Information Bubbles

Breaking form here as I wanted to discuss the primary.

I’m a Texan for those who don’t know me, with a mostly liberal friends circle. I noticed something downright bizarre in 2014 – several of my friends said Wendy Davis would be elected governor. This wasn’t the typical political bravado. They were claiming that not only would Davis beat Greg Abbott, but that it wouldn’t be close.

Wendy Davis lost the election by over twenty points.

What happened? Why the certainty of a victory? Was the polling wrong?

Nope. The belief that Wendy Davis would win wasn’t based on evidence. It came from an information bubble. (You’ll find the term more commonly called a filter bubble, but I don’t like the way that sounds).

Simply put, this is a bias all of us have. (I’m actually referring to the combination of a few biases here, but let’s keep it simple.) To put it simply, we tend to self separate into like groups. Your coworkers, your friends, and your family are going to tend to share groupings with you – race, income, values. Obviously there is variation there, but it holds in general. This affects the sorts of information sources you have and technology has exacerbated the problem. The algorithms know what sort of information you want to see and therefore show you that news while shunting other information.

In short, we end up in information silos – surrounded by things we agree with and lacking contact with things we don’t. This happens to ALL of us, but I’m directing this piece to the reaction to Super Tuesday I’ve seen from Sanders supporters.

I saw several things being said. “Biden can’t win because he won’t turn out numbers at the polls.” “Sanders is going to energize young voters and is therefore the better candidate.” “Sanders is better at coalition building.” Like those who thought Wendy Davis was going to walk into the Texas governor’s mansion, none of these ideas are based in the data. Let’s take a brief look at how people voted last week.

There’s a focus on who “wins” each state but as the Democrats don’t have a winner take all system, “winning” a state doesn’t mean much if it’s a narrow victory. If you narrowly win a large state but your opponent has blowout victories in small states, they’re going to pull ahead of you.

Sanders was ahead in four of the fourteen super Tuesday states. Some of these are still counting right now, but as of now, he’s ahead in California by 8%, Colorado by 12%, Utah by 16%, and Vermont by 29%. Vermont only has 16 delegates. 29 in Utah, Colorado with 67, and the crown jewel of Super Tuesday, California, has 415 delegates. This means Sanders picked up moderately more delegates in California and Colorado compared to Biden.

But that isn’t the entire story in those states. In 2016, Sanders had 2.4 million primary votes in California. With 88% of the vote in, he’s at 1.2 million votes this year. (It’s difficult to compare Colorado and Utah numbers as they had a caucus in 2016 and “normal” voting in 2020.) Sanders is still likely to win California, but not by a large margin and with fewer votes than he had 4 years ago. The energized turnout the Sanders campaign touts as a reason to elect him did not happen.

Now let’s look at Biden. Biden is currently leading in 10 of the 14 Super Tuesday states, but more importantly, he’s leading most of those by WIDE margins. He’s leading Alabama (52 delegates) by 47%, Arkansas (31 delegates) by 18%, Maine (24 delegates) by 2%, Massachusetts (91 delegates) by 7%, Minnesota (75 delegates) by 9%, North Carolina (110 delegates) by 19%, Oklahoma (37 delegates) by 14%, Tennessee (64 delegates) by 17%, Texas (228 delegates) by 5%, and Virginia (99 delegates) by 30%.

Winning those medium sized states by large margins is worth more than winning California by a smaller margin. Despite California’s size, Biden had a net delegate increase in Virginia which offsets Sander’s net delegate increase in California.

But let’s look at turnout again. Every state Biden won saw increased turnout except for Oklahoma (ignoring states that had a caucus in 2016). Here’s how many more people voted this year than 2016: Alabama +50, 000, Arkansas +11,000, Massachusetts +200,000, NC +200,0000, Oklahoma -30,000, Tennessee +145,000, TX +600,0000, VA +517,000.

The states Biden won on Tuesday mostly saw MASSIVE gains in turnout. The argument that he can’t turn out voters is not evidence based.

So what about subgroups – did Sanders turn out his base demographics and get other demos to vote for him?

The answer is no with one exception – Sanders has done VERY well with Hispanic voters. I’ve not researched why that’s the case but he absolutely deserves credit for that. The percentage of Hispanics he won is much higher than it was in 2016.

The problem is the youth vote and the African American vote.

The argument for Sanders’s electability hinges on young people being energized and turning out in large numbers. Political scientists have been skeptical of this as youth turnout is traditionally awful. The political scientists were right.

Sanders EASILY won the youth vote, but there comparatively weren’t many of them. None of the Super Tuesday states exceeded 20% of the voters falling into the youngest demo. In North Carolina which saw a 17% surge in turnout, youth turnout DROPPED by 9%. There were states that saw increases but they were moderate. Simply put, young people continued to not vote, which is typical of all election cycles.

Now for African Americans, Biden blew Sanders out of the water. These are the margins by which Biden won the African American vote last week as compared to Sanders (not every state has data) : AL +62%, CA +17%, MA +9%, MN +4%, NC +45%, TN +38%, VA +52%. Those numbers do see a dramatic shift when broken down by age, but that runs into the same problem Sanders has in general – young African Americans largely did not vote. Older African Americans did turn out and they overwhelmingly turned out for Biden – in increased numbers from 2016.

There are still paths for Sanders to win but they’re increasingly unlikely and may close tonight depending on his performance in Michigan and Washington.

But the arguments that Sanders will turn out voters and can form broad coalitions simply are not supported by data. Neither are the arguments that Biden can’t increase turnout and can’t form a broad coalition. Biden has successfully done both while expending relatively few resources. Sanders simply has not.

There isn’t blame here. If your cohort groups are mostly Sanders supporters, you are naturally going to extrapolate that into a wider popularity. Several journalists have been shocked by the overwhelming support Biden is receiving from southern African Americans. Their surprise is for a simple reason – they don’t interact with Southern African Americans much. (I’m excluding Texas from “southern” here. We are our own thing.)

Sanders needed to make inroads with groups like these in order to win the nomination. He hasn’t. He needed to see an increase in youth voters. He hasn’t. I’ve seen lots of posts about inspiration, rally attendance, etc… but none of those things matter at all when they don’t translate to votes.

(All information was taken from the Washington Post and New York Times 2012 and 2016 primary pages)


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s