ABC Mixes Polls and Makes Clinton Look Better

In the wake of the New York bombing, ABC and NBC both reported polling data Sept. 19 on how voters thought presidential candidates would handle terrorism. Both networks had the same data but showed different results.

If you watched ABC, Hillary Clinton seemed like the favored candidate:

ABC News

ABC News

But not necessarily if you watched NBC:

NBC News

NBC News

How did this happen?

Both of NBC’s figures came from a CNN poll conducted from Sept. 1 to Sept. 4:

CNN/ORC International Poll

CNN/ORC International Poll

ABC used the same information from the CNN poll to report on who voters thought would be better suited for commander in chief. But for the question of handling terrorism, the network used an ABC News/Washington Post poll released Sept. 11.

ABC News/Washington Post poll

So ABC showed that those polled preferred Clinton, both for handling terrorism and as commander in chief. ABC could not have used the Washington Post poll alone because it didn’t include a question about who would be a better commander in chief.

But ABC did not include the CNN poll results on nearly the same question about terrorism.

David Yepsen, director of the Paul Simon Public Policy Institute at Southern Illinois University, said both polls showed candidates were fairly evenly divided on the questions, so there wasn’t an issue or display of bias.

“The [margins of error] on all these numbers tells me this is pretty much tied in both polls,” said Yepsen, who worked as a political reporter at the Des Moines Register. “Neither candidate owns the electorate on the issue. I see no problem with using two different polls in a story.”

The broadcast showed the polls were taken from different sources, but the anchor did not say anything to that effect. If you look at the ABC graphics we showed above, the sources are difficult to spot.

For the viewers who saw both broadcasts, what are they to think?

Candy Crowley, former CNN chief political correspondent and anchor, said she saw no problem with ABC using its own poll of who voters could trust to handle terrorism.

“If they conduct them, news outlets want to use their own polls assuming they are still within a valid time frame (polls expire quickly),” Crowley wrote in an email. “Polls are very expensive as well as prestigious- two good reasons to air one’s own polls.”

On the other hand, Norman Ornstein, resident scholar at the American Enterprise Institute and contributing editor for The Atlantic, said the anchor should have pointed out the polls were different.

“Mixing and matching, especially with different dates, requires great care,” wrote Ornstein in an email. “I would say the anchor should have noted the difference, and noted that the CNN poll question replicated in the ABC poll had a somewhat different result, which might have resulted from different poll methodology or from different timing. Or ABC should not have used the CNN poll at all.”

More from the experts:

Candy Crowley:

• This doesn’t seem to me to rise to the level of an ethical problem unless ABC chose its polls with the intent of being deceptive about Hillary Clinton’s strength. I doubt that was the intent of whoever chose to highlight these numbers.
• If they do not have recent polling of their own on a particular subject or if their own poll is dated, news outlets want to/should use the most recent, reliable poll numbers available.   Not knowing the dates of the ABC poll, or the internal discussion, my guess is ABC used its own newish poll for the terrorism question . Then, lacking the CIC question in their own recent poll, they went to the most recent one they could find with roughly the question they wanted.
• If ABC took a poll asking the same terrorism question as CNN at around the same time (again, i don’t know the ABC poll dates), I see no reason ABC should air the CNN results on terrorism simply because it was airing the CIC numbers from CNN.
• I always like to err on the side of more information to the viewer, but I have to say, I wonder if verbally explaining the two different polls wouldn’t be more confusing.  Still, simply saying  “ABC’s most recent poll does not ask the commander in chief question, but A CNN survey conducted earlier this month found that ..blah blah blah”  …wouldn’t have hurt anything.  Particularly if in hindsight people will find it suspicious!  But again, networks like to ring the bell for their own polls, not others.

David Yepsen:

• I see no problem with using two different polls in a story.  Yes, I wonder why they did and the only guess would be is that they liked the wording on different questions. Again, the stories both make the point that the candidates are pretty evenly divided on these questions.
• ABC probably could have made the graphics clearer.  Still, they did disclose … ABC has a respected polling unit and isn’t known for screwing up.