Tuesday, November 07, 2006

Updates: early numbers

This year, as it turns out, there are few leaks of exit poll data, so far. So the best I can do is give you some raw numbers, without much context or interpretation:

In the hottest Senate races, here is what is happening, according to exit poll data that is emerging here and there in the media:


SENATE EXIT POLLS AS OF 5:30 EST:
The Democrats are leading in 8 key races:
Virginia (D52-R47)
Rhode Island (D53-R46)
Pennsylvania (D57-R42)
Ohio (D57-R43)
New Jersey (D52-R45)
Montana (D53-R46)
Missouri (D50-R48)
Maryland (D53-R46)…

The Republican candidates are leading in 2 races:
Tennessee (R51-D48)
Arizona (R50-D46)

All of this data needs to be read with a great deal of caution, based on the past few election cycles. That's because Democrats in general seem to poll higher in exit polls than they do once the final votes are counted (*).

Nevertheless, the media consortium that owns the exit polling system has tried to tighten up its weaknesses, supposedly. Therefore, these may be more predictive this time around. (Ask me around midnight PST.)

With all caveats on the table, I will say these early exit poll trends suggest the Democrats may have a better chance at winning control of the Senate than I predicted earlier today.

I'll try to post Housse data oncee I can gather it and evaluate it.

One other note: There appear to be delays, and possibly polling problems in many states again this year, including both Florida and Ohio.

Deja Vu all over again?

***

Here, BTW, is what the Republican Party suggests about interpreting exit poll data:


BEWARE OF EXIT POLLS

Biased And Inaccurate Predictions Have Led To Poor
GOP Exit Poll Showings In Past Three National Elections
_______________________________________


FAST FACTS ON EXIT POLLING

* Election Experts Believe Exit Polls Give An Edge And Sway Towards Democrat Candidates.

* National Exit Polls Will Skew In Favor Of Democrats This Year, Due To Large Numbers Of Uncontested Democrat Seats In The House Of Representatives.

* Early Exit Polling Returns In 2004 Were Widely Inaccurate, Declaring Sen. John Kerry (D-MA) To Be The Next President Of The United States And Republicans Barely Holding A One Seat Majority In The U.S. Senate.

* In The 2002 Midterm Elections, Exit Polling Produced Unusable Data.

* In 2000, Exit Polling Malfunctioned And Incorrectly Projecting Vice President Al Gore As The Winner Of The Crucial Battleground State Of Florida.

EXIT POLLS TRADITIONALLY HAVE A DEMOCRAT BIAS

Exit Polls Give Democrats An Edge:

Exit Polling Always Tends To "Give An Edge To Democratic Candidates." President of Mitofsky International, Warren Mitofsky: "Mitofsky said exit polls have always tended to give an edge to Democratic candidates ..." (John Cook, "Early Exit Polls Overstated Kerry Results, Media Group Says," Chicago Tribune, 1/20/05)

* "[Mitofsky] Said That For Reasons That Remain Unclear, Democratic Voters Are More Likely Than Republicans To Agree To Interview Requests From Pollsters." (John Cook, "Early Exit Polls Overstated Kerry Results, Media Group Says," Chicago Tribune, 1/20/05)

Democrats More Likely Than Republicans To Respond To Exit Polls:

October 2006 Fox News/Opinion Dynamics Poll Found Democrats Were More Likely To Respond To Exit Polls Than Republicans.

* 72% Of Democrats Responded They Were Very Or Somewhat Likely To Fill Out Questionnaire, Compared To 66% Of Republicans. (Fox News/Opinion Dynamics Poll, 10/26/06)

* 44% Of Democrats Claimed They Were Very Likely To Fill Out Exit Poll Survey, Compared To 35% Of Republicans. (Fox News/Opinion Dynamics Poll, 10/26/06)

* Democrats (37%-10%) And Republicans (25%-18%) Agreed That Democrats Are More Likely To Share How They Voted With A Pollster They Do Not Know. (Fox News/Opinion Dynamics Poll, 10/26/06)

In 2006, Exit Polls Skew In Favor Of Democrats, Due To Large Numbers Of Uncontested Seats In The House:

In 2006, There Are Over 40 Uncontested Democrat Seats And 10 Uncontested GOP Seats, Which Will Overstate National Democrat House Vote In Exit Polling. (National Journal Website, www.nationaljournal.com, Accessed 11/6/06)

* In 2004, There Were 39 Uncontested Republican Seats, Compared To 30 Uncontested Democrat Seats. (National Journal Website, www.nationaljournal.com, Accessed 11/6/06)

