43’s 44% in 2004 – Zombie Exit Polling

June 22, 2019

Now that we are entering another presidential election cycle, the unfortunate recycling of dubious narratives is also beginning anew. The latest example is the Lazarushian claim that George W. Bush received 44% of the Hispanic vote in the 2004 election.

Witness this week’s The Wall Street Journal print story (available online) by Alex Leary: Trump Launches New Effort to Lure Hispanic Voters. Contained within:

“Mr. Trump, who launched his 2016 campaign with fiery talk about illegal immigration, received 28% of the Hispanic vote in 2016, beating some projections and roughly achieving what Mitt Romney did in 2012.”

“Still, that is dismal compared with the 44% achieved by George W. Bush in 2004 and Mr. Trump’s base remains overwhelmingly reliant on white voters. But if he can boost his numbers, it could be decisive in states like Florida.”

It wouldn’t be bad to begin with an acknowledgement that exit polls are estimates, not facts. As such, each claim should be credited to the source so readers are fully informed. This would also be trust-enhancing for journalists – a worthwhile bonus, and one that can be easily accomplished here.

The 2016 figure credited to Trump by The Wall Street Journal comes from the Pew Research Center:

While Clinton under performed among Latinos compared with 2012, Republican Donald Trump won 28% of the Latino vote, a similar share to 2012, when Mitt Romney won 27%, and to 2008, when John McCain won 31%, according to exit polls. (It is important to note that the national exit poll is a survey with an overall margin of error of plus or minus 3 percentage points for the national result.)

Note the disclaimer that Pew’s experts make: the poll is a survey with a margin of error!

So is it safe to assume that The Wall Street Journal also used Pew’s estimates for George W. Bush’s returns in the 2004 election?

Nope. That figure is 40%, not 44%:

Pew Research Center wrote in 2005:

“An analysis of census and exit poll data suggests that President Bush took 40 percent of the Hispanic vote in 2004 rather than the 44 percent originally reported from the major news media exit poll.”

Steven Malanga of City Journal wrote of the post-election academic consensus in 2007:

“The media have also exaggerated Bush’s Hispanic support. Exit polls taken during 2004 showed a wide variation in Bush’s share of the Hispanic vote, ranging from a high of 44 percent in some polls to a low of 33 percent. Most in the media have gone with the higher number. Yet subsequent academic studies have discredited it and have since estimated Bush’s actual level of Hispanic support at somewhere between 35 percent and 37 percent.”

It’s an open secret that the media chose the highest number right from the jump and ran with it. More than a decade after the 44% estimate was abandoned, there are still reporters who are willing to run with the claim. Why?

Perhaps unsurprisingly, top Bush political strategist Karl Rove continued to cite the 44% exit poll – including in The Wall Street Journal‘s opinion pages.

Could a favorable view of the pro-immigration increases and pro-amnesty “comprehensive immigration reform” positions of 2004 President Bush factor into news divisions’ decisions to go with the higher (but discredited) number? You be the judge.

Again, the repudiation of that exit poll began almost immediately. Mike Tolson of the Houston Chronicle reported just days after the 2004 election:

“But a prominent Latino organization claims the numbers are as incredible as they appear.

“‘It ain’t true,’ said Antonio Gonzalez, president of the Willie C. Velazquez Institute, which researches Latino voting patterns. ‘Their poll showed more Latinos voting than there are registered Latino voters. That tells you everything you need to know.'”

So, advocates and experts seem to agree that the 44% number is not mathematically reliable. To this day, LatinoUSA and Latino Vote Matters both claim no more than 40%. The New York Times, Politico, and Vox employ Pew’s 40% figure. How is 44% still getting ink? It’s unclear whether there is a single exit pollster who would defend the veracity of its methodology. In fact, alternative (Ivy League!) estimates about the unauthorized population being higher than Pew says have more proponents* than the 44% exit poll number. It’s a safe bet you won’t be seeing their numbers in The Wall Street Journal, though – and certainly not presented as a fact.

Our goal is to analyze media – to see what’s being covered, how it’s being covered, and consider why it’s being covered the way it is. It’s in vogue in certain quarters to criticize the media, and far too often criticism is deserved. But our desire is strengthening media credibility, not weakening it. We discuss to persuade, because the only way public trust will flow to the media is through their buy-in of high, consistent standards, and their self-policing of those standards. The best watchdog is an encouraging one that roots for the success of the watched and hopes not to be needed.

Let’s hope Leary’s story was an outlier.

*Note: The immigration-reduction experts at Center for Immigration Studies do not find the methodology of the MIT-Yale study to be plausible.

Additional Reporting Including This Debunked Poll

Opinion Pieces Including This Debunked Poll

**Note: Every publication listed has been contacted about the issues with the exit poll and declined to change their coverage.

ANDREW GOOD is the Assistant Director of the Media Standards Program for NumbersUSA