Truth and meme-ing

Memes are clearly the most popular way to spread political propaganda, which makes sense, since they’re designed specifically for sharing on social media. Memes were one of the tools the Russian government used to steer the 2016 US presidential election, and they’re still beloved by activists on the Right and Left alike. They make short, punchy, shareable arguments about an issue without the person sharing having to back it with evidence or reason—perfect for the social media age, in other words. But they also seldom tell the truth, even when they’re factually accurate. Yesterday I saw a perfect example of that.

When I saw the meme above yesterday, I saw it for what it was—propaganda—and also knew that that it was factually true. However, something about it bothered me, though at the time I couldn’t quite put my finger on why. It turns out, there’s a LOT that’s wrong with it.

The meme is made up of edited screenshots of two different Tweets: The top part of the meme is taken from a Tweet by the USA’s National Public Radio (stylised as “npr”), which refers people to an article on their site, “Democrats Consider: Is A White, Straight Man The Safe Bet Against Trump?”. The bottom part of the meme is taken from a Tweet by Oliver Willis, who describes himself as “Senior Writer at Shareblue”. Both Tweets are genuine, and the editing of the pictures didn’t alter their points.

The problem with the meme’s point—Willis’ Tweet—is that it’s a “gee whiz statistic”, something that sounds very profound and important—until you think about it and realise it says absolutely nothing, even though it’s factually true: Obama’s popular vote in 2008 was 69,498,516 (and it remains the highest popular vote that any presidential candidate has ever received). In second place was Obama in 2012 with 65,915,795, and then Hillary Clinton in 2016 with 65,853,514. However, in 2008 John McCain received just under 60 million votes, in 2012, Mitt Romney got just under 61 million, and in 2016 the Republican got just under 63 million. In full context of the elections, the 65 million plus votes the Democrats got is less astounding than it sounds, apart from 2008, which was an historic year with record levels of interest and engagement.

Also, the fact that the Democratic candidates have each won more than 65 million votes in each of the past three presidential elections isn’t the least bit surprising: The vote has been increasing as the population increases. It also doesn’t tell the whole story: The number of votes going to the Republican candidates has also been increasing, while Democrats’ vote totals have been decreasing.

Third, and most important of all, the popular vote is irrelevant: Americans don’t elect their president, they elect the people who elect their president through the Electoral College. And that is why this meme is so silly: While the point it makes is factually true, it’s also completely irrelevant.

To win a presidential election, a candidate need not receive the most popular votes—the current occupant of the White House didn’t, after all. Instead, they just need to win enough votes in the right states. Those states are called “swing states” because they can swing back and forth from supporting the Republican or Democratic candidate. They matter because both parties have states that are firmly supporting them, usually putting them beyond the reach of the other party, but neither candidate can win with their strong states alone—they need to win some of the swing states.

So, the only states that actually matter in a presidential election are the ones that up for grabs to some extent or other: Colorado, Florida, Iowa, Michigan, Minnesota, Ohio, Nevada, New Hampshire, North Carolina, Pennsylvania, Virginia, and Wisconsin. Those dozen or so states are the ones the candidates visit, and where they spend their money. The other 38 states just sit back and watch the show.

Among those swing states are six that are often crucial in presidential elections: Florida, Michigan, Ohio, North Carolina, Pennsylvania, and Wisconsin. They were certainly important in all the presidential elections so far in the 21st Century.

Here are the popular vote percentages in those six states for the past three presidential elections:

Apart from the historic year of 2008, the state percentages track reasonably close to the nationwide percentages, though in all three elections both candidates outperformed the average in some states, underperformed in others.

However, when we look at the popular vote in those same states in 2016, something different emerges:

What this chart shows is that a shift of 77,774 votes in three states, with a combined total of 48 Electoral College votes, would have elected Hillary Clinton US President. This is because after the election was done, the Republican won office with 306 Electoral College votes to Hillary Clinton’s 232 (due to Faithless Electors, the final totals were reduced to 304 and 227, respectively). That means that if those three states alone had gone for Clinton, on Election Night she would have been elected president with 280 Electoral College votes. And that is why Swing States matter so much.

There are plenty of reasons why the 2016 election was so fraught, not the least that both the Republican and Democratic nominees were unpopular. The evidence for this isn’t voter turnout, which was pretty similar in both 2016 and 2012: 55.7% in 2016 and 54.9% in 2012 (both were significantly lower than the 58.2% turnout in 2008). Instead, the evidence comes from the performance of the two main minor party candidates, the Libertarians’ Gary Johnson and the Greens’ Jill Stein.