* In 2002, There Were 45 Uncontested Republican Seats, Compared To 36 Uncontested Democrat Seats. (National Journal Website, www.nationaljournal.com, Accessed 11/6/06)

IN 2004, EXIT POLLING PREDICTED OVERWHELMING DEMOCRAT VICTORIES

In 2004, Exit Polls "Got All Of The Bush States Wrong" - Predictions Were "Most Skewed" Since Exit Polling Began:

Exit Polls Showed Across-The-Board Failure. "[T]he networks did get the exit polls wrong. Not just one of them. They got all of the Bush states wrong." (Dick Morris, "Those Faulty Exit Polls Were Sabotage," The Hill, 11/4/04)

* "It Was Dej? Vu All Over Again With Major Embarrassment For Exit Pollsters Since It Was Clear That They Way Underestimated Bush's Support In States Like Virginia That The President Won Handily Once The Real Vote Came In." (Deborah Orin et. al., "Voters Choose Double Dubya," New York Post, 11/3/04)

USA Today: "In fact, the 2004 numbers were the most skewed since joint exit polling began in the 1980s." (Editorial, "Exit Polls' Cloud Crystal Ball," USA Today, 1/20/05)

* The [Cleveland, OH] Plain Dealer: "Add in poor weather, data programming errors and other technical glitches, and the end product, calculated to give major news operations an inside glimmer as to which way the vote was going, instead produced the most inaccurate information in the past five presidential elections." (Editorial, "The Pollsters Were Pole-Axed," The [Cleveland, OH] Plain Dealer, 1/23/05)

* San Francisco Chronicle: "If exit polls have a role at all, they need to be staged carefully and handled with caution. Along with improved methods, the report [by firms who came up with flawed 2004 exit polls] suggested that no numbers be released to news organizations until near the end of Election Day. That's the least poll-takers can do." (Editorial, "Exit-Poll Errors," San Francisco Chronicle, 1/21/05)

In 2004, National Election Pool (NEP) Incorrectly Projected Victories For Sen. John Kerry (D-MA):

"ABC, AP, CBS, CNN, Fox, And NBC ... Created The National Election Pool To Provide Tabulated Vote Counts And Exit Poll Surveys ... These Six Major News Organization, In A Joint Decision ... Appointed Edison Media Research And Mitofsky International As The Sole Provider Of Exit Polls ..." (Exit-Poll.net Website, http://www.exit-poll.net/, Accessed 10/24/06)

"In The 32 States With Exit Poll Estimates For Both A Presidential Race And A Senate Race The Average Error On The Difference Between The Top Two Candidates Was 5.0 Points In The Democratic Direction For President And 3.6 Points In The Democratic Direction For Senate." (Edison Media Research And Mitofsky International, "Evaluation Of Edison/Mitofsky Election System 2004 For The National Election Pool," 1/19/05, p. 20)

* Iowa: NEP Projected Sen. Kerry Winning By 1% - President Bush Carried Iowa By .7%;
* Nevada: NEP Projected Sen. Kerry Winning By 1.4% - President Bush Carried Nevada By 2.6%;
* New Mexico: NEP Projected Sen. Kerry Winning By 4.2%- President Bush Carried New Mexico By .8%;
* Ohio: NEP Projected Sen. Kerry Winning By 6.5% - President Bush Carried By 2.1%;
* Virginia: NEP Projected Sen. Kerry Winning By 0.5% - President Bush Carried Virginia By 8.2%. (Edison Media Research And Mitofsky International, "Evaluation Of Edison/Mitofsky Election System 2004 For The National Election Pool," 1/19/05, p. 21-22)

NEP's Exit Poll Projections Underestimated President Bush's Support In Several Key States:

* Minnesota: NEP Projected Sen. Kerry Winning By 14.3% -- Sen. Kerry Won By 3.5%;
* New Hampshire: NEP Projected Sen. Kerry Winning By 15% -- Sen. Kerry Won By 1.4%;
* North Carolina: NEP Projected President Bush Winning By 3.6% -- President Bush Won By 12.4%;
* Pennsylvania: NEP Projected Sen. Kerry Winning By 13.8% -- Sen. Kerry Won By 2.3%;
* Wisconsin: NEP Projected Sen. Kerry Winning By 5.7% -- Sen. Kerry Won By .4%. (Edison Media Research And Mitofsky International, "Evaluation Of Edison/Mitofsky Election System 2004 For The National Election Pool," 1/19/05, p. 21-22)

