Monday, December 18, 2017

There's no "real" reason that Roy Moore lost

On Tuesday Doug Jones defeated Roy Moore to became the first Democratic Senator from Alabama in a quarter of a century, and political pundits are eager to give their two cents on how he pulled it off.

Some of them claim that Jones’s victory was due to the high turnout of Black voters, and Black women in particular

Others say it was due to those GOP voters who refused to back an accused child molester who had twice been kicked off the state supreme court

While others think that the credit should go to Doug Jones himself

When something momentous happens in politics there’s nothing the media loves more than fighting about what “the real reason” for it was. But these debates are always kind of dumb, because in reality political events never happen for just one reason.

In a close election, everything matters

Winning a close election  - like the one we just had in Alabama  - is like drawing a royal flush in poker. It would be dumb to say that “drawing the queen of diamonds” was the real reason you got a royal flush. You could say the exact same thing about the ten, jack, king, or ace of diamonds. You need all of them together to make the hand. None of them alone could have done it without all of the others.

Let’s set the stage - Doug Jones won the Alabama senate race because around 670,000 Alabamians voted for him, compared to only 650,000 who voted for Roy Moore - a difference of around 20,000 votes. In comparison, when Alabama’s other Senator, Republican Richard Shelby, ran for reelection in 2016, he won by over 600,000 votes.

So did Jones win “because” of Black women? Definitely! Here’s a chart of exit polls from the Washington post showing that 97% of Black women voted for Doug Jones.

From the Washington Post

Since Black women made up 18% of all voters, and around 1,344,000 people cast a ballot in the election, some simple math1 tells us that around 230,000 black women voted for Jones. If just 20,000 of those Black women had stayed home - or voted for Moore - Jones would have lost.

But obviously Black women - or Blacks in general - couldn’t have done it alone. There were only around 390,000 Black voters total in this election, and Jones needed over 670,000 votes to win. In order for Jones to win he needed at least some whites to vote for him as well.

Let’s look instead a totally different group: White born-again Christians.

From the Washington Post

Only 18% of this group supported Jones, but since almost 600,000 white born-again Christians cast a ballot in the race (44% of all voters),  that still translates to over 100,000 votes. If just 10,000 white-born again Christians  - just 1% of all white evangelicals who cast a ballot - had switched their vote from Jones to Moore - Jones would have lost.2 So even though most evangelicals didn't vote for Jones, the votes of those who did were absolutely critical to victory.

Of course - we could do make a similar argument that Jones’ victory was “because” of female Republicans, White Democrats, 18-29 year olds, or whoever. If the percentage of any of these groups who voted for Jones was just slightly lower, or just a few thousand more of any of them had stayed home, Roy Moore would be the Senator of Alabama.

Different voters, different stories

Now, it’s true that the levels of support Jones got from these different groups  - both in terms of how people voted and who showed up to vote in the first place - mean very different things.

For example, the sheer number of Blacks who showed up to vote for Jones is probably more impressive than the percentage - which probably would have been close to 90% not matter who the Republican candidate was. It looks like 30% more Blacks showed up to vote in this special election than did in the 2014 midterm election. This says something about the impact of turnout operations - even in the face of widespread voter suppression tactics.

On the other hand, I think that the fact that even 18% of white, born-again Christians in Alabama cast a vote for a pro-choice Democrat is pretty amazing. It is sad that this number isn't higher - considering all the awful things Moore had done (or is alleged to have done) but in reality, it's shocking that it's as high as it is. It's hard to know for sure, but it seems likely that if Jones had been running against a "normal" conservative Republican, the percentage of White evangelicals in Alabama who would have voted for a pro-choice candidate would be closer to zero. And remember that just a one percentage point change in this number would have cost Jones the election.

These are all important stories to explore - and it's good for us to argue about what each one means (maybe my interpretations are bogus!) But instead of doing that we often just argue about which one is the “real” reason that Moore lost and Jones won.

We do this again and again whenever anything momentous happens in politics.

What does it mean for something to happen "because" of something else?

When we argue about "the real reason" that something happened, we usually have a mental image in our head that looks something like this

There's this one thing that really caused that event to happen, and we need to figure out which thing that is. Some people think that other things are the real reason, but they're wrong.

But the real picture probably looks more like this:

Every important thing in politics - or life really - is simultaneously “caused” by a whole bunch of different things, which are themselves caused by a whole bunch of other things, which may even be caused by that first thing. It’s all a big mess.

Social scientists try to untangle the mess, bit by bit. We study each of those little arrows one by one, to try and see where they go, and how they interact with other arrows. Just because we’re looking at one arrow doesn’t mean we’re ignoring the others. But we might from time to time say “hey, there’s an arrow over here that people should pay more attention to.”

