Love, Hate, and Polls in the Time of Polarization
Love, Hate, and Polls in the Time of Polarization
The conversation around Presidential polling is outdated and oversimplistic
This Sunday saw the traditional release of a slew of pre-State of the Union polls. In one sense. Joe Biden’s report card was quite good. In the ABC/Washington Post, NBC/Wall Street Journal, and CBS/YouGov polls, a majority of Americans approve of Biden’s job performance. Even larger majorities approve of his efforts to control the Coronavirus and get people vaccinated.
But there is another, more alarming way to look at these numbers. In the last three months, Joe Biden has mailed checks to millions of Americans, and more than 200 million vaccine doses have been stuck in people’s arms. Yet, according to FiveThirtyEight, Biden’s approval rating is 53.8 percent — just about two points higher than the 51.3 percent of the vote he received in November.
Giving people cash and saving them from a potentially deadly virus is only worth two points!
To be clear, these results say nothing about Joe Biden, the popularity of his agenda, or his communications strategy. But it says everything about the rapid increase in political polarization. The coverage and conversation around these polls demonstrate a dangerously outdated view of American politics and fail to account for how polarization changes public opinion and political strategy.
Stop the Historical Comparisons
It was a tough morning for the staffer in charge of the ABC News Twitter account. A tweet that was almost certainly scheduled to go out automatically after the embargo lifted on their new poll caused a bit of a firestorm online.
The primary problem with this since-deleted tweet is that it is inaccurate. Biden’s 52 percent approval rating is not the lowest since 1945. It is, however, the third-lowest in that time period. The lowest? Donald Trump at 42 percent.
In the article, ABC News unhelpfully points out that “for the 14 presidents from Truman to Biden, the 100-day average is 66 percent.” Once again, that tells us almost nothing about Joe Biden’s performance, political standing, or the prospects for the 2022 midterms. Polarization has been getting worse with every subsequent President, which means fewer members of the opposite party are willing to tell pollsters that they approve of the job the President is doing.
Pollsters measure polarization in terms of the gap between how Republicans and Democrats rate the president’s job performance. For example, according to Gallup, Donald Trump’s third year in office was the most polarizing in history because 89 percent of Republicans approved of Trump, and only 7 percent of Democrats did. To give you a sense of the trend toward greater polarization, the top ten most polarizing years include Trump’s first year, five years Obama was President, and three years of George W. Bush’s Presidency. The other two Trump years would have made this list, but his approval among Republicans was lower because he kept trying to kick people off their healthcare and “both sides” Nazism. By the end of Bush’s Presidency, he became the “uniter, not a divider” he promised to be by uniting the country in agreement over their disapproval of his disastrous presidency.
As an example of how much things have changed, Bill Clinton’s approval rating among Republicans was 41 percent in a Gallup poll immediately after being impeached by a Republican Congress.
Math is math. It is simply mathematically impossible for Biden to have a 60 percent approval rating when only 13 percent of Republicans tell pollsters they approve of the job he has done. These numbers say a lot more about polarization than Biden’s performance.
Partisanship now dictates Presidential approval. In the ABC/Washington Post poll, Biden gets much higher marks for the specific things he is doing — 64 percent approve of his handling of the pandemic, 65 percent approve his American Rescue Plan, and 58 percent approve of his proposal to raise corporate taxes. While the crosstabs were not available (or at least I couldn’t find them), it is mathematically safe to say that his approval on these specific measures is much higher among Republicans than his overall approval rating. In other words, more Republicans approve of the actual job Joe Biden is doing than are willing to tell pollsters because partisanship trumps reality. To be fair, the same was true with Democrats during the Trump era when it came to the economy. Trump’s overall approval among Democrats was in the single digits, but his approval on the economy was often north of 20 percent.
Comparing Biden’s numbers to presidents from previous decades is a disservice to the reader. It gives an incorrect impression and sets impossible expectations.
It’s Better to be (not) Feared than Loved
The media and political strategists are obsessed with Presidential approval numbers. As Perry Bacon wrote in a piece for FiveThirtyEight:
First of all, we don’t yet have a lot of other data to rely on … Second and more importantly, presidential approval ratings in recent years have been a decent indicator of what will happen in the midterms. In the last four (2006, 2010, 2014, 2018), the incumbent president’s disapproval rating was higher than his approval, and in all four cases, the president’s party lost a sizable bloc of House seats.
I am not here to argue that approval ratings do not matter. They do. If Biden’s approval rating is high going into the midterms, I will feel cautiously optimistic. If it is low, I will be hiding under my desk doom scrolling through Twitter1. But the focus on the topline approval rating oversimplifies the complex factors at play in the election. There are real limits to relying on approval rating as a predictor of electoral success. Donald Trump’s approval rating in 2018 was so abysmal that his party lost 40 House seats, yet Republicans gained two seats in the Senate.
The conversation around Biden’s poll numbers and presidential approval more generally undervalue two important political dynamics.
First, political analysis often presumes a static electorate from year to year where different results come from voters changing their minds. In reality, the biggest difference is about who turns out. Obama got crushed in 2010 because many of his voters stayed home and won in 2012 because they turned out. Of course, persuasion matters. There are swing voters, but if you don’t turn out your base, there aren’t enough swing voters to make up the difference.
Second, we live in an era of negative partisanship — where hatred for the other party is the biggest driving factor in political action. This is why Biden’s policies can poll in the seventies, and his approval rating can be in the low fifties. Hatred towards Trump was the number one factor in Democratic turnout in 2020.
Therefore, as we think about 2022, we should focus a little more on Biden’s disapproval rating. In the aforementioned ABC/Washington Post poll, only 42 percent of respondents disapprove of Biden’s job performance. Based on recent history, this number is impressively low. At this point in his Presidency, Trump’s disapproval was 53 percent. Biden’s number is only three points higher than Bill Clinton’s at the 100-day mark in a radically less polarized era.
Biden hasn’t gotten Republican voters to like him, but he has prevented them from hating him — a truly remarkable achievement. And one that bodes well for 2022 if it continues. Republicans need high turnout in the midterms, which might be difficult to achieve without Trump on the ballot if they can’t turn Biden into a scary and hated figure.
In the future, I think we should focus less on the overall approval number and look at the gap between approval and disapproval. Politicians that can be loved by their own party while being hated by fewer people in the other party might end up having the most political success.
I recognize this seems like a patently obvious observation, but it is rarely considered in poll analysis. Ultimately, analysts and strategists need to figure out which messages, policies, and candidate attributes drive turnout among their voters and depress it for the other side. Those answers will be impossible to find if we continue to simplistically obsess about Presidential approval ratings.
Who am I kidding? I will be doom scrolling under all scenarios.
You state that it may seem patently obvious - I think your analysis really digs into the complexities that are frequently missing from discussion of polling on the cable shows. My own feeling is how can we rely on polling data that is also only gathering information from such a small number of people. If I learned anything from that quantitative analysis class, it was that sample size and diversity is important, and that outliers tell us important stuff.