NEP's Exit Poll Projections Had Republicans Winning Only 51 Senate Seats Instead Of 55 Seats They Hold Now:

* Alaska: NEP Projected Former Gov. Tony Knowles (D-AK) Senate Victory; Sen. Lisa Murkowski (R-AK) Won;
* Florida: NEP Projected Dem Betty Castor (D-FL) Senate Victory; Sen. Mel Martinez (R-FL) Won;
* Kentucky: NEP Projected Dem Dan Mongiardo (D-KY) Senate Victory; Sen. Jim Bunning (R-KY) Won;
* North Carolina: NEP Projected Dem Erskine Bowles (D-NC) Senate Victory; Sen. Richard Burr (R-NC) Won. (Edison Media Research And Mitofsky International, "Evaluation Of Edison/Mitofsky Election System 2004 For The National Election Pool," 1/19/05, p. 23)

Political Professionals Agree 2004 Highlighted Exit Polling's Failures:

Zogby International's John Zogby: "I'm not sure that I will ever believe an exit poll again ... How could they have been so way off? They were worse than virtually every pre-election poll." (John Cook, "Early Exit Polls Overstated Kerry Results, Media Group Says," Chicago Tribune, 1/20/05)

* Zogby: "The sum total of what we got today is enough to suggest that there should never be exit polls again." (Donald Lambro, "Polling Firms Blame Youth, Leaks For Errors," The Washington Times, 1/20/05)

The Washington Post's Director Of Polling Richard Morin: "[T]he 2004 election may have finally stripped exit polling of its reputation as the crown jewel of political surveys, somehow immune from the myriad problems that affect telephone polls and other types of public opinion surveys." (Richard Morin, Op-Ed, "Surveying The Damage," The Washington Post, 11/21/04)

* Morin: "Instead, this face-to-face, catch-the-voters-on-the-way-out poll has been revealed for what it is: just another poll, with all the problems and imperfections endemic to the craft." (Richard Morin, Op-Ed, "Surveying The Damage," The Washington Post, 11/21/04)

President Of Mason-Dixon Polling Firm Brad Corker: "Exit Polls Are Often Wrong". "[B]rad Coker, president of the Mason-Dixon polling firm that called Bush's 2.5-percentage-point win in Ohio practically right on the nose for The Plain Dealer, says ... exit polls are often wrong." (Ted Diadiun, "Rest Assured, We Checked Out Election 2004 Thoroughly," The [Cleveland] Plain Dealer, 6/18/06)

"Herb Asher, An Ohio State University Political Scientist, Said Election Results Don't Necessarily Reflect Exit Polls." (John Nolan, "Forty Voters Want November Election Results Thrown Out," The Associated Press, 12/14/04)

* Asher: "We all know that exit polls can be wrong. Exit polls are basically a sample." (John Nolan, "Forty Voters Want November Election Results Thrown Out," The Associated Press, 12/14/04)

CNN's Bill Schneider: "The lesson here is put not your faith in exit polls ... particularly if the exit poll is close ... Exit polls are designed for analysis. ... They are not very good ..." (William Douglas, "Turnout: The Early Exit Polls Mostly Wrong," Detroit Free Press, 11/4/04)

Then-CNN Anchor Judy Woodruff: "People want to jump on (exit polls) because they are the first little sliver, little shred of evidence ... It's dangerous to seize on those numbers and assume anything - and yet that's what happened." (Michelle Mittelstadt, "Exit Poll Group Assailed For Erroneous Early Results," The Dallas Morning News, 11/4/04)

Pollster Andrew Kohut: "[D]oing [exit polls] on the fly has led us astray." (Michelle Mittelstadt, "Exit Poll Group Assailed For Erroneous Early Results," The Dallas Morning News, 11/4/04)

EXIT POLLS WERE ALSO WIDELY INACCURATE IN 2002 AND 2000 ELECTIONS

FLASHBACK FACT: In 2002 Midterms, Exit Polls Were Scrapped Due To Inaccuracy:

VNS Consortium Scrapped 2002 National Exit Polls Because It Could Not Guarantee Accuracy. "Voter News Service abandoned its state and national exit poll plans for Election Night, saying it could not guarantee the accuracy of the analysis which media organizations use to help explain why people voted as they did." ("VNS Abandons National Exit Poll Operation, A Setback For Revamped Elections System," The Associated Press, 11/6/02)