Alabama gets complicated

Let's go back to the Alabama Senate race. Doug Jones did about 30 points better than Democrats have typically done in Alabama - and he needed every one of those 30 points to win.

If we want to understand how this happened, then - in addition to looking at specific groups of voters, like Black women or White evangelicals - we should also look at broader national trends that had nothing to do with Roy Moore at all.

One of the "other things" that helped Jones is that this is a really good time to be a Democrat running for Congress!

Democrats are around 10 points ahead of Republicans in the "generic ballot" - a survey question which asks Americans whether they would rather vote for a "generic" Democrat or Republican in a Congressional election.

Democrats have also been doing really well in special elections this year. This might be a surprise to you - because the media tends to only care about who won these races and misses the story about how big the margin of victory was. For example, back in June, Republican Ralph Norman defeated Democrat Archie Parnell in a special election for South Carolina's 5th congressional district. Media reports focused on the fact that Parnell lost (and tried to find the "real" reason why!). But Parnell only lost by 3 points in a district that Republicans usually win by close to 20 points!

We have seen similar trends in most other special elections. Overall, in the six special elections this year before Alabama, the Democratic candidate did an average of 13 points better than we'd expect based on previous election results.

Although we can't be sure, this suggests that - despite everything that Roy Moore said and did - Jones would have done about 10 points worse (and thus lost) if the election had happened in a year when Democrats weren't enjoying such a huge boost in popularity.

Why are Democrats doing so well right now? Well, like the Alabama result itself, there isn't just one "real" reason, but a bunch of different reason that are all tied up with each other.

One thing that's going on is that Americans always become more liberal when a Republican is in the White House (although they also become more conservative when a Democrat is in the White House). That's one of the reasons that the opposition party (in this case, Democrats) tend to do really well in Midterm elections.

But right now the effect is probably even stronger than usual because of Trump's low approval ratings (which would be even lower if not for the healthy economy).

So why did Doug Jones win?

 Doug Jones won because of Black women. Doug Jones won because of white evangelicals. Doug Jones won because of young voters. Doug Jones won because of Roy Moore.  Doug Jones won because a Republican was in the White House. Doug Jones won because Donald Trump is really unpopular.

All of these statements are true - and they're all important in different ways. You could argue that we should be talking about some of the more than others, but it's important to recognize that none of them are the "real reason" that Jones won - because reality doesn't work like that.

The "real story of 2016"

Pundits write these "real reason" stories all the time. There was a deluge of them after after the 2016 election. Everywhere you looked there was someone telling you the real reason Clinton lost and Trump won. But just like Jones, Trump just barely won. If just 400,000 voters in a few swing states had changed their minds, Hillary Clinton would be President right now.

Many of the factors that these pundits point to were important. If American voters had been just a little bit less racist or sexist - Trump would have probably lost. If Clinton had been just a little bit more popular - Trump would have probably lost. If the American economy were doing just a little bit better, Trump would have probably lost. If James Comey hadn't released his infamous letter about Clinton's emails - and the media hasn't spent so much time reporting on it - Trump would have probably lost. Trump only won because all of these different cards came up just the way he needed to.

Just like Alabama, these different reasons mean different things. Some things - like how the media covers people like Trump, - are easier to change than others - like the US economy, or structural racism.  But they're also complexly intertwined with each other. How much of Clinton's unpopularity was driven by sexism? How much is white racism and hostility to immigrants driven by economic frustrations? These are tough questions that social scientists are scrambling to figure out.

In reality, there's rarely just one "because"

Trying to untangle the various factors that led to Trump's victory, Moore's loss, or any important political event -  is an important task.  Arguing about which is the "real" reason for these events is not.

If we want to get better at understanding politics, we need to develop a more sophisticated notion of what it means for something to happen "because" of something else. The better we get at that, the better we will be at understanding the craziness that is American politics in the early twenty-first century.


1 18% of 1,344,000 is 242,000. And 97% of 242,000 is 234,000. Note that these numbers are just estimates, since the percentages I'm using come from the Washington Post's exit polls, which have a pretty large "margin of error" around them. So I'm rounding to the nearest ten thousand.

2 A tricky thing about electoral math is that if a person switches their vote from candidate A to candidate B then that actually closes the gap by two votes. So if 10,000 evangelicals had voted for Moore instead of Jones, then there would have been 10,000 more vote for Moore, and 10,000 less for Jones - which would have eaten up the entire 20,000 vote gap. In contrast, if 10,000 evangelicals (or Black women, or whoever) had just stayed home instead of voting for Jones, then that would cost Jones 10,000 votes, but Moore would have the same number, so Jones would still win. In short - More would have won if any 10,000 voters had changed their vote, or if any 20,000 had stayed home - or any combination thereof.

No comments:

Post a Comment