is possible that when confronted with a computer calling them, some people don't like the "Press 1 for the Democrat or Press 2 for the Republican" routine and may hang up, but it is not obvious why Republicans should hang up more often than Democrats. Besides, Rasmussen was one of the most accurate pollsters in 2004, robo or no robo.

Maybe something is wrong with Siena College's poll. I heard from one reader that the college polls tend to be quite long. In addition to all the political questions, there are student-supplied questions and then the usual questions about age, party preference, occupation, size of your city, whether you own a home, income, etc. It is possible that some respondents get annoyed toward the end and hang up. The poor student caller is supposed to discard the data and call someone else. But after spending 20 minutes on the phone and then getting dropped, some students might just decide to answer the final questions themselves. Having the caller make up the answers is called "curbstoning" (a reference to when people went door-to-door to take polls and an interviewer might have "done" the polls sitting on the curb). It is one of polling's dirty little secrets. Although they rarely boast about it, the robopollsters have the advantage that the computer does not get frustrated when someone hangs up. It just picks another random number and calls someone else. The computer is prepared to call people until hell freezes over or it gets 600 valid interviews, whichever happens first. But I doubt that these differences in methodology can account for such a huge difference in NY-20. I think RT Strategies poll is wrong, but I don't know why.

Politics

Supposed you are the wife of a politician who has been caught in flagrante delicto. What do you do? In Hillary's case you make him sleep on the sofa for a year and then run for President. Carol Sherwood, the wife on Rep. Don Sherwood (R-PA), who choked his mistress of 5 years and thus achieved a bit of notoriety, wrote a letter accusing Sherwood's opponent, Democrat Chris Carney of negative campaigning. What did Carney do? He ran TV ads essentially saying that while Sherwood talks a lot about how important family values are to him, in real life he has a long-term mistress and then tries to choke her to death. So Mrs. Sherwood goes public and says anybody can make a mistake. Right. At least Bill was nice to Monica and even offered her a cigar.

Projected New House*:     227 Democrats     206 Republicans     2 Ties

WWW www.electoral-vote.com