* "The Exit Poll Failure Was A Major Setback For VNS - A Consortium Consisting Of ABC, CBS, NBC, CNN, Fox And The AP. VNS Had Completely Rebuilt Its System In Response To The 2000 Election, When Television Networks Twice Used Its Information To Make Wrong Calls In The Decisive Florida Vote For The Presidential Election." ("VNS Abandons National Exit Poll Operation, A Setback For Revamped Elections System," The Associated Press, 11/6/02)

* Zogby International's John Zogby: "The early [2002] exit-poll data was awful. ... And it came with the caveat that it was unreliable. Can you imagine a doctor saying that about a diagnosis? What a mess!" (Beth Gillin, "Media Organizations Discuss What Went Wrong With Exit-Poll Service," The Philadelphia Inquirer, 11/7/02)

FLASHBACK FACT: In 2000 Presidential Campaign, Exit Polls Were Badly Flawed In Critical State Of Florida:

In 2000, Exit Polls Were Badly Flawed In Critical State Of Florida. "[S]ome bad assumptions led Voter News Service, the television network exit poll consortium, to make and then retract two dramatic election-night predictions on the winner of the presidential race in Florida." (Richard Morin and Claudia Deane, "Why The Fla. Exit Polls Were Wrong," The Washington Post, 11/8/00)

* Networks Blame Incorrect Projections On Erroneous Exit Polling. "[The networks] concluded the problems were largely due to bad information supplied by Voter News Service, an exit poll consortium run by television networks and Associated Press." (Elizabeth Jensen and Megan Garvey, "TV Election Gaffes Called Statistical, Not Political," Los Angeles Times, 2/9/01)

o CBS' Director Of Surveys Kathy Frankovic: "In the Tampa area, the exit poll results from the [sample] precincts turned out to be more Democratic than the vote turned out to be." (Richard Morin and Claudia Deane, "Why The Fla. Exit Polls Were Wrong," The Washington Post, 11/8/00)

Experts Lament Exit Polling's "Poor" Predictions In 2000. "Paul Biemer, a statistician hired to review the VNS methodology, reported that 49 percent of those surveyed leaving the polls declined to respond. In the parlance of his trade, Mr. Biemer said 'a non-response rate of this magnitude is a potential source of statistical bias in the model projections ...' In other words, exit polls are poor predictors." (Jules Witcover, Op-Ed, "Voters Don't Need Instant TV Results," The [Baltimore, MD] Sun, 2/16/01)

Faulty Media Reports In 2000 Impacted Florida Panhandle And Congressional Races:

"About 7:50 P.M. EST Tuesday, Less Than An Hour After Most Of The Polls Closed In Florida, Both CBS And NBC Declared Mr. Gore The Winner In That State, A Decision Based Largely Upon Exit Polls Conducted By Voter News Service. Fox News, CNN And ABC Followed Within Minutes." (John Godfrey, "Tauzin Plans Hearings On Gaffe," The Washington Times, 11/10/00)

* "A Survey Conducted By John McLaughlin & Associates Found That The Early And Incorrect News Network Announcements Declaring Al Gore The Winner In Florida May Have Influenced Thousands Of Last-Minute Voters In The Central Time Zone Section Of The Florida Panhandle Not To Vote. The Premature Announcement Discouraged Many Registered Voters Who, According To [The] Survey's Results, Would Have Voted Like The Rest Of Their Neighbors - Overwhelmingly For George W. Bush." (McLaughlin & Associates Website, "Panhandle Poll Summary," www.mclaughlinonline.com, Press Release, 11/20/00)







A Product Of The RNC Research Department


--More TK --

2 comments:

Robert W. said...

David,

Sorry for the late posting but I went to Technorati.com, searched for "telephone polling inaccuracies" and found your posting.

I got in a big argument today with a longtime professional pollster who uses telephone polling to get most of her results. Hasn't telephone polling been proven to be wildly inaccurate?

Robert W.

David Weir said...

As you can see, I added every caveat in the book when I released these numbers. I'm not a professional pollster but a journalist, and I have learned over many election cycles that these early numbers, though fascinating to insiders, can be misleading to the public. What was going on in the overall sense in this particular election was the possibility that Democrats would win control of Congress. What I attempted to do was collect, distribute, and interpret numbers from many sources. I think I posted eight times or so. Although my readership was small, I feel I gave anyone who checked a pretty accurate view of the outcome as it emerged. So, I agree that telephone polls and exit polls are mot entirely reliable indicators of outcomes. However, when mapped with demograophic data and other databases, it is possible for a skeptical journalist to provide a resaonable roadmap on election night. Call it triangulation.