The votes for both Johnson and Stein skyrocketed in 2016 as compared to 2012: In 2016, Johnson’s nationwide vote total was up three and a half times what it was in 2012 (4,489,341 in 2016 v. 1,275,971 in 2012), and Stein’s was 3.1 times higher than in 2012 (1,457,218 v. 469,627 in 2012). While some centrists may have voted either way, it’s fair to assume that most of Johnson’s increase came from Republicans who couldn’t stomach their Party’s nominee. Similarly, Stein’s increase came from those who could not vote for either the Libertarian or Democratic nominees, but not all of them were necessarily Democrats: Her totals also included disaffected supporters of Sen. Bernie Sanders.

This defection to minor parties is, ultimately, what cost the election for Clinton. Look again at the chart for the popular vote in 2016, and in two of those three states with Clinton’s smallest margin of loss, Clinton could have made up the deficit with votes from Stein alone, and Stein still would have performed better than in 2012. The one state where that wouldn’t have been true is Pennsylvania, where Stein’s votes could have erased Hillary’s deficit, however, the leftover would have been less than Stein received in 2012.

Here’s what all this tell us. First, Clinton ran a terrible campaign. Had she not been in such a week position going into the election, those three states wouldn’t have been lost, nor the election. To be sure, part of the reason she had so much trouble was because of eight years of intense efforts by the Republican Party to destroy her through persecution and inquisitions, and because the Russian government worked with the Republican campaign to ensure Clinton lost. But those factors alone were not enough, and focusing on them lets her campaign off the hook for having made some colossally stupid decisions in the final weeks, giving the Republicans nearly free reign in those three states. The result is history.

Now, look again at the percentages from 2008-16: We see that it’d steadily become harder for a Democrat to win. We know that part of the reason for that is that far too many Democratic voters simply stayed home in 2012 and 2016, so getting them to vote is crucial in 2020. But what else is hidden in plain view?

North Carolina is probably lost, unless the candidate has overwhelming support like Obama in 2008—and even then, an absolutely terrible candidate like the Republicans had in 2016 still did better than Obama did in 2008. Sure, it’s still possible for the Democrat to win the state, given the right candidate and the right issues, but resources may be better spent elsewhere.

Beyond that, the figures do NOT indicate what kind of candidate would be best. It’s easy to see why some Democrats want a centrist nominee because they’ll need to win the swing states, and, considering the Republican won them all in 2016 despite having such a terrible person as their candidate, it’s not crazy to conclude a moderate is more electable in those states, and so, to the presidency.

On the other hand, Democrats on the Left think that a more Left-leaning candidate would be better in order to provide a clearer alternative to the Republican and his regime. Their logic is supported by the enormous energy that propelled Obama to victory in 2008, something the Left believes could happen again, but only if the nominee is Left-leaning.

There are problems with the logic of both.

Overall, Conservatives outnumber Liberals 35% to 26%, according to a Gallup report back in January. At the same time, the gap between Conservatives and Liberals has narrowed from 19 points in 1992 to 9 points today. So, both sides are right.

More relevant to why, precisely, so many Democrats say they want a Left-leaning nominee is that now, for the first time, a bare majority—51%—of Democrats identify as Liberal. Moreover, the percentage of Democrats that identify as Liberal has been rising steadily since 1994, and faster in recent years. Republicans, meanwhile, have steadily become more conservative since 1994, which surprises no one: Nearly three-quarters now say they’re Conservative.

Buried in all this, and in the meme that began this discussion, is the belief that the Democratic nominee “must” be a woman or minority to win election, but that’s just an ideological position. Take women voters, for example: They’re pretty evenly distributed between Liberals, Moderates, and Conservatives. Black people tend to be Moderate (41%), then Liberal (31%), and then Conservative (22%). Which means that both women voters as a group and black voters as a group are more conservative than Democrats, but more Liberal than Republicans, and it's more than a little insulting to suggest that they'll put their own political leanings aside to vote someone just because the candidate is "like" them.

Based on ideology alone, and all other things being equal, it appears that a Centrist Democrat could have broader appeal than a Liberal—in the General Election. A Left-leaning candidate could do better in the primaries, where voters are more Liberal than the general population, but if they can’t credibly tilt back to the Centre after nomination, they could face a hard time in the General Election.

All of which means that the meme was factually correct, but it’s irrelevant because that’s not how presidents are elected. It’s real point, that a woman or minority can do well is also true, but ignores the fact that neither Obama nor Clinton ran as actual Liberals, but, rather, more of somewhat Left-of-Centre candidates, and the ideological make-up of the general electorate suggests that this mattered.

Have said all of that, there are any number of things could completely change the current dynamic. One candidate could pull well ahead of the others, some could start dropping out, or a new candidate could enter the race and electrify everyone. Anything’s possible, as 2016 proved.

Unfortunately, memes can’t adequately express variables, uncertainties, or nuance. It’s best to give meme’s a miss and stick with all the facts, messy as they can be, and not rely on some cartoonish reduction of them.

Additional references:

"List of United States presidential elections by popular vote margin" – Wikipedia
"2008 United States presidential election" – Wikipedia
"2012 United States presidential election" – Wikipedia
"2016 United States presidential election" – Wikipedia