Chris Cillizza

Lies, Damn Lies and Exit Polls

Some reporters live for numbers and calculate the horse race. Literally, that seems to be their job at newspapers. One is Chris Cillizza of the Washington Post. Part of his directive (meaning I blame not him but his editors) is to declare “winners and losers” and to provide interpretation on numbers.

This story is one where his interpretations are, for the most part, over-hyped. A more precise way of saying it is that he’s wrong and provides conclusions that cannot be supported, though to his credit, he warns that drawing conclusions based on a small sample set is dicey. Which makes one ask, then why do it? To wit, (Numerals track the issues Cillizza raised.)

  1. Cillizza says that Reagan and Romney won the white vote by 20 points. Then he breathlessly adds that Trump’s margin was a “record”: “On Tuesday, Trump one-upped them both — literally. He won the white vote 58 percent to 37 percent.” That’s 21%, well within the margin of error (4%), meaning that Trump may have gotten a few percentage points less than the R&R boys. Why couldn’t Cillizza say, Trump matched those previous margins?
  1. Then Cillizza points out that there wasn’t an increase in women voters.

Women made up 52 percent of the overall electorate in 2016 — down from 53 percent in 2012. And Hillary Clinton’s 12-point margin over Trump among women was pretty darn close to the 11-point win among women that Obama claimed over Romney four years ago.

So it was a “record” when Trump won maybe one percent more of white voters, but when Clinton won one percent more of the women’s vote, it was “darn close.” Again, margin of error for both surveys (I’m assuming 2012 had the same margin of error) was 4%.

  1. Cillizza tells us there was no surge of Latino votes. “In 2012, Hispanics made up 10 percent of the overall electorate. That bumped up, marginally, to 11 percent in 2016.” That one percent increase is now “marginal.”
  1. Later in the piece we learn that Trump outperformed Romney with evangelicals, 81% to 78%. Four percent margin of error means Romney could have had 82% and Trump 77%.
  1. Cillizza argues that Trump didn’t bring new voters to the table.

Just 10 percent of voters said that the 2016 election was their first time voting. Of that group, Clinton won 56 percent to 40 percent over Trump. Of course, new voters often overlap with younger voters who are eligible to vote for the first time; Clinton won among 18- to 24-year-olds by 21 points.

What we don’t know: How many new voters were first time eligible voters? Maybe only 2%. Of the other 8%, how many voted for Trump vs. Clinton? We don’t know, so this is metric is meaningless.

  1. Cillizza thinks “wanting change” trumped “right experience.”

Provided with four candidate qualities and asked which mattered most to their vote, almost 4 in 10 (39 percent) said a candidate who “can bring needed change.” (A candidate who “has the right experience” was the second most important character trait.)

What’s wrong with this? One thing is we don’t know is where he got the 39% figure. It is not in the poll data he cites, which is here. Secondly, he doesn’t tell us the figure for “right experience.” Was it 38% and therefore well within the margin of error?

  1. Perhaps this startles Cillizza most: Voters didn’t think Clinton won all three debates as clearly as Cillizza did. Count me among them.

Finally, I must remind you: These are exit polls. Which are not very reliable.

Some foolish journalists might write entire posts that assume that the black share of the electorate was 15 percent in Ohio. In reality, the exit polls just aren’t precise enough to justify making distinctions between an electorate that’s 15 percent black and, say, 13 percent black.

Let’s say, as the talking heads often do, we’ll just leave